Reference-Free, Long-Horizon Trajectory Optimization for Aggressive Autonomous Driving in Milliseconds
Prayag Sharma, Jon Goh, Franck Djeumou
AI summary
Problem
Existing autonomous driving planners rely on offline references or simplified dynamics, failing to solve full nonlinear, reference-free optimal control problems in real time for safety-critical, high-speed maneuvers.
Approach
The authors systematically optimize numerical integrators, interior-point solver configurations, and initial guess generation to solve long-horizon optimal control problems with full vehicle dynamics in milliseconds.
Key results
- Lower-order A-stable implicit integrators outperform higher-order methods by up to two orders of magnitude
- Robust real-time performance requires specific barrier updates and Hessian handling in interior-point solvers
- A cheap dynamic-equilibrium initial guess reduces feasibility error by up to four orders of magnitude
- The framework generates 90 km/h collision avoidance trajectories in under 72 ms, validating drifting as necessary
Why it matters
Enables autonomous vehicles to safely exploit their full dynamic limits in unpredictable, high-speed scenarios without relying on pre-computed paths or simplified models.
Abstract
Autonomous vehicles must generate long-horizon and dynamically feasible trajectories in real time—even when operating at the limits of vehicle handling—to ensure safe oper- ation in adverse conditions. However, existing work rarely quan- tifies the computational demands of generating such trajectories without prior references, warm starts and often defaults to low- fidelity models, compromising accuracy and control authority. We investigate the modeling and solver design choices that enable real-time solution of long-horizon, reference-free optimal control problems (OCPs) using full vehicle dynamics. To this end, we analyze vehicle stiffness properties to justify the OCP’s integration scheme and show that lower-order A-stable methods consistently outperform alternatives, with solve time differences reaching two orders of magnitude. We show that robust nonlinear solver performance hinges on understanding barrier parameter update strategies and safeguarding techniques for Hessian indefiniteness, inherent in some interior point methods. Lastly, we propose a computationally efficient method for generating initial guesses using dynamic equilibrium, unlocking real-time performance and reducing initial infeasibility by up to four orders of magnitude. Extensive benchmarking and high- fidelity BeamNG simulation demonstrate compute times as low as 55 ms over a 260 m horizon, including high-speed obstacle avoidance scenarