Showing 1–50 of 53 results for author: Haeri, S

Nonlinear Model Predictive Control via Sequential Convex Programming for Drone-to-Drone Docking

Authors: Neeraj Balachandar, Shriram Hari, Vishnu R. Unni

Abstract: Autonomous mid-air docking of multi-rotor vehicles under disturbance-driven target motion poses a constrained non-linear trajectory optimization challenge. This work formulates the docking task as a finite-horizon optimal control problem based on a reduced-order nonlinear model augmented with disturbance states. The resulting problem is solved using sequential convex programming within a receding-… ▽ More Autonomous mid-air docking of multi-rotor vehicles under disturbance-driven target motion poses a constrained non-linear trajectory optimization challenge. This work formulates the docking task as a finite-horizon optimal control problem based on a reduced-order nonlinear model augmented with disturbance states. The resulting problem is solved using sequential convex programming within a receding-horizon framework to generate dynamically feasible docking trajectories. State estimation with noisy measurements is incorporated to enable robust relative motion prediction, while trajectory execution is validated in a high-fidelity rigid-body MuJoCo simulation environment. The proposed framework is evaluated for stationary and constant-velocity target motions, demonstrating reliable convergence to the docking interface while satisfying geometric capture constraints. Quantitatively, the method maintains negligible docking-cone violations and terminal state errors within prescribed tolerances, and achieves consistent, safe docking performance for cone half-angles as low as 10 degrees. Robust operation is observed for wind disturbance levels up to a standard deviation of 0.5, while preserving bounded approach velocities and stable control effort. These results demonstrate the effectiveness of the SCP-based trajectory optimization framework for disturbance-robust aerial docking under estimation uncertainty. △ Less

Submitted 11 August, 2026; originally announced August 2026.

Comments: Accepted at IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM) 2026

SOLAR: AI-Powered Speed-of-Light Performance Analysis

Authors: Qijing Huang, Sana Damani, Zhifan Ye, Athinagoras Skiadopoulos, Siva Kumar Sastry Hari, Jason Clemons, Sahil Modi, Jingquan Wang, Aditya Kane, Edward C Lin, Humphrey Shi, Christos Kozyrakis

Abstract: How fast could a deep-learning model run on target hardware, and how far is today's implementation from that limit? These questions are central to software, hardware, and algorithm optimizations. Speed-of-Light (SOL) analysis answers them by computing a workload's theoretical minimum execution time on a given architecture. Yet deriving SOL bounds remains manual, error-prone, and disconnected from… ▽ More How fast could a deep-learning model run on target hardware, and how far is today's implementation from that limit? These questions are central to software, hardware, and algorithm optimizations. Speed-of-Light (SOL) analysis answers them by computing a workload's theoretical minimum execution time on a given architecture. Yet deriving SOL bounds remains manual, error-prone, and disconnected from rapid model development. To close this gap, we introduce SOLAR, a framework that automatically derives validated SOL bounds from PyTorch and JAX source code. SOLAR leverages both generative and deterministic components in its flow: an LLM frontend translates any source programs into an executable Affine Loop IR, validated by output comparison; a deterministic flow lifts the IR into an einsum graph; and an analytical backend computes unfused, fused, and cache-aware SOL bounds. SOLAR provides comprehensive operator and language coverage, produces validated bounds with zero observed SOL violations, and offers multi-fidelity analysis that tightens bounds and surfaces optimization insights. We evaluate SOLAR across KernelBench, JAX/Flax models, and robotics workloads. These experiments demonstrate four use cases: headroom analysis at multiple fidelity levels, identifying optimization opportunities, cross-platform exploration, and inverse-roofline hardware provisioning. △ Less

Submitted 24 June, 2026; originally announced June 2026.

Regulating Branch Parallelism in LLM Serving

Authors: Swapnil Gandhi, Siva Hari, William J. Dally, Christos Kozyrakis

Abstract: Recent methods expose intra-request parallelism in LLM outputs, allowing independent branches to decode concurrently. Existing serving systems execute these branches eagerly or under fixed caps. We show that both are brittle: eager admission inflates the shared decode step, degrading co-batched requests in serial stages, while conservative fixed caps forgo the throughput that motivated exposing br… ▽ More Recent methods expose intra-request parallelism in LLM outputs, allowing independent branches to decode concurrently. Existing serving systems execute these branches eagerly or under fixed caps. We show that both are brittle: eager admission inflates the shared decode step, degrading co-batched requests in serial stages, while conservative fixed caps forgo the throughput that motivated exposing branches in the first place. We call the excess step latency caused by admitted branches the branch externality and show that the safe width depends on batch composition, context lengths, and accumulated slack, all of which change continuously over a workload trace. We introduce TAPER, a per-step admission controller that treats extra branches as opportunistic work, admitted only when the predicted branch externality fits within the batch's current slack budget. Per-step regulation is practical because branch-level scheduling decouples compute from memory: branches share the request's prefix KV, so expanding or contracting width requires no memory reclamation. On Qwen3-32B, TAPER improves goodput by $1.77\times$ over IRP-Off and by $1.48\times$ over IRP-Eager, while maintaining over $95\%$ SLO attainment. △ Less

Submitted 7 May, 2026; originally announced May 2026.

Improving Efficiency of GPU Kernel Optimization Agents using a Domain-Specific Language and Speed-of-Light Guidance

Authors: Siva Kumar Sastry Hari, Vignesh Balaji, Sana Damani, Qijing Huang, Christos Kozyrakis

Abstract: Optimizing GPU kernels with LLM agents is an iterative process over a large design space. Every candidate must be generated, compiled, validated, and profiled, so fewer trials will save both runtime and cost. We make two key observations. First, the abstraction level that agents operate at is important. If it is too low, the LLM wastes reasoning on low-impact details. If it is too high, it may mis… ▽ More Optimizing GPU kernels with LLM agents is an iterative process over a large design space. Every candidate must be generated, compiled, validated, and profiled, so fewer trials will save both runtime and cost. We make two key observations. First, the abstraction level that agents operate at is important. If it is too low, the LLM wastes reasoning on low-impact details. If it is too high, it may miss important optimization choices. Second, agents cannot easily tell when they reach the point of diminishing returns, wasting resources as they continue searching. These observations motivate two design principles to improve efficiency: (1) a compact domain-specific language (DSL) that can be learned in context and lets the model reason at a higher level while preserving important optimization levers, and (2) Speed-of-Light (SOL) guidance that uses first-principles performance bounds to steer and budget search. We implement these principles in $μ$CUTLASS, a DSL with a compiler for CUTLASS-backed GPU kernels that covers kernel configuration, epilogue fusion, and multi-stage pipelines. We use SOL guidance to estimate headroom and guide optimization trials, deprioritize problems that are near SOL, and flag kernels that game the benchmark. On 59 KernelBench problems with the same iteration budgets, switching from generating low-level code to DSL code using GPT-5-mini turns a 0.40x geomean regression into a 1.27x speedup over PyTorch. Adding SOL-guided steering raises this to 1.56x. Across model tiers, $μ$CUTLASS + SOL-guidance lets weaker models outperform stronger baseline agents at lower token cost. SOL-guided budgeting saves 19-43% of tokens while retaining at least 95% of geomean speedup, with the best policy reaching a 1.68x efficiency gain. Lastly, SOL analysis helps detect benchmark-gaming cases, where kernels may appear fast while failing to perform the intended computation. △ Less

Submitted 30 March, 2026; originally announced March 2026.

SOL-ExecBench: Speed-of-Light Benchmarking for Real-World GPU Kernels Against Hardware Limits

Authors: Edward Lin, Sahil Modi, Siva Kumar Sastry Hari, Qijing Huang, Zhifan Ye, Nestor Qin, Fengzhe Zhou, Yuan Zhang, Jingquan Wang, Sana Damani, Dheeraj Peri, Ouye Xie, Aditya Kane, Moshe Maor, Michael Behar, Triston Cao, Rishabh Mehta, Vartika Singh, Vikram Sharma Mailthody, Terry Chen, Zihao Ye, Hanfeng Chen, Tianqi Chen, Vinod Grover, Wei Chen , et al. (8 additional authors not shown)

Abstract: As agentic AI systems become increasingly capable of generating and optimizing GPU kernels, progress is constrained by benchmarks that reward speedup over software baselines rather than proximity to hardware-efficient execution. We present SOL-ExecBench, a benchmark of 235 CUDA kernel optimization problems extracted from 124 production and emerging AI models spanning language, diffusion, vision, a… ▽ More As agentic AI systems become increasingly capable of generating and optimizing GPU kernels, progress is constrained by benchmarks that reward speedup over software baselines rather than proximity to hardware-efficient execution. We present SOL-ExecBench, a benchmark of 235 CUDA kernel optimization problems extracted from 124 production and emerging AI models spanning language, diffusion, vision, audio, video, and hybrid architectures, targeting NVIDIA Blackwell GPUs. The benchmark covers forward and backward workloads across BF16, FP8, and NVFP4, including kernels whose best performance is expected to rely on Blackwell-specific capabilities. Unlike prior benchmarks that evaluate kernels primarily relative to software implementations, SOL-ExecBench measures performance against analytically derived Speed-of-Light (SOL) bounds computed by SOLAR, our pipeline for deriving hardware-grounded SOL bounds, yielding a fixed target for hardware-efficient optimization. We report a SOL Score that quantifies how much of the gap between a release-defined scoring baseline and the hardware SOL bound a candidate kernel closes. To support robust evaluation of agentic optimizers, we additionally provide a sandboxed harness with GPU clock locking, L2 cache clearing, isolated subprocess execution, and static analysis based checks against common reward-hacking strategies. SOL-ExecBench reframes GPU kernel benchmarking from beating a mutable software baseline to closing the remaining gap to hardware Speed-of-Light. △ Less

Submitted 19 March, 2026; originally announced March 2026.

