Learning Maximal Safe Sets Using Hypernetworks for MPC-Based Local Trajectory Planning in Unknown Environments
Bojan Derajic, Mohamed-Khalil Bouzidi, Sebastian Bernhard, Wolfgang Hoenig
AI summary
Problem
Computing maximal safe sets for obstacle avoidance in unknown environments is computationally intractable in real-time, while existing learning-based methods lack generalization and speed for complex nonlinear dynamics.
Approach
A hypernetwork maps local environment observations to parameters for a main neural network that approximates the maximal safe set, which is used as a terminal constraint in an MPC planner.
Key results
- Up to 52% increase in success rate over baselines
- Real-time inference speed of 5–13 ms per planning step
- Successful physical robot deployment in scenarios where baselines fail
- Novel radially-weighted loss improves safe-set boundary approximation
Why it matters
Enables robust, real-time safe navigation for autonomous robots in unstructured environments without requiring prior global maps or simplifying dynamic assumptions.
Abstract
This paper presents a novel learning-based approach for online estimation of maximal safe sets for local trajectory planning in unknown static environments. The neural represen- tation of a set is used as the terminal set constraint for a model predictive control (MPC) local planner, resulting in improved recursive feasibility and safety. To achieve real-time performance and desired generalization properties, we employ the idea of hypernetworks. We use the Hamilton-Jacobi (HJ) reachability analysis as the source of supervision during the training process, allowing us to consider general nonlinear dynamics and arbitrary constraints. The proposed method is extensively evaluated against relevant baselines in simulations for different environments and robot dynamics. The results show an increase in success rate of up to 52% compared to the best baseline while maintaining com- parable execution speed. Additionally, we deploy our proposed method, NTC-MPC, on a physical robot and demonstrate its ability to safely avoid obstacles in scenarios where the baselines fail.