Graph neural network speeds vehicle obstacle avoidance decisions by half
Imitation Learning of Avoidance Modes for Efficient Multimodal Model Predictive Control
Robotics
Summary
Avoiding obstacles while driving requires deciding which side to pass each one, and traditional methods check every possible option, which takes a lot of time. The authors created SideNet-MPC, a smart system that learns from the traditional method to quickly predict the best side to avoid obstacles. It uses a graph that considers how obstacles block each other, reducing work from many checks to just two per decision. Their approach runs almost twice as fast as the original and keeps most of its safety and success.
What this means in practice
- •For autonomous vehicle engineers: Cut obstacle avoidance computation time nearly in half while maintaining safe decision-making in real-time driving scenarios.
- •For robotics control teams: Implement fast side-choice predictions for navigating multiple obstacles without the exponential cost of enumerating all avoidance options.
Authors
Changyu Lee, Byung-Kwan Lee, Kanghoon Yoon
Abstract
Multimodal model predictive control (MM-MPC) avoids local minima among moving obstacles by solving one optimal control problem (OCP) per combination of avoidance sides, so its computational cost grows exponentially with the number of decision-critical obstacles. We propose SideNet-MPC, which replaces this enumeration with a mode decision learned by imitating MM-MPC as an expert. A graph neural network is trained to predict the expert's minimum-cost mode, one avoidance side per obstacle. Its edges encode, for each obstacle pair, the gap predicted for the moment the vehicle first reaches either obstacle, so a neighbor that blocks a preferred passage can reverse the proposed side. The expert's selection rule compares its minimum-cost mode with the previously executed mode; SideNet-MPC keeps this rule and substitutes the network proposal for the minimum-cost mode. Hence at most two OCPs are solved per control period, and every applied input still comes from an OCP with control barrier function constraints rather than from the network. In $100$ seed-matched closed-loop Monte Carlo runs of mixed traffic, SideNet-MPC reaches $93\%$ of the expert's success rate and needs \SI{18.6}{ms} per control period on a single thread, 48\% less than the expert's \SI{35.5}{ms} on sixteen parallel threads. In a controlled gap sweep, its proposed side reverses in the same gap interval as the expert's, whereas proposers without pairwise edges, trained on the same data, never reverse and all deadlock at the blocked gates of a deterministic scene.