Dynamic Modeling and Attitude Control of a Reaction-Wheel-Based Low-Gravity Bipedal Hopper

Authors: Shriram Hari, M Venkata Sai Nikhil, R Prasanth Kumar

Abstract: Planetary bodies characterized by low gravitational acceleration, such as the Moon and near-Earth asteroids, impose unique locomotion constraints due to diminished contact forces and extended airborne intervals. Among traversal strategies, hopping locomotion offers high energy efficiency but is prone to mid-flight attitude instability caused by asymmetric thrust generation and uneven terrain inter… ▽ More Planetary bodies characterized by low gravitational acceleration, such as the Moon and near-Earth asteroids, impose unique locomotion constraints due to diminished contact forces and extended airborne intervals. Among traversal strategies, hopping locomotion offers high energy efficiency but is prone to mid-flight attitude instability caused by asymmetric thrust generation and uneven terrain interactions. This paper presents an underactuated bipedal hopping robot that employs an internal reaction wheel to regulate body posture during the ballistic flight phase. The system is modeled as a gyrostat, enabling analysis of the dynamic coupling between torso rotation and reaction wheel momentum. The locomotion cycle comprises three phases: a leg-driven propulsive jump, mid-air attitude stabilization via an active momentum exchange controller, and a shock-absorbing landing. A reduced-order model is developed to capture the critical coupling between torso rotation and reaction wheel dynamics. The proposed framework is evaluated in MuJoCo-based simulations under lunar gravity conditions (g = 1.625 m/s^2). Results demonstrate that activation of the reaction wheel controller reduces peak mid-air angular deviation by more than 65% and constrains landing attitude error to within 3.5 degrees at touchdown. Additionally, actuator saturation per hop cycle is reduced, ensuring sufficient control authority. Overall, the approach significantly mitigates in-flight attitude excursions and enables consistent upright landings, providing a practical and control-efficient solution for locomotion on irregular extraterrestrial terrains. △ Less

Submitted 11 March, 2026; originally announced March 2026.

Comments: Preprint. Under review

KernelBlaster: Continual Cross-Task CUDA Optimization via Memory-Augmented In-Context Reinforcement Learning

Authors: Kris Shengjun Dong, Sahil Modi, Dima Nikiforov, Sana Damani, Edward Lin, Siva Kumar Sastry Hari, Christos Kozyrakis

Abstract: Optimizing CUDA code across multiple generations of GPU architectures is challenging, as achieving peak performance requires an extensive exploration of an increasingly complex, hardware-specific optimization space. Traditional compilers are constrained by fixed heuristics, whereas finetuning Large Language Models (LLMs) can be expensive. However, agentic workflows for CUDA code optimization have… ▽ More Optimizing CUDA code across multiple generations of GPU architectures is challenging, as achieving peak performance requires an extensive exploration of an increasingly complex, hardware-specific optimization space. Traditional compilers are constrained by fixed heuristics, whereas finetuning Large Language Models (LLMs) can be expensive. However, agentic workflows for CUDA code optimization have limited ability to aggregate knowledge from prior exploration, leading to biased sampling and suboptimal solutions. We propose KernelBlaster, a Memory-Augmented In-context Reinforcement Learning (MAIC-RL) framework designed to improve CUDA optimization search capabilities of LLM-based GPU coding agents. KernelBlaster enables agents to learn from experience and make systematically informed decisions on future tasks by accumulating knowledge into a retrievable Persistent CUDA Knowledge Base. We propose a novel profile-guided, textual-gradient-based agentic flow for CUDA generation and optimization to achieve high performance across generations of GPU architectures. KernelBlaster guides LLM agents to systematically explore high-potential optimization strategies beyond naive rewrites. Compared to the PyTorch baseline, our method achieves geometric mean speedups of 1.43x, 2.50x, and 1.50x on KernelBench Levels 1, 2, and 3, respectively. We release KernelBlaster as an open-source agentic framework, accompanied by a test harness, verification components, and a reproducible evaluation pipeline. △ Less

Submitted 15 February, 2026; originally announced February 2026.

Comments: 15 pages, 33 pages with appendix

ProofWright: Towards Agentic Formal Verification of CUDA

Authors: Bodhisatwa Chatterjee, Drew Zagieboylo, Sana Damani, Siva Hari, Christos Kozyrakis

Abstract: Large Language Models (LLMs) are increasingly used to automatically generate optimized CUDA kernels, substantially improving developer productivity. However, despite rapid generation, these kernels often contain subtle correctness bugs and lack formal safety guarantees. Runtime testing is inherently unreliable - limited input coverage and reward hacking can mask incorrect behavior - while manual f… ▽ More Large Language Models (LLMs) are increasingly used to automatically generate optimized CUDA kernels, substantially improving developer productivity. However, despite rapid generation, these kernels often contain subtle correctness bugs and lack formal safety guarantees. Runtime testing is inherently unreliable - limited input coverage and reward hacking can mask incorrect behavior - while manual formal verification is reliable but cannot scale to match LLM output rates, creating a critical validation bottleneck. We present ProofWright, an agentic verification framework that bridges this gap by integrating automated formal verification with LLM-based code generation. ProofWright provides end-to-end guarantees of memory safety, thread safety, and semantic correctness for LLM-generated CUDA kernels. On KernelBench L1, ProofWright verifies safety properties for 74% of generated kernels, uncovers subtle correctness errors missed by conventional testing, and establishes semantic equivalence for a class of element-wise kernels. With a modest overhead of 3 minutes per kernel, ProofWright demonstrates that scalable, automated formal verification of LLM-generated GPU code is feasible - offering a path toward trustworthy high-performance code generation without sacrificing developer productivity. △ Less

Submitted 18 March, 2026; v1 submitted 15 November, 2025; originally announced November 2025.

GO-GAN: Geometry Optimization Generative Adversarial Network for Achieving Optimized Structures with Targeted Physical Properties

Authors: A. Padmaprabhan, Shriram Hari, Nived Philip Thomas, Khaish Singh Chadha, Sai Sidhardh, Viswanath Chinthapenta, Prabhat Kumar

Abstract: This paper presents GO-GAN, a novel Generative Adversarial Network (GAN) architecture for geometry optimization (GO), specifically to generate structures based on user-specified input parameters. The architecture for GO-GAN proposed here combines a \texttt{Pix2Pix} GAN with a new input mechanism, involving a dynamic batch gradient descent-based training loop that leverages dataset symmetries. The… ▽ More This paper presents GO-GAN, a novel Generative Adversarial Network (GAN) architecture for geometry optimization (GO), specifically to generate structures based on user-specified input parameters. The architecture for GO-GAN proposed here combines a \texttt{Pix2Pix} GAN with a new input mechanism, involving a dynamic batch gradient descent-based training loop that leverages dataset symmetries. The model, implemented here using \texttt{TensorFlow} and \texttt{Keras}, is trained using input images representing scalar physical properties generated by a custom MatLab code. After training, GO-GAN rapidly generates optimized geometries from input images representing scalar inputs of the physical properties. Results demonstrate GO-GAN's ability to produce acceptable designs with desirable variations. These variations are followed by the influence of discriminators during training and are of practical significance in ensuring adherence to specifications while enabling creative exploration of the design space. △ Less

Submitted 1 February, 2025; originally announced February 2025.

Comments: iNCMDAO 2024

Remote Manipulation of Multiple Objects with Airflow Field Using Model-Based Learning Control

Authors: Artur Kopitca, Shahriar Haeri, Quan Zhou

Abstract: Non-contact manipulation is a promising methodology in robotics, offering a wide range of scientific and industrial applications. Among the proposed approaches, airflow stands out for its ability to project across considerable distances and its flexibility in actuating objects of varying materials, sizes, and shapes. However, predicting airflow fields at a distance-and the motion of objects within… ▽ More Non-contact manipulation is a promising methodology in robotics, offering a wide range of scientific and industrial applications. Among the proposed approaches, airflow stands out for its ability to project across considerable distances and its flexibility in actuating objects of varying materials, sizes, and shapes. However, predicting airflow fields at a distance-and the motion of objects within them-remains notoriously challenging due to their nonlinear and stochastic nature. Here, we propose a model-based learning approach using a jet-induced airflow field for remote multi-object manipulation on a surface. Our approach incorporates an analytical model of the field, learned object dynamics, and a model-based controller. The model predicts an air velocity field over an infinite surface for a specified jet orientation, while the object dynamics are learned through a robust system identification algorithm. Using the model-based controller, we can automatically and remotely, at meter-scale distances, control the motion of single and multiple objects for different tasks, such as path-following, aggregating, and sorting. △ Less

Submitted 15 May, 2025; v1 submitted 4 December, 2024; originally announced December 2024.

Comments: 8 pages, 7 figures

Journal ref: IEEE/ASME Transactions on Mechatronics Volume: 30, Issue: 4, Aug. 2025

Command-line Risk Classification using Transformer-based Neural Architectures

Authors: Paolo Notaro, Soroush Haeri, Jorge Cardoso, Michael Gerndt

Abstract: To protect large-scale computing environments necessary to meet increasing computing demand, cloud providers have implemented security measures to monitor Operations and Maintenance (O&M) activities and therefore prevent data loss and service interruption. Command interception systems are used to intercept, assess, and block dangerous Command-line Interface (CLI) commands before they can cause dam… ▽ More To protect large-scale computing environments necessary to meet increasing computing demand, cloud providers have implemented security measures to monitor Operations and Maintenance (O&M) activities and therefore prevent data loss and service interruption. Command interception systems are used to intercept, assess, and block dangerous Command-line Interface (CLI) commands before they can cause damage. Traditional solutions for command risk assessment include rule-based systems, which require expert knowledge and constant human revision to account for unseen commands. To overcome these limitations, several end-to-end learning systems have been proposed to classify CLI commands. These systems, however, have several other limitations, including the adoption of general-purpose text classifiers, which may not adapt to the language characteristics of scripting languages such as Bash or PowerShell, and may not recognize dangerous commands in the presence of an unbalanced class distribution. In this paper, we propose a transformer-based command risk classification system, which leverages the generalization power of Large Language Models (LLM) to provide accurate classification and the ability to identify rare dangerous commands effectively, by exploiting the power of transfer learning. We verify the effectiveness of our approach on a realistic dataset of production commands and show how to apply our model for other security-related tasks, such as dangerous command interception and auditing of existing rule-based systems. △ Less

Submitted 2 December, 2024; originally announced December 2024.

