Papers
2
Total Citations
10
H-Index
2
About
Mack Tang pushes the boundaries of autonomous robotics and bio-inspired locomotion, with research spanning reinforcement learning for real-world navigation and novel legged robot design. In his highly cited work "SACPlanner" (2023, 6 citations), Tang pioneered the integration of Soft Actor-Critic (SAC) with data augmentation techniques (RAD, DrQ) to achieve near-perfect collision avoidance training in just 10,000 episodes—a dramatic improvement over prior methods. This work demonstrated that RL-based local planners can produce reliable, real-world trajectories, bridging the simulation-to-reality gap for mobile robots. Tang’s equally impactful "StaccaToe" (2024, 4 citations) introduced a human-scale, single-leg robot with an actuated toe and co-actuation architecture inspired by human biomechanics. By mimicking the human leg’s distal control, StaccaToe achieves unprecedented agility for a monopedal system, rivaling human-like locomotion. This design builds on the HyperLeg platform, showcasing Tang’s talent for translating biological principles into practical hardware. His work has already influenced both the RL and robotics communities, offering scalable solutions for autonomous navigation and dynamic legged locomotion.
Research Focus
Key Achievements
Top Papers
- 1
- 2StaccaToe: A Single-Leg Robot that Mimics the Human Leg and Toe4 citations · 2024