Deval Shah

Papers

Total Citations

H-Index

About

Deval Shah is a rising researcher at the intersection of autonomous robotics and energy-efficient hardware design. Her work focuses on solving the critical bottleneck of motion planning—the computationally intensive process that enables robots to navigate dynamic environments safely. Shah’s most cited paper, “Energy-Efficient Realtime Motion Planning” (2023, 13 citations), tackles the challenge that over 90% of motion planning computation is consumed by collision detection. She proposes novel approaches to dramatically reduce this overhead, enabling real-time performance with minimal energy draw. Building on this, her 2024 work on “Collision Prediction for Robotics Accelerators” (3 citations) explores how neural networks can learn from human experts to predict collisions, further streamlining autonomous navigation. Shah also addresses the reliability of these systems in safety-critical applications. In “Characterizing and Improving Resilience of Accelerators to Memory Errors in Autonomous Robots” (2023, 3 citations), she identifies the vulnerability of motion planning accelerators to soft errors and develops targeted mitigation strategies, avoiding the inefficiency of blanket solutions. Her research is vital for deploying trustworthy, high-performance robots in real-world settings, from autonomous vehicles to industrial automation.