LSST: Learned Single-Shot Trajectory and Reconstruction Network for MR Imaging

Authors: Hemant Kumar Aggarwal, Sudhanya Chatterjee, Dattesh Shanbhag, Uday Patil, K. V. S. Hari

Abstract: Single-shot magnetic resonance (MR) imaging acquires the entire k-space data in a single shot and it has various applications in whole-body imaging. However, the long acquisition time for the entire k-space in single-shot fast spin echo (SSFSE) MR imaging poses a challenge, as it introduces T2-blur in the acquired images. This study aims to enhance the reconstruction quality of SSFSE MR images by… ▽ More Single-shot magnetic resonance (MR) imaging acquires the entire k-space data in a single shot and it has various applications in whole-body imaging. However, the long acquisition time for the entire k-space in single-shot fast spin echo (SSFSE) MR imaging poses a challenge, as it introduces T2-blur in the acquired images. This study aims to enhance the reconstruction quality of SSFSE MR images by (a) optimizing the trajectory for measuring the k-space, (b) acquiring fewer samples to speed up the acquisition process, and (c) reducing the impact of T2-blur. The proposed method adheres to physics constraints due to maximum gradient strength and slew-rate available while optimizing the trajectory within an end-to-end learning framework. Experiments were conducted on publicly available fastMRI multichannel dataset with 8-fold and 16-fold acceleration factors. An experienced radiologist's evaluation on a five-point Likert scale indicates improvements in the reconstruction quality as the ACL fibers are sharper than comparative methods. △ Less

Submitted 8 August, 2024; originally announced September 2024.

Leveraging Open-Source Large Language Models for encoding Social Determinants of Health using an Intelligent Router

Authors: Akul Goel, Surya Narayanan Hari, Belinda Waltman, Matt Thomson

Abstract: Social Determinants of Health (SDOH), also known as Health-Related Social Needs (HSRN), play a significant role in patient health outcomes. The Centers for Disease Control and Prevention (CDC) introduced a subset of ICD-10 codes called Z-codes to recognize and measure SDOH. However, Z-codes are infrequently coded in a patient's Electronic Health Record (EHR), and instead, in many cases, need to be… ▽ More Social Determinants of Health (SDOH), also known as Health-Related Social Needs (HSRN), play a significant role in patient health outcomes. The Centers for Disease Control and Prevention (CDC) introduced a subset of ICD-10 codes called Z-codes to recognize and measure SDOH. However, Z-codes are infrequently coded in a patient's Electronic Health Record (EHR), and instead, in many cases, need to be inferred from clinical notes. Previous research has shown that large language models (LLMs) show promise on extracting unstructured data from EHRs, but it can be difficult to identify a single model that performs best on varied coding tasks. Further, clinical notes contain protected health information posing a challenge for the use of closed-source language models from commercial vendors. The identification of open-source LLMs that can be run within health organizations and exhibit high performance on SDOH tasks is an important issue to solve. Here, we introduce an intelligent routing system for SDOH coding that uses a language model router to direct medical record data to open-source LLMs that demonstrate optimal performance on specific SDOH codes. This intelligent routing system exhibits state of the art performance of 96.4% accuracy averaged across 13 codes, including homelessness and food insecurity, outperforming closed models such as GPT-4o. We leveraged a publicly-available, deidentified dataset of medical record notes to run the router, but we also introduce a synthetic data generation and validation paradigm to increase the scale of training data without needing privacy-protected medical records. Together, we demonstrate an architecture for intelligent routing of inputs to task-optimal language models to achieve high performance across a set of medical coding sub-tasks. △ Less

Submitted 15 January, 2026; v1 submitted 29 May, 2024; originally announced May 2024.

MPGemmFI: A Fault Injection Technique for Mixed Precision GEMM in ML Applications

Authors: Bo Fang, Xinyi Li, Harvey Dam, Cheng Tan, Siva Kumar Sastry Hari, Timothy Tsai, Ignacio Laguna, Dingwen Tao, Ganesh Gopalakrishnan, Prashant Nair, Kevin Barker, Ang Li

Abstract: Emerging deep learning workloads urgently need fast general matrix multiplication (GEMM). To meet such demand, one of the critical features of machine-learning-specific accelerators such as NVIDIA Tensor Cores, AMD Matrix Cores, and Google TPUs is the support of mixed-precision enabled GEMM. For DNN models, lower-precision FP data formats and computation offer acceptable correctness but significan… ▽ More Emerging deep learning workloads urgently need fast general matrix multiplication (GEMM). To meet such demand, one of the critical features of machine-learning-specific accelerators such as NVIDIA Tensor Cores, AMD Matrix Cores, and Google TPUs is the support of mixed-precision enabled GEMM. For DNN models, lower-precision FP data formats and computation offer acceptable correctness but significant performance, area, and memory footprint improvement. While promising, the mixed-precision computation on error resilience remains unexplored. To this end, we develop a fault injection framework that systematically injects fault into the mixed-precision computation results. We investigate how the faults affect the accuracy of machine learning applications. Based on the error resilience characteristics, we offer lightweight error detection and correction solutions that significantly improve the overall model accuracy if the models experience hardware faults. The solutions can be efficiently integrated into the accelerator's pipelines. △ Less

Submitted 9 November, 2023; originally announced November 2023.

Herd: Using multiple, smaller LLMs to match the performances of proprietary, large LLMs via an intelligent composer

Authors: Surya Narayanan Hari, Rex Liu, Matt Thomson

Abstract: Currently, over a thousand LLMs exist that are multi-purpose and are capable of performing real world tasks, including Q&A, text summarization, content generation, etc. However, accessibility, scale and reliability of free models prevents them from being widely deployed in everyday use cases. To address the first two issues of access and scale, organisations such as HuggingFace have created model… ▽ More Currently, over a thousand LLMs exist that are multi-purpose and are capable of performing real world tasks, including Q&A, text summarization, content generation, etc. However, accessibility, scale and reliability of free models prevents them from being widely deployed in everyday use cases. To address the first two issues of access and scale, organisations such as HuggingFace have created model repositories where users have uploaded model weights and quantized versions of models trained using different paradigms, as well as model cards describing their training process. While some models report performance on commonly used benchmarks, not all do, and interpreting the real world impact of trading off performance on a benchmark for model deployment cost, is unclear. Here, we show that a herd of open source models can match or exceed the performance of proprietary models via an intelligent router. We show that a Herd of open source models is able to match the accuracy of ChatGPT, despite being composed of models that are effectively 2.5x smaller. We show that in cases where GPT is not able to answer the query, Herd is able to identify a model that can, at least 40% of the time. △ Less

Submitted 21 September, 2024; v1 submitted 30 October, 2023; originally announced October 2023.

cuRobo: Parallelized Collision-Free Minimum-Jerk Robot Motion Generation

Authors: Balakumar Sundaralingam, Siva Kumar Sastry Hari, Adam Fishman, Caelan Garrett, Karl Van Wyk, Valts Blukis, Alexander Millane, Helen Oleynikova, Ankur Handa, Fabio Ramos, Nathan Ratliff, Dieter Fox

Abstract: This paper explores the problem of collision-free motion generation for manipulators by formulating it as a global motion optimization problem. We develop a parallel optimization technique to solve this problem and demonstrate its effectiveness on massively parallel GPUs. We show that combining simple optimization techniques with many parallel seeds leads to solving difficult motion generation pro… ▽ More This paper explores the problem of collision-free motion generation for manipulators by formulating it as a global motion optimization problem. We develop a parallel optimization technique to solve this problem and demonstrate its effectiveness on massively parallel GPUs. We show that combining simple optimization techniques with many parallel seeds leads to solving difficult motion generation problems within 50ms on average, 60x faster than state-of-the-art (SOTA) trajectory optimization methods. We achieve SOTA performance by combining L-BFGS step direction estimation with a novel parallel noisy line search scheme and a particle-based optimization solver. To further aid trajectory optimization, we develop a parallel geometric planner that plans within 20ms and also introduce a collision-free IK solver that can solve over 7000 queries/s. We package our contributions into a state of the art GPU accelerated motion generation library, cuRobo and release it to enrich the robotics community. Additional details are available at https://curobo.org △ Less

Submitted 3 November, 2023; v1 submitted 26 October, 2023; originally announced October 2023.

Comments: revised technical report, 62 pages, Website: https://curobo.org

VaPr: Variable-Precision Tensors to Accelerate Robot Motion Planning

Authors: Yu-Shun Hsiao, Siva Kumar Sastry Hari, Balakumar Sundaralingam, Jason Yik, Thierry Tambe, Charbel Sakr, Stephen W. Keckler, Vijay Janapa Reddi

Abstract: High-dimensional motion generation requires numerical precision for smooth, collision-free solutions. Typically, double-precision or single-precision floating-point (FP) formats are utilized. Using these for big tensors imposes a strain on the memory bandwidth provided by the devices and alters the memory footprint, hence limiting their applicability to low-power edge devices needed for mobile rob… ▽ More High-dimensional motion generation requires numerical precision for smooth, collision-free solutions. Typically, double-precision or single-precision floating-point (FP) formats are utilized. Using these for big tensors imposes a strain on the memory bandwidth provided by the devices and alters the memory footprint, hence limiting their applicability to low-power edge devices needed for mobile robots. The uniform application of reduced precision can be advantageous but severely degrades solutions. Using decreased precision data types for important tensors, we propose to accelerate motion generation by removing memory bottlenecks. We propose variable-precision (VaPr) search optimization to determine the appropriate precision for large tensors from a vast search space of approximately 4 million unique combinations for FP data types across the tensors. To obtain the efficiency gains, we exploit existing platform support for an out-of-the-box GPU speedup and evaluate prospective precision converter units for GPU types that are not currently supported. Our experimental results on 800 planning problems for the Franka Panda robot on the MotionBenchmaker dataset across 8 environments show that a 4-bit FP format is sufficient for the largest set of tensors in the motion generation stack. With the software-only solution, VaPr achieves 6.3% and 6.3% speedups on average for a significant portion of motion generation over the SOTA solution (CuRobo) on Jetson Orin and RTX2080 Ti GPU, respectively, and 9.9%, 17.7% speedups with the FP converter. △ Less

Submitted 11 October, 2023; originally announced October 2023.

