Papers
16
Total Citations
259
H-Index
6
About
Amrit Singh Bedi is a robotics and machine learning researcher whose work spans online optimization, multi-robot systems, and language-guided autonomous navigation. His most influential contribution, "Online Learning With Inexact Proximal Online Gradient Descent Algorithms" (2019, 109 citations), established efficient low-complexity optimization frameworks for nondifferentiable, time-varying problems critical to robotics and signal processing. This foundational work shaped subsequent research in tracking moving agents through decentralized multi-robot systems, where his inexact gradient descent methods enabled resource-constrained robots to collaborate on complex tasks like search-and-rescue and intrusion detection. Bedi has made significant strides in distributed robotics, developing ASAPP, the first asynchronous algorithm for distributed pose graph optimization in multi-robot SLAM (45 citations), and pioneering reinforcement learning approaches for warehouse automation through multi-robot task allocation systems like RTAW and DC-MRTA. More recently, his research has ventured into trustworthy human-robot interaction, integrating large language models into navigation systems via TrustNavGPT and LANCAR, addressing uncertainty and context-awareness in unstructured environments. His cumulative body of work reflects a coherent trajectory from theoretical optimization to practical, deployable robotic intelligence, making him a notable voice at the intersection of learning theory and autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1Online Learning With Inexact Proximal Online Gradient Descent Algorithms109 citations · 2019
- 2Asynchronous and Parallel Distributed Pose Graph Optimization45 citations · 2020
- 3Tracking Moving Agents via Inexact Online Gradient Descent Algorithm39 citations · 2018
- 4
- 5
- 6
- 7
- 8
- 9
- 10