Context-Triggered Contingency Games for Strategic Multi-Agent Interaction
Kilian Schweppe, Anne-Kathrin Schmuck
AI summary
Problem
Autonomous agents struggle to balance long-term strategic goals with short-term dynamic adaptation in uncertain environments, often causing stalls or safety violations. Existing methods either ignore game-theoretic coupling across control layers or lack real-time computational efficiency.
Approach
The framework links high-level strategic games derived from temporal logic specifications with low-level dynamic contingency games, solved in real-time using a novel factor-graph-based approximate solver.
Key results
- Integration of strategy templates and contingency games into a context-triggered architecture
- Development of a novel factor-graph-based solver enabling real-time contingency game MPC
- Guaranteed satisfaction of high-level LTL objectives while optimizing low-level interactions online
- Experimental validation demonstrating efficient, reliable interaction in autonomous driving and robotic navigation
Why it matters
Enables safe, goal-directed, and computationally efficient multi-agent control for autonomous vehicles and robots operating in complex, uncertain environments.
Abstract
We address the challenge of reliable and efficient interaction in autonomous multi-agent systems, where agents must balance long-term strategic objectives with short-term dynamic adaptation. We propose context-triggered contingency games, a novel integration of strategic games derived from temporal logic specifications with dynamic contingency games solved in real-time. Our two-layered architecture leverages strategy templates to guarantee satisfaction of high-level objec- tives, while a new factor-graph–based solver enables scalable, real-time model predictive control of dynamic interactions. The resulting framework ensures both safety and progress in uncertain, interactive environments. We validate our approach through simulations and hardware experiments in autonomous driving and robotic navigation, demonstrating efficient, reliable, and adaptive multi-agent interaction.