Comments: 7 pages, 5 figures, 8 tables, to be published in 2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

ALBERTA: ALgorithm-Based Error Resilience in Transformer Architectures

Authors: Haoxuan Liu, Vasu Singh, Michał Filipiuk, Siva Kumar Sastry Hari

Abstract: Vision Transformers are being increasingly deployed in safety-critical applications that demand high reliability. It is crucial to ensure the correctness of their execution in spite of potential errors such as transient hardware errors. We propose a novel algorithm-based resilience framework called ALBERTA that allows us to perform end-to-end resilience analysis and protection of transformer-based… ▽ More Vision Transformers are being increasingly deployed in safety-critical applications that demand high reliability. It is crucial to ensure the correctness of their execution in spite of potential errors such as transient hardware errors. We propose a novel algorithm-based resilience framework called ALBERTA that allows us to perform end-to-end resilience analysis and protection of transformer-based architectures. First, our work develops an efficient process of computing and ranking the resilience of transformers layers. We find that due to the large size of transformer models, applying traditional network redundancy to a subset of the most vulnerable layers provides high error coverage albeit with impractically high overhead. We address this shortcoming by providing a software-directed, checksum-based error detection technique aimed at protecting the most vulnerable general matrix multiply (GEMM) layers in the transformer models that use either floating-point or integer arithmetic. Results show that our approach achieves over 99% coverage for errors that result in a mismatch with less than 0.2% and 0.01% computation and memory overheads, respectively. Lastly, we present the applicability of our framework in various modern GPU architectures under different numerical precisions. We introduce an efficient self-correction mechanism for resolving erroneous detection with an average of less than 2% overhead per error. △ Less

Submitted 5 February, 2024; v1 submitted 5 October, 2023; originally announced October 2023.

Tryage: Real-time, intelligent Routing of User Prompts to Large Language Models

Authors: Surya Narayanan Hari, Matt Thomson

Abstract: The introduction of the transformer architecture and the self-attention mechanism has led to an explosive production of language models trained on specific downstream tasks and data domains. With over 200, 000 models in the Hugging Face ecosystem, users grapple with selecting and optimizing models to suit multifaceted workflows and data domains while addressing computational, security, and recency… ▽ More The introduction of the transformer architecture and the self-attention mechanism has led to an explosive production of language models trained on specific downstream tasks and data domains. With over 200, 000 models in the Hugging Face ecosystem, users grapple with selecting and optimizing models to suit multifaceted workflows and data domains while addressing computational, security, and recency concerns. There is an urgent need for machine learning frameworks that can eliminate the burden of model selection and customization and unleash the incredible power of the vast emerging model library for end users. Here, we propose a context-aware routing system, Tryage, that leverages a language model router for optimal selection of expert models from a model library based on analysis of individual input prompts. Inspired by the thalamic router in the brain, Tryage employs a perceptive router to predict down-stream model performance on prompts and, then, makes a routing decision using an objective function that integrates performance predictions with user goals and constraints that are incorporated through flags (e.g., model size, model recency). Tryage allows users to explore a Pareto front and automatically trade-off between task accuracy and secondary goals including minimization of model size, recency, security, verbosity, and readability. Across heterogeneous data sets that include code, text, clinical data, and patents, the Tryage framework surpasses Gorilla and GPT3.5 turbo in dynamic model selection identifying the optimal model with an accuracy of 50.9% , compared to 23.6% by GPT 3.5 Turbo and 10.8% by Gorilla. Conceptually, Tryage demonstrates how routing models can be applied to program and control the behavior of multi-model LLM systems to maximize efficient use of the expanding and evolving language model ecosystem. △ Less

Submitted 23 August, 2023; v1 submitted 22 August, 2023; originally announced August 2023.

Algebraic Reasoning About Timeliness

Authors: Seyed Hossein Haeri, Peter W. Thompson, Peter Van Roy, Magne Haveraaen, Neil J. Davies, Mikhail Barash, Kevin Hammond, James Chapman

Abstract: Designing distributed systems to have predictable performance under high load is difficult because of resource exhaustion, non-linearity, and stochastic behaviour. Timeliness, i.e., delivering results within defined time bounds, is a central aspect of predictable performance. In this paper, we focus on timeliness using the DELTA-Q Systems Development paradigm (DELTA-QSD, developed by PNSol), which… ▽ More Designing distributed systems to have predictable performance under high load is difficult because of resource exhaustion, non-linearity, and stochastic behaviour. Timeliness, i.e., delivering results within defined time bounds, is a central aspect of predictable performance. In this paper, we focus on timeliness using the DELTA-Q Systems Development paradigm (DELTA-QSD, developed by PNSol), which computes timeliness by modelling systems observationally using so-called outcome expressions. An outcome expression is a compositional definition of a system's observed behaviour in terms of its basic operations. Given the behaviour of the basic operations, DELTA-QSD efficiently computes the stochastic behaviour of the whole system including its timeliness. This paper formally proves useful algebraic properties of outcome expressions w.r.t. timeliness. We prove the different algebraic structures the set of outcome expressions form with the different DELTA-QSD operators and demonstrate why those operators do not form richer structures. We prove or disprove the set of all possible distributivity results on outcome expressions. On our way for disproving 8 of those distributivity results, we develop a technique called properisation, which gives rise to the first body of maths for improper random variables. Finally, we also prove 14 equivalences that have been used in the past in the practice of DELTA-QSD. An immediate benefit is rewrite rules that can be used for design exploration under established timeliness equivalence. This work is part of an ongoing project to disseminate and build tool support for DELTA-QSD. The ability to rewrite outcome expressions is essential for efficient tool support. △ Less

Submitted 21 August, 2023; originally announced August 2023.

Comments: In Proceedings ICE 2023, arXiv:2308.08920

ACM Class: B.8.2; C.4; D.2.4; D.2.8; F.3.2; F.3.1; F.4.1; F.4.3; I.1.1

Journal ref: EPTCS 383, 2023, pp. 35-54

Unfolding for Joint Channel Estimation and Symbol Detection in MIMO Communication Systems

Authors: Swati Bhattacharya, K. V. S. Hari, Yonina C. Eldar

Abstract: This paper proposes a Joint Channel Estimation and Symbol Detection (JED) scheme for Multiple-Input Multiple-Output (MIMO) wireless communication systems. Our proposed method for JED using Alternating Direction Method of Multipliers (JED-ADMM) and its model-based neural network version JED using Unfolded ADMM (JED-U-ADMM) markedly improve the symbol detection performance over JED using Alternating… ▽ More This paper proposes a Joint Channel Estimation and Symbol Detection (JED) scheme for Multiple-Input Multiple-Output (MIMO) wireless communication systems. Our proposed method for JED using Alternating Direction Method of Multipliers (JED-ADMM) and its model-based neural network version JED using Unfolded ADMM (JED-U-ADMM) markedly improve the symbol detection performance over JED using Alternating Minimization (JED-AM) for a range of MIMO antenna configurations. Both proposed algorithms exploit the non-smooth constraint, that occurs as a result of the Quadrature Amplitude Modulation (QAM) data symbols, to effectively improve the performance using the ADMM iterations. The proposed unfolded network JED-U-ADMM consists of a few trainable parameters and requires a small training set. We show the efficacy of the proposed methods for both uncorrelated and correlated MIMO channels. For certain configurations, the gain in SNR for a desired BER of $10^{-2}$ for the proposed JED-ADMM and JED-U-ADMM is upto $4$ dB and is also accompanied by a significant reduction in computational complexity of upto $75\%$, depending on the MIMO configuration, as compared to the complexity of JED-AM. △ Less

Submitted 21 August, 2023; v1 submitted 17 August, 2023; originally announced August 2023.

Comments: 14 pages, 19 figures, submitted to IEEE Transactions on Signal Processing

Safety-Critical Scenario Generation Via Reinforcement Learning Based Editing

Authors: Haolan Liu, Liangjun Zhang, Siva Kumar Sastry Hari, Jishen Zhao

Abstract: Generating safety-critical scenarios is essential for testing and verifying the safety of autonomous vehicles. Traditional optimization techniques suffer from the curse of dimensionality and limit the search space to fixed parameter spaces. To address these challenges, we propose a deep reinforcement learning approach that generates scenarios by sequential editing, such as adding new agents or mod… ▽ More Generating safety-critical scenarios is essential for testing and verifying the safety of autonomous vehicles. Traditional optimization techniques suffer from the curse of dimensionality and limit the search space to fixed parameter spaces. To address these challenges, we propose a deep reinforcement learning approach that generates scenarios by sequential editing, such as adding new agents or modifying the trajectories of the existing agents. Our framework employs a reward function consisting of both risk and plausibility objectives. The plausibility objective leverages generative models, such as a variational autoencoder, to learn the likelihood of the generated parameters from the training datasets; It penalizes the generation of unlikely scenarios. Our approach overcomes the dimensionality challenge and explores a wide range of safety-critical scenarios. Our evaluation demonstrates that the proposed method generates safety-critical scenarios of higher quality compared with previous approaches. △ Less

Submitted 6 March, 2024; v1 submitted 25 June, 2023; originally announced June 2023.

Circuit simulation using explicit methods

Authors: Mahesh B. Patil, V. V. S. Pavan Kumar Hari

Abstract: Use of explicit methods for simulating electrical circuits, especially for power electronics applications, is described. Application of the forward Euler method to a half-wave rectifier is discussed, and the limitations of a fixed-step method are pointed out. Implementation of the Runge-Kutta-Fehlberg (RKF) method, which allows variable time steps, for the half-wave rectifier circuit is discussed,… ▽ More Use of explicit methods for simulating electrical circuits, especially for power electronics applications, is described. Application of the forward Euler method to a half-wave rectifier is discussed, and the limitations of a fixed-step method are pointed out. Implementation of the Runge-Kutta-Fehlberg (RKF) method, which allows variable time steps, for the half-wave rectifier circuit is discussed, and its advantages pointed out. Formulation of circuit equations for the purpose of simulation using the RKF method is described for some more examples. Stability and accuracy issues related to power electronic circuits are brought out, and mechanisms to address them are presented. Future plans related to this work are described. △ Less

Submitted 11 January, 2023; originally announced January 2023.

