Chance-Constrained Convex MPC for Robust Quadruped Locomotion under Parametric and Additive Uncertainties
Ananya Trivedi, Sarvesh Prajapati, Mark Zolotas, Michael Everett, Taskin Padir
AI summary
Problem
Existing quadrupedal locomotion controllers struggle to maintain stability under unknown payloads and complex terrains, often requiring extensive tuning or lacking real-time computational tractability.
Approach
The method models payload and terrain uncertainties as stochastic distributions and reformulates friction cone safety constraints as chance constraints, solving the resulting control problem as a fast, convex quadratic program.
Key results
- 100% simulation success rate across multiple gaits
- Real-time control execution at ~500 Hz
- Successful hardware locomotion with payloads exceeding 50% of body weight
- First real-time stochastic MPC deployment on quadrupedal hardware
Why it matters
It bridges the gap between theoretical stochastic control and practical deployment, enabling safer, more adaptable legged robots for real-world applications without manual tuning.
Abstract
Recent advances in quadrupedal locomotion have fo- cused on improving stability and performance across diverse en- vironments. However, existing methods often lack adequate safety analysis and struggle to adapt to varying payloads and complex terrains, typically requiring extensive tuning. To overcome these challenges, we propose a Chance-Constrained Model Predictive Control (CCMPC) framework that explicitly models payload and terrain variability as distributions of parametric and additive disturbances within the single rigid body dynamics model. Our approach ensures safe and consistent performance under uncer- tain dynamics by expressing the model’s friction cone constraints, which define the feasible set of ground reaction forces, as chance constraints. Moreover, we solve the resulting stochastic control problem using a computationally efficient quadratic programming formulation. Extensive Monte Carlo simulations of quadrupedal locomotion across varying payloads and complex terrains demon- strate that CCMPC significantly outperforms two competitive benchmarks: Linear MPC and MPC with hand-tuned safety mar- gins to maintain stability, reduce foot slippage, and track the center of mass. Hardware experiments on the Unitree Go1 robot show successful locomotion across various indoor and outdoor terrains with unknown loads exceeding 50% of the robot’s body weight, despite no additional parameter tuning.