Apratim Mukherjee
Papers
2
Total Citations
7
H-Index
2
About
Apratim Mukherjee is a researcher advancing the frontier of decentralized multi-agent control, with a focus on applying deep reinforcement learning (DRL) to robot swarms and physically connected aggregates. His work directly tackles the critical challenge of non-stationarity—the instability that arises when multiple robots update their policies concurrently. In his most cited paper, *Decentralized Multi-Agent Reinforcement Learning with Global State Prediction* (2023, 4 citations), he proposes a novel framework that enables agents to predict global state dynamics, mitigating this instability and paving the way for scalable swarm coordination. Complementing this, his study *A Study of Reinforcement Learning Algorithms for Aggregates of Minimalistic Robots* (2022, 3 citations) explores DRL control for groups of physically linked robots that must maintain a prescribed shape. By addressing the unique constraints of minimalistic, connected systems, Mukherjee’s work bridges the gap between single-robot DRL successes and the complex, real-world demands of collective robotics. His contributions are foundational for applications in modular robotics, environmental monitoring, and distributed automation, offering a scalable path toward intelligent, decentralized robotic teams.
Research Focus
Key Achievements
Top Papers
- 1
- 2