Comments: 13 pages, 22 figures

Cooperative Coverage with a Leader and a Wingmate in Communication-Constrained Environments

Authors: Sai Krishna Kanth Hari, Sivakumar Rathinam, Swaroop Darbha, David W. Casbeer

Abstract: We consider a mission framework in which two unmanned vehicles (UVs), a leader and a wingmate, are required to provide cooperative coverage of an environment while being within a short communication range. This framework finds applications in underwater and/or military domains, where certain constraints are imposed on communication by either the application or the environment. An important objecti… ▽ More We consider a mission framework in which two unmanned vehicles (UVs), a leader and a wingmate, are required to provide cooperative coverage of an environment while being within a short communication range. This framework finds applications in underwater and/or military domains, where certain constraints are imposed on communication by either the application or the environment. An important objective of missions within this framework is to minimize the total travel and communication costs of the leader-wingmate duo. In this paper, we propose and formulate the problem of finding routes for the UVs that minimize the sum of their travel and communication costs as a network optimization problem of the form of a binary program (BP). The BP is computationally expensive, with the time required to compute optimal solutions increasing rapidly with the problem size. To address this challenge, here, we propose two algorithms, an approximation algorithm and a heuristic algorithm, to solve large-scale instances of the problem swiftly. We demonstrate the effectiveness and the scalability of these algorithms through an analysis of extensive numerical simulations performed over 500 instances, with the number of targets in the instances ranging from 6 to 100. △ Less

Submitted 5 October, 2022; originally announced October 2022.

Zhuyi: Perception Processing Rate Estimation for Safety in Autonomous Vehicles

Authors: Yu-Shun Hsiao, Siva Kumar Sastry Hari, Michał Filipiuk, Timothy Tsai, Michael B. Sullivan, Vijay Janapa Reddi, Vasu Singh, Stephen W. Keckler

Abstract: The processing requirement of autonomous vehicles (AVs) for high-accuracy perception in complex scenarios can exceed the resources offered by the in-vehicle computer, degrading safety and comfort. This paper proposes a sensor frame processing rate (FPR) estimation model, Zhuyi, that quantifies the minimum safe FPR continuously in a driving scenario. Zhuyi can be employed post-deployment as an onli… ▽ More The processing requirement of autonomous vehicles (AVs) for high-accuracy perception in complex scenarios can exceed the resources offered by the in-vehicle computer, degrading safety and comfort. This paper proposes a sensor frame processing rate (FPR) estimation model, Zhuyi, that quantifies the minimum safe FPR continuously in a driving scenario. Zhuyi can be employed post-deployment as an online safety check and to prioritize work. Experiments conducted using a multi-camera state-of-the-art industry AV system show that Zhuyi's estimated FPRs are conservative, yet the system can maintain safety by processing only 36% or fewer frames compared to a default 30-FPR system in the tested scenarios. △ Less

Submitted 6 May, 2022; originally announced May 2022.

Comments: 2022 Design Automation Conference (DAC), July 10-14, 2022, San Francisco

Engineering flexible machine learning systems by traversing functionally-invariant paths

Authors: Guruprasad Raghavan, Bahey Tharwat, Surya Narayanan Hari, Dhruvil Satani, Matt Thomson

Abstract: Transformers have emerged as the state of the art neural network architecture for natural language processing and computer vision. In the foundation model paradigm, large transformer models (BERT, GPT3/4, Bloom, ViT) are pre-trained on self-supervised tasks such as word or image masking, and then, adapted through fine-tuning for downstream user applications including instruction following and Ques… ▽ More Transformers have emerged as the state of the art neural network architecture for natural language processing and computer vision. In the foundation model paradigm, large transformer models (BERT, GPT3/4, Bloom, ViT) are pre-trained on self-supervised tasks such as word or image masking, and then, adapted through fine-tuning for downstream user applications including instruction following and Question Answering. While many approaches have been developed for model fine-tuning including low-rank weight update strategies (eg. LoRA), underlying mathematical principles that enable network adaptation without knowledge loss remain poorly understood. Here, we introduce a differential geometry framework, functionally invariant paths (FIP), that provides flexible and continuous adaptation of neural networks for a range of machine learning goals and network sparsification objectives. We conceptualize the weight space of a neural network as a curved Riemannian manifold equipped with a metric tensor whose spectrum defines low rank subspaces in weight space that accommodate network adaptation without loss of prior knowledge. We formalize adaptation as movement along a geodesic path in weight space while searching for networks that accommodate secondary objectives. With modest computational resources, the FIP algorithm achieves comparable to state of the art performance on continual learning and sparsification tasks for language models (BERT), vision transformers (ViT, DeIT), and the CNNs. Broadly, we conceptualize a neural network as a mathematical object that can be iteratively transformed into distinct configurations by the path-sampling algorithm to define a sub-manifold of weight space that can be harnessed to achieve user goals. △ Less

Submitted 3 September, 2023; v1 submitted 30 April, 2022; originally announced May 2022.

Comments: 22 pages

An open-source simulation package for power electronics education

Authors: Mahesh B. Patil, V. V. S. Pavan Kumar Hari, Ruchita D. Korgaonkar, Kumar Appaiah

Abstract: Extension of the open-source simulation package GSEIM for power electronics applications is presented. Recent developments in GSEIM, including those oriented specifically towards power electronic circuits, are described. Some examples of electrical element templates, which form a part of the GSEIM library, are discussed. Representative simulation examples in power electronics are presented to brin… ▽ More Extension of the open-source simulation package GSEIM for power electronics applications is presented. Recent developments in GSEIM, including those oriented specifically towards power electronic circuits, are described. Some examples of electrical element templates, which form a part of the GSEIM library, are discussed. Representative simulation examples in power electronics are presented to bring out important features of the simulator. Advantages of GSEIM for educational purposes are discussed. Finally, plans regarding future developments in GSEIM are presented. △ Less

Submitted 25 April, 2022; originally announced April 2022.

Comments: 8 pages, 11 figures

A Contrastive Learning Approach to Auroral Identification and Classification

Authors: Jeremiah W. Johnson, Swathi Hari, Donald Hampton, Hyunju K. Connor, Amy Keesee

Abstract: Unsupervised learning algorithms are beginning to achieve accuracies comparable to their supervised counterparts on benchmark computer vision tasks, but their utility for practical applications has not yet been demonstrated. In this work, we present a novel application of unsupervised learning to the task of auroral image classification. Specifically, we modify and adapt the Simple framework for C… ▽ More Unsupervised learning algorithms are beginning to achieve accuracies comparable to their supervised counterparts on benchmark computer vision tasks, but their utility for practical applications has not yet been demonstrated. In this work, we present a novel application of unsupervised learning to the task of auroral image classification. Specifically, we modify and adapt the Simple framework for Contrastive Learning of Representations (SimCLR) algorithm to learn representations of auroral images in a recently released auroral image dataset constructed using image data from Time History of Events and Macroscale Interactions during Substorms (THEMIS) all-sky imagers. We demonstrate that (a) simple linear classifiers fit to the learned representations of the images achieve state-of-the-art classification performance, improving the classification accuracy by almost 10 percentage points over the current benchmark; and (b) the learned representations naturally cluster into more clusters than exist manually assigned categories, suggesting that existing categorizations are overly coarse and may obscure important connections between auroral types, near-earth solar wind conditions, and geomagnetic disturbances at the earth's surface. Moreover, our model is much lighter than the previous benchmark on this dataset, requiring in the area of fewer than 25\% of the number of parameters. Our approach exceeds an established threshold for operational purposes, demonstrating readiness for deployment and utilization. △ Less

Submitted 28 September, 2021; v1 submitted 28 September, 2021; originally announced September 2021.

Comments: 6 pages, 5 figures, 1 table

Journal ref: Proceedings of the 20th IEEE International Conference on Machine Learning and Applications, Dec. 2021

Bounds on Optimal Revisit Times in Persistent Monitoring Missions with a Distinct \& Remote Service Station

Authors: Sai Krishna Kanth Hari, Sivakumar Rathinam, Swaroop Darbha, Krishna Kalyanam, Satyanarayana Gupta Manyam, David Casbeer

Abstract: Persistent monitoring missions require an up-to-date knowledge of the changing state of the underlying environment. UAVs can be gainfully employed to continually visit a set of targets representing tasks (and locations) in the environment and collect data therein for long time periods. The enduring nature of these missions requires the UAV to be regularly recharged at a service station. In this pa… ▽ More Persistent monitoring missions require an up-to-date knowledge of the changing state of the underlying environment. UAVs can be gainfully employed to continually visit a set of targets representing tasks (and locations) in the environment and collect data therein for long time periods. The enduring nature of these missions requires the UAV to be regularly recharged at a service station. In this paper, we consider the case in which the service station is not co-located with any of the targets. An efficient monitoring requires the revisit time, defined as the maximum of the time elapsed between successive revisits to targets, to be minimized. Here, we consider the problem of determining UAV routes that lead to the minimum revisit time. The problem is NP-hard, and its computational difficulty increases with the fuel capacity of the UAV. We develop an algorithm to construct near-optimal solutions to the problem quickly, when the fuel capacity exceeds a threshold. We also develop lower bounds to the optimal revisit time and use these bounds to demonstrate (through numerical simulations) that the constructed solutions are, on an average, at most 0.01% away from the optimum. △ Less

Submitted 21 May, 2021; originally announced May 2021.

Comments: Submitted to IEEE TRO

Generating and Characterizing Scenarios for Safety Testing of Autonomous Vehicles

Authors: Zahra Ghodsi, Siva Kumar Sastry Hari, Iuri Frosio, Timothy Tsai, Alejandro Troccoli, Stephen W. Keckler, Siddharth Garg, Anima Anandkumar

Abstract: Extracting interesting scenarios from real-world data as well as generating failure cases is important for the development and testing of autonomous systems. We propose efficient mechanisms to both characterize and generate testing scenarios using a state-of-the-art driving simulator. For any scenario, our method generates a set of possible driving paths and identifies all the possible safe drivin… ▽ More Extracting interesting scenarios from real-world data as well as generating failure cases is important for the development and testing of autonomous systems. We propose efficient mechanisms to both characterize and generate testing scenarios using a state-of-the-art driving simulator. For any scenario, our method generates a set of possible driving paths and identifies all the possible safe driving trajectories that can be taken starting at different times, to compute metrics that quantify the complexity of the scenario. We use our method to characterize real driving data from the Next Generation Simulation (NGSIM) project, as well as adversarial scenarios generated in simulation. We rank the scenarios by defining metrics based on the complexity of avoiding accidents and provide insights into how the AV could have minimized the probability of incurring an accident. We demonstrate a strong correlation between the proposed metrics and human intuition. △ Less

Submitted 12 March, 2021; originally announced March 2021.

