Matthew Andrews
Papers
1
Total Citations
6
H-Index
1
About
Matthew Andrews is a robotics researcher whose work centers on reinforcement learning (RL) for real-world autonomous navigation, with a particular focus on collision avoidance and local planning. His major contribution lies in bridging the gap between simulated RL training and practical deployment on physical robots. In his highly cited work, "SACPlanner: Real-World Collision Avoidance with a Soft Actor Critic Local Planner and Polar State Representations" (2023, 6 citations), Andrews demonstrated that enhancements to the Soft Actor Critic (SAC) algorithm—specifically RAD and DrQ—enable near-perfect training performance after just 10,000 episodes, producing smooth, collision-free trajectories on real-world platforms. This achievement is notable for its efficiency and robustness, addressing a key challenge in transferring RL policies from simulation to reality. Andrews’s research has significant implications for autonomous systems in dynamic environments, from warehouse robots to self-driving vehicles. His work is recognized for its practical impact, offering a scalable, data-efficient approach to safe navigation that inspires further advances in learning-based robotics.
Research Focus
Key Achievements
Top Papers
- 1