Making Convolutions Resilient via Algorithm-Based Error Detection Techniques

Authors: Siva Kumar Sastry Hari, Michael B. Sullivan, Timothy Tsai, Stephen W. Keckler

Abstract: The ability of Convolutional Neural Networks (CNNs) to accurately process real-time telemetry has boosted their use in safety-critical and high-performance computing systems. As such systems require high levels of resilience to errors, CNNs must execute correctly in the presence of hardware faults. Full duplication provides the needed assurance but incurs a prohibitive 100% overhead. Algorithmic t… ▽ More The ability of Convolutional Neural Networks (CNNs) to accurately process real-time telemetry has boosted their use in safety-critical and high-performance computing systems. As such systems require high levels of resilience to errors, CNNs must execute correctly in the presence of hardware faults. Full duplication provides the needed assurance but incurs a prohibitive 100% overhead. Algorithmic techniques are known to offer low-cost solutions, but the practical feasibility and performance of such techniques have never been studied for CNN deployment platforms (e.g., TensorFlow or TensorRT on GPUs). In this paper, we focus on algorithmically verifying Convolutions, which are the most resource-demanding operations in CNNs. We use checksums to verify convolutions, adding a small amount of redundancy, far less than full-duplication. We first identify the challenges that arise in employing Algorithm-Based Error Detection (ABED) for Convolutions in optimized inference platforms that fuse multiple network layers and use reduced-precision operations, and demonstrate how to overcome them. We propose and evaluate variations of ABED techniques that offer implementation complexity, runtime overhead, and coverage trade-offs. Results show that ABED can detect all transient hardware errors that might otherwise corrupt output and does so while incurring low runtime overheads (6-23%), offering at least 1.6X throughput to workloads compared to full duplication. △ Less

Submitted 8 June, 2020; originally announced June 2020.

Estimating Silent Data Corruption Rates Using a Two-Level Model

Authors: Siva Kumar Sastry Hari, Paolo Rech, Timothy Tsai, Mark Stephenson, Arslan Zulfiqar, Michael Sullivan, Philip Shirvani, Paul Racunas, Joel Emer, Stephen W. Keckler

Abstract: High-performance and safety-critical system architects must accurately evaluate the application-level silent data corruption (SDC) rates of processors to soft errors. Such an evaluation requires error propagation all the way from particle strikes on low-level state up to the program output. Existing approaches that rely on low-level simulations with fault injection cannot evaluate full application… ▽ More High-performance and safety-critical system architects must accurately evaluate the application-level silent data corruption (SDC) rates of processors to soft errors. Such an evaluation requires error propagation all the way from particle strikes on low-level state up to the program output. Existing approaches that rely on low-level simulations with fault injection cannot evaluate full applications because of their slow speeds, while application-level accelerated fault testing in accelerated particle beams is often impractical. We present a new two-level methodology for application resilience evaluation that overcomes these challenges. The proposed approach decomposes application failure rate estimation into (1) identifying how particle strikes in low-level unprotected state manifest at the architecture-level, and (2) measuring how such architecture-level manifestations propagate to the program output. We demonstrate the effectiveness of this approach on GPU architectures. We also show that using just one of the two steps can overestimate SDC rates and produce different trends---the composition of the two is needed for accurate reliability modeling. △ Less

Submitted 27 April, 2020; originally announced May 2020.

HarDNN: Feature Map Vulnerability Evaluation in CNNs

Authors: Abdulrahman Mahmoud, Siva Kumar Sastry Hari, Christopher W. Fletcher, Sarita V. Adve, Charbel Sakr, Naresh Shanbhag, Pavlo Molchanov, Michael B. Sullivan, Timothy Tsai, Stephen W. Keckler

Abstract: As Convolutional Neural Networks (CNNs) are increasingly being employed in safety-critical applications, it is important that they behave reliably in the face of hardware errors. Transient hardware errors may percolate undesirable state during execution, resulting in software-manifested errors which can adversely affect high-level decision making. This paper presents HarDNN, a software-directed ap… ▽ More As Convolutional Neural Networks (CNNs) are increasingly being employed in safety-critical applications, it is important that they behave reliably in the face of hardware errors. Transient hardware errors may percolate undesirable state during execution, resulting in software-manifested errors which can adversely affect high-level decision making. This paper presents HarDNN, a software-directed approach to identify vulnerable computations during a CNN inference and selectively protect them based on their propensity towards corrupting the inference output in the presence of a hardware error. We show that HarDNN can accurately estimate relative vulnerability of a feature map (fmap) in CNNs using a statistical error injection campaign, and explore heuristics for fast vulnerability assessment. Based on these results, we analyze the tradeoff between error coverage and computational overhead that the system designers can use to employ selective protection. Results show that the improvement in resilience for the added computation is superlinear with HarDNN. For example, HarDNN improves SqueezeNet's resilience by 10x with just 30% additional computations. △ Less

Submitted 25 February, 2020; v1 submitted 22 February, 2020; originally announced February 2020.

Comments: 14 pages, 5 figures, a short version accepted for publication in First Workshop on Secure and Resilient Autonomy (SARA) co-located with MLSys2020

Secure Wireless Internet of Things Communication using Virtual Private Networks

Authors: Ishaan Lodha, Lakshana Kolur, K. Sree Hari, Honnavalli Prasad

Abstract: The Internet of Things (IoT) is an exploding market as well as a important focus area for research. Security is a major issue for IoT products and solutions, with several massive problems that are still commonplace in the field. In this paper, we have successfully minimized the risk of data eavesdropping and tampering over the network by securing these communications using the concept of tunneling… ▽ More The Internet of Things (IoT) is an exploding market as well as a important focus area for research. Security is a major issue for IoT products and solutions, with several massive problems that are still commonplace in the field. In this paper, we have successfully minimized the risk of data eavesdropping and tampering over the network by securing these communications using the concept of tunneling. We have implemented this by connecting a router to the internet via a Virtual Private network while using PPTP and L2TP as the underlying protocols for the VPN and exploring their cost benefits, compatibility and most importantly, their feasibility. The main purpose of our paper is to try to secure IoT networks without adversely affecting the selling point of IoT. △ Less

Submitted 30 November, 2019; originally announced December 2019.

Comments: 8 pages

3D Conditional Generative Adversarial Networks to enable large-scale seismic image enhancement

Authors: Praneet Dutta, Bruce Power, Adam Halpert, Carlos Ezequiel, Aravind Subramanian, Chanchal Chatterjee, Sindhu Hari, Kenton Prindle, Vishal Vaddina, Andrew Leach, Raj Domala, Laura Bandura, Massimo Mascaro

Abstract: We propose GAN-based image enhancement models for frequency enhancement of 2D and 3D seismic images. Seismic imagery is used to understand and characterize the Earth's subsurface for energy exploration. Because these images often suffer from resolution limitations and noise contamination, our proposed method performs large-scale seismic volume frequency enhancement and denoising. The enhanced imag… ▽ More We propose GAN-based image enhancement models for frequency enhancement of 2D and 3D seismic images. Seismic imagery is used to understand and characterize the Earth's subsurface for energy exploration. Because these images often suffer from resolution limitations and noise contamination, our proposed method performs large-scale seismic volume frequency enhancement and denoising. The enhanced images reduce uncertainty and improve decisions about issues, such as optimal well placement, that often rely on low signal-to-noise ratio (SNR) seismic volumes. We explored the impact of adding lithology class information to the models, resulting in improved performance on PSNR and SSIM metrics over a baseline model with no conditional information. △ Less

Submitted 15 November, 2019; originally announced November 2019.

Comments: To be Presented at the NeurIPS 2019, Second Workshop on Machine Learning and the Physicial Sciences, Vancouver, Canada

An Approximation Algorithm for a Task Allocation, Sequencing and Scheduling Problem involving a Human-Robot Team

Authors: Sai Krishna Hari, Abhishek Nayak, Sivakumar Rathinam

Abstract: This article presents an approximation algorithm for a task allocation, sequencing and scheduling problem involving a team of human operators and robots. Specifically, we present an algorithm with an approximation ratio as a function of the number of human operators ($m$) and the number of robots ($k$) in the team. The approximation ratios are $\frac{7}{2} -\frac{5}{2k}$,… ▽ More This article presents an approximation algorithm for a task allocation, sequencing and scheduling problem involving a team of human operators and robots. Specifically, we present an algorithm with an approximation ratio as a function of the number of human operators ($m$) and the number of robots ($k$) in the team. The approximation ratios are $\frac{7}{2} -\frac{5}{2k}$, $\frac{5}{2} -\frac{1}{k}$ and $\frac{7}{2} -\frac{1}{k}$ when $m=1$, $m\geq k\geq 2$ and $k>m\geq 2$ respectively. We also present computational results to corroborate the performance of the proposed approximation algorithm. △ Less

Submitted 11 September, 2019; v1 submitted 2 July, 2019; originally announced July 2019.

ML-based Fault Injection for Autonomous Vehicles: A Case for Bayesian Fault Injection

Authors: Saurabh Jha, Subho S. Banerjee, Timothy Tsai, Siva K. S. Hari, Michael B. Sullivan, Zbigniew T. Kalbarczyk, Stephen W. Keckler, Ravishankar K. Iyer

Abstract: The safety and resilience of fully autonomous vehicles (AVs) are of significant concern, as exemplified by several headline-making accidents. While AV development today involves verification, validation, and testing, end-to-end assessment of AV systems under accidental faults in realistic driving scenarios has been largely unexplored. This paper presents DriveFI, a machine learning-based fault inj… ▽ More The safety and resilience of fully autonomous vehicles (AVs) are of significant concern, as exemplified by several headline-making accidents. While AV development today involves verification, validation, and testing, end-to-end assessment of AV systems under accidental faults in realistic driving scenarios has been largely unexplored. This paper presents DriveFI, a machine learning-based fault injection engine, which can mine situations and faults that maximally impact AV safety, as demonstrated on two industry-grade AV technology stacks (from NVIDIA and Baidu). For example, DriveFI found 561 safety-critical faults in less than 4 hours. In comparison, random injection experiments executed over several weeks could not find any safety-critical faults △ Less

Submitted 1 July, 2019; originally announced July 2019.

Comments: Accepted at 2019 49th Annual IEEE/IFIP International Conference on Dependable Systems and Networks

Kayotee: A Fault Injection-based System to Assess the Safety and Reliability of Autonomous Vehicles to Faults and Errors

Authors: Saurabh Jha, Timothy Tsai, Siva Hari, Michael Sullivan, Zbigniew Kalbarczyk, Stephen W. Keckler, Ravishankar K. Iyer

Abstract: Fully autonomous vehicles (AVs), i.e., AVs with autonomy level 5, are expected to dominate road transportation in the near-future and contribute trillions of dollars to the global economy. The general public, government organizations, and manufacturers all have significant concern regarding resiliency and safety standards of the autonomous driving system (ADS) of AVs . In this work, we proposed an… ▽ More Fully autonomous vehicles (AVs), i.e., AVs with autonomy level 5, are expected to dominate road transportation in the near-future and contribute trillions of dollars to the global economy. The general public, government organizations, and manufacturers all have significant concern regarding resiliency and safety standards of the autonomous driving system (ADS) of AVs . In this work, we proposed and developed (a) `Kayotee' - a fault injection-based tool to systematically inject faults into software and hardware components of the ADS to assess the safety and reliability of AVs to faults and errors, and (b) an ontology model to characterize errors and safety violations impacting reliability and safety of AVs. Kayotee is capable of characterizing fault propagation and resiliency at different levels - (a) hardware, (b) software, (c) vehicle dynamics, and (d) traffic resilience. We used Kayotee to study a proprietary ADS technology built by Nvidia corporation and are currently applying Kayotee to other open-source ADS systems. △ Less

Submitted 1 July, 2019; originally announced July 2019.

Comments: Presented at Automotive Reliability and Testing (ART) 2018 colocated with International Testing Conference

Persistent Monitoring of Dynamically Changing Environments Using an Unmanned Vehicle

Authors: Sai Krishna Kanth Hari, Sivakumar Rathinam, Swaroop Darbha, Krishnamoorthy Kalyanam, Satyanarayana Gupta Manyam, David Casbeer

Abstract: We consider the problem of planning a closed walk $\mathcal W$ for a UAV to persistently monitor a finite number of stationary targets with equal priorities and dynamically changing properties. A UAV must physically visit the targets in order to monitor them and collect information therein. The frequency of monitoring any given target is specified by a target revisit time, $i.e.$, the maximum allo… ▽ More We consider the problem of planning a closed walk $\mathcal W$ for a UAV to persistently monitor a finite number of stationary targets with equal priorities and dynamically changing properties. A UAV must physically visit the targets in order to monitor them and collect information therein. The frequency of monitoring any given target is specified by a target revisit time, $i.e.$, the maximum allowable time between any two successive visits to the target. The problem considered in this paper is the following: Given $n$ targets and $k \geq n$ allowed visits to them, find an optimal closed walk $\mathcal W^*(k)$ so that every target is visited at least once and the maximum revisit time over all the targets, $\mathcal R(\mathcal W(k))$, is minimized. We prove the following: If $k \geq n^2-n$, $\mathcal R(\mathcal W^*(k))$ (or simply, $\mathcal R^*(k)$) takes only two values: $\mathcal R^*(n)$ when $k$ is an integral multiple of $n$, and $\mathcal R^*(n+1)$ otherwise. This result suggests significant computational savings - one only needs to determine $\mathcal W^*(n)$ and $\mathcal W^*(n+1)$ to construct an optimal solution $\mathcal W^*(k)$. We provide MILP formulations for computing $\mathcal W^*(n)$ and $\mathcal W^*(n+1)$. Furthermore, for {\it any} given $k$, we prove that $\mathcal R^*(k) \geq \mathcal R^*(k+n)$. △ Less

Submitted 3 June, 2019; v1 submitted 7 August, 2018; originally announced August 2018.

Worlds of Events: Deduction with Partial Knowledge about Causality

Authors: Seyed Hossein Haeri, Peter Van Roy, Carlos Baquero, Christopher Meiklejohn

Abstract: Interactions between internet users are mediated by their devices and the common support infrastructure in data centres. Keeping track of causality amongst actions that take place in this distributed system is key to provide a seamless interaction where effects follow causes. Tracking causality in large scale interactions is difficult due to the cost of keeping large quantities of metadata; even m… ▽ More Interactions between internet users are mediated by their devices and the common support infrastructure in data centres. Keeping track of causality amongst actions that take place in this distributed system is key to provide a seamless interaction where effects follow causes. Tracking causality in large scale interactions is difficult due to the cost of keeping large quantities of metadata; even more challenging when dealing with resource-limited devices. In this paper, we focus on keeping partial knowledge on causality and address deduction from that knowledge. We provide the first proof-theoretic causality modelling for distributed partial knowledge. We prove computability and consistency results. We also prove that the partial knowledge gives rise to a weaker model than classical causality. We provide rules for offline deduction about causality and refute some related folklore. We define two notions of forward and backward bisimilarity between devices, using which we prove two important results. Namely, no matter the order of addition/removal, two devices deduce similarly about causality so long as: (1) the same causal information is fed to both. (2) they start bisimilar and erase the same causal information. Thanks to our establishment of forward and backward bisimilarity, respectively, proofs of the latter two results work by simple induction on length. △ Less

Submitted 10 August, 2016; originally announced August 2016.

Comments: In Proceedings ICE 2016, arXiv:1608.03131

ACM Class: C.2.4; F.4.1

Journal ref: EPTCS 223, 2016, pp. 113-127

Fusion of Sparse Reconstruction Algorithms for Multiple Measurement Vectors

Authors: Deepa K. G., Sooraj K. Ambat, K. V. S. Hari

Abstract: We consider the recovery of sparse signals that share a common support from multiple measurement vectors. The performance of several algorithms developed for this task depends on parameters like dimension of the sparse signal, dimension of measurement vector, sparsity level, measurement noise. We propose a fusion framework, where several multiple measurement vector reconstruction algorithms partic… ▽ More We consider the recovery of sparse signals that share a common support from multiple measurement vectors. The performance of several algorithms developed for this task depends on parameters like dimension of the sparse signal, dimension of measurement vector, sparsity level, measurement noise. We propose a fusion framework, where several multiple measurement vector reconstruction algorithms participate and the final signal estimate is obtained by combining the signal estimates of the participating algorithms. We present the conditions for achieving a better reconstruction performance than the participating algorithms. Numerical simulations demonstrate that the proposed fusion algorithm often performs better than the participating algorithms. △ Less

Submitted 6 April, 2015; originally announced April 2015.

A Risk Minimization Framework for Channel Estimation in OFDM Systems

Authors: Karthik Upadhya, Chandra Sekhar Seelamantula, K. V. S. Hari

Abstract: We address the problem of channel estimation for cyclic-prefix (CP) Orthogonal Frequency Division Multiplexing (OFDM) systems. We model the channel as a vector of unknown deterministic constants and hence, do not require prior knowledge of the channel statistics. Since the mean-square error (MSE) is not computable in practice, in such a scenario, we propose a novel technique using Stein's lemma to… ▽ More We address the problem of channel estimation for cyclic-prefix (CP) Orthogonal Frequency Division Multiplexing (OFDM) systems. We model the channel as a vector of unknown deterministic constants and hence, do not require prior knowledge of the channel statistics. Since the mean-square error (MSE) is not computable in practice, in such a scenario, we propose a novel technique using Stein's lemma to obtain an unbiased estimate of the mean-square error, namely the Stein's unbiased risk estimate (SURE). We obtain an estimate of the channel from noisy observations using linear and nonlinear denoising functions, whose parameters are chosen to minimize SURE. Based on computer simulations, we show that using SURE-based channel estimate in equalization offers an improvement in signal-to-noise ratio of around 2.25 dB over the maximum-likelihood channel estimate, in practical channel scenarios, without assuming prior knowledge of channel statistics. △ Less

Submitted 22 October, 2014; originally announced October 2014.

A Fast Eigen Solution for Homogeneous Quadratic Minimization with at most Three Constraints

Authors: Dinesh Dileep Gaurav, K. V. S. Hari

Abstract: We propose an eigenvalue based technique to solve the Homogeneous Quadratic Constrained Quadratic Programming problem (HQCQP) with at most 3 constraints which arise in many signal processing problems. Semi-Definite Relaxation (SDR) is the only known approach and is computationally intensive. We study the performance of the proposed fast eigen approach through simulations in the context of MIMO rel… ▽ More We propose an eigenvalue based technique to solve the Homogeneous Quadratic Constrained Quadratic Programming problem (HQCQP) with at most 3 constraints which arise in many signal processing problems. Semi-Definite Relaxation (SDR) is the only known approach and is computationally intensive. We study the performance of the proposed fast eigen approach through simulations in the context of MIMO relays and show that the solution converges to the solution obtained using the SDR approach with significant reduction in complexity. △ Less

Submitted 1 August, 2013; originally announced August 2013.

Comments: 15 pages, The same content without appendices is accepted and is to be published in IEEE Signal Processing Letters

Low Complexity Joint Estimation of Synchronization Impairments in Sparse Channel for MIMO-OFDM System

Authors: Renu Jose, Sooraj K. Ambat, K. V. S. Hari

Abstract: Low complexity joint estimation of synchronization impairments and channel in a single-user MIMO-OFDM system is presented in this letter. Based on a system model that takes into account the effects of synchronization impairments such as carrier frequency offset, sampling frequency offset, and symbol timing error, and channel, a Maximum Likelihood (ML) algorithm for the joint estimation is proposed… ▽ More Low complexity joint estimation of synchronization impairments and channel in a single-user MIMO-OFDM system is presented in this letter. Based on a system model that takes into account the effects of synchronization impairments such as carrier frequency offset, sampling frequency offset, and symbol timing error, and channel, a Maximum Likelihood (ML) algorithm for the joint estimation is proposed. To reduce the complexity of ML grid search, the number of received signal samples used for estimation need to be reduced. The conventional channel estimation methods using Least-Squares (LS) fail for the reduced sample under-determined system, which results in poor performance of the joint estimator. The proposed ML algorithm uses Compressed Sensing (CS) based channel estimation method in a sparse fading scenario, where the received samples used for estimation are less than that required for an LS based estimation. The performance of the estimation method is studied through numerical simulations, and it is observed that CS based joint estimator performs better than LS based joint estimator △ Less

Submitted 28 April, 2013; originally announced April 2013.

Comments: 7 pages, 4 figures, under review in AEU - International Journal of Electronics and Communications (Elsevier) (paper id-AEUE-D-12-00625)

Spatial Modulation in Zero-Padded Single Carrier Communication

Authors: Rakshith Rajashekar, K. V. S. Hari

Abstract: In this paper, we consider the Spatial Modulation (SM) system in a frequency selective channel under single carrier (SC) communication scenario and propose zero-padding instead of cyclic prefix considered in the existing literature. We show that the zero-padded single carrier (ZP-SC) SM system offers full multipath diversity under maximum-likelihood (ML) detection, unlike the cyclic prefixed SM sy… ▽ More In this paper, we consider the Spatial Modulation (SM) system in a frequency selective channel under single carrier (SC) communication scenario and propose zero-padding instead of cyclic prefix considered in the existing literature. We show that the zero-padded single carrier (ZP-SC) SM system offers full multipath diversity under maximum-likelihood (ML) detection, unlike the cyclic prefixed SM system. Further, we show that the order of ML decoding complexity in the proposed ZP-SC SM system is independent of the frame length and depends only on the number of multipath links between the transmitter and the receiver. Thus, we show that the zero-padding in the SC SM system has two fold advantage over cyclic prefixing: 1) gives full multipath diversity, and 2) offers relatively low ML decoding complexity. Furthermore, we extend the partial interference cancellation receiver (PIC-R) proposed by Guo and Xia for the decoding of STBCs in order to convert the ZP-SC system into a set of flat-fading subsystems. We show that the transmission of any full rank STBC over these subsystems achieves full transmit, receive as well as multipath diversity under PIC-R. With the aid of this extended PIC-R, we show that the ZP-SC SM system achieves receive and multipath diversity with a decoding complexity same as that of the SM system in flat-fading scenario. △ Less

Submitted 16 January, 2013; v1 submitted 6 December, 2012; originally announced December 2012.

Comments: This paper has been withdrawn by the authors

Maximum Likelihood Algorithms for Joint Estimation of Synchronization Impairments and Channel in MIMO-OFDM System

Authors: Renu Jose, K. V. S. Hari

Abstract: Maximum Likelihood (ML) algorithms, for the joint estimation of synchronization impairments and channel in Multiple Input Multiple Output-Orthogonal Frequency Division Multiplexing (MIMO-OFDM) system, are investigated in this work. A system model that takes into account the effects of carrier frequency offset, sampling frequency offset, symbol timing error, and channel impulse response is formulat… ▽ More Maximum Likelihood (ML) algorithms, for the joint estimation of synchronization impairments and channel in Multiple Input Multiple Output-Orthogonal Frequency Division Multiplexing (MIMO-OFDM) system, are investigated in this work. A system model that takes into account the effects of carrier frequency offset, sampling frequency offset, symbol timing error, and channel impulse response is formulated. Cramér-Rao Lower Bounds for the estimation of continuous parameters are derived, which show the coupling effect among different impairments and the significance of the joint estimation. We propose an ML algorithm for the estimation of synchronization impairments and channel together, using grid search method. To reduce the complexity of the joint grid search in ML algorithm, a Modified ML (MML) algorithm with multiple one-dimensional searches is also proposed. Further, a Stage-wise ML (SML) algorithm using existing algorithms, which estimate fewer number of parameters, is also proposed. Performance of the estimation algorithms is studied through numerical simulations and it is found that the proposed ML and MML algorithms exhibit better performance than SML algorithm. △ Less

Submitted 27 October, 2012; v1 submitted 19 October, 2012; originally announced October 2012.

Comments: 18 pages, 5 figures, Submitted to IET Communications

Structured Dispersion Matrices from Space-Time Block Codes for Space-Time Shift Keying

Authors: Rakshith Rajashekar, K. V. S. Hari, L. Hanzo

Abstract: Coherent Space-Time Shift Keying (CSTSK) is a recently developed generalized shift-keying framework for Multiple-Input Multiple-Output systems, which uses a set of Space-Time matrices termed as Dispersion Matrices (DM). CSTSK may be combined with a classic signaling set (eg. QAM, PSK) in order to strike a flexible tradeoff between the achievable diversity and multiplexing gain. One of the key bene… ▽ More Coherent Space-Time Shift Keying (CSTSK) is a recently developed generalized shift-keying framework for Multiple-Input Multiple-Output systems, which uses a set of Space-Time matrices termed as Dispersion Matrices (DM). CSTSK may be combined with a classic signaling set (eg. QAM, PSK) in order to strike a flexible tradeoff between the achievable diversity and multiplexing gain. One of the key benefits of the CSTSK scheme is its Inter-Channel Interference (ICI) free system that makes single-stream Maximum Likelihood detection possible at low-complexity. In the existing CSTSK scheme, DMs are chosen by maximizing the mutual information over a large set of complex valued, Gaussian random matrices through numerical simulations. We refer to them as Capacity-Optimized (CO) DMs. In this contribution we establish a connection between the STSK scheme as well as the Space-Time Block Codes (STBC) and show that a class of STBCs termed as Decomposable Dispersion Codes (DDC) enjoy all the benefits that are specific to the STSK scheme. Two STBCs belonging to this class are proposed, a rate-one code from Field Extensions and a full-rate code from Cyclic Division Algebras, that offer structured DMs with desirable properties such as full-diversity, and a high coding gain. We show that the DMs derived from these codes are capable of achieving a performance than CO-DMs, and emphasize the importance of DMs having a higher coding gain than CO-DMs in scenarios having realistic, imperfect channel state information at the receiver. △ Less

Submitted 9 October, 2012; originally announced October 2012.

Comments: 30 pages 9 figures. Ignore the 31st page which has a copy of Fig. 5

Power Allocation in Amplify and Forward Relays with a Power Constrained Relay

Authors: Dinesh Dileep Gaurav, K. V. S. Hari

Abstract: We consider a two-hop Multiple-Input Multiple-Output channel with a source, a single Amplify and Forward relay, and the destination. We consider the problem of designing precoders at the source and the relay, and the receiver matrix at the destination. In particular, we address the problem of optimal power allocation scheme at the source which minimizes the source transmit power while satisfying a… ▽ More We consider a two-hop Multiple-Input Multiple-Output channel with a source, a single Amplify and Forward relay, and the destination. We consider the problem of designing precoders at the source and the relay, and the receiver matrix at the destination. In particular, we address the problem of optimal power allocation scheme at the source which minimizes the source transmit power while satisfying a given Quality of Service requirement at the destination, and a power constraint at the relay. We consider two types of receiver at the destination, a Zero Forcing receiver and an Minimum Mean Square Error receiver. Simulation Results are provided in the end which compare the performance of both the receivers. △ Less

Submitted 8 November, 2012; v1 submitted 26 September, 2012; originally announced September 2012.

Comments: 9 pages, 2 figures, This is to present the new version with updated content

Low Complexity Maximum Likelihood Detection in Spatial Modulation Systems

Authors: Rakshith Rajashekar, K. V. S. Hari

Abstract: Spatial Modulation (SM) is a recently developed low-complexity Multiple-Input Multiple-Output scheme that uses antenna indices and a conventional signal set to convey information. It has been shown that the Maximum-Likelihood (ML) detection in an SM system involves joint detection of the transmit antenna index and the transmitted symbol, and hence, the ML search complexity grows linearly with the… ▽ More Spatial Modulation (SM) is a recently developed low-complexity Multiple-Input Multiple-Output scheme that uses antenna indices and a conventional signal set to convey information. It has been shown that the Maximum-Likelihood (ML) detection in an SM system involves joint detection of the transmit antenna index and the transmitted symbol, and hence, the ML search complexity grows linearly with the number of transmit antennas and the size of the signal set. In this paper, we show that the ML search complexity in an SM system becomes independent of the constellation size when the signal set employed is a square- or a rectangular-QAM. Further, we show that Sphere Decoding (SD) algorithms become essential in SM systems only when the number of transmit antennas is large and not necessarily when the employed signal set is large. We propose a novel {\em hard-limiting} enabled sphere decoding detector whose complexity is lesser than that of the existing detector and a generalized detection scheme for SM systems with {\em arbitrary} number of transmit antennas. We support our claims with simulation results that the proposed detectors are ML-optimal and offer a significantly reduced complexity. △ Less

Submitted 16 January, 2013; v1 submitted 27 June, 2012; originally announced June 2012.

Comments: This paper has been withdrawn by the authors

ML Decoding Complexity Reduction in STBCs Using Time-Orthogonal Pulse Shaping

Authors: Rakshith Rajashekar, K. V. S. Hari

Abstract: Motivated by the recent developments in the Space Shift Keying (SSK) and Spatial Modulation (SM) systems which employ Time-Orthogonal Pulse Shaping (TOPS) filters to achieve transmit diversity gains, we propose TOPS for Space-Time Block Codes (STBC). We show that any STBC whose set of weight matrices partitions into P subsets under the equivalence relation termed as Common Support Relation can be… ▽ More Motivated by the recent developments in the Space Shift Keying (SSK) and Spatial Modulation (SM) systems which employ Time-Orthogonal Pulse Shaping (TOPS) filters to achieve transmit diversity gains, we propose TOPS for Space-Time Block Codes (STBC). We show that any STBC whose set of weight matrices partitions into P subsets under the equivalence relation termed as Common Support Relation can be made P -group decodable by properly employing TOPS waveforms across space and time. Furthermore, by considering some of the well known STBCs in the literature we show that the order of their Maximum Likelihood decoding complexity can be greatly reduced by the application of TOPS. △ Less

Submitted 25 April, 2012; originally announced April 2012.

Comments: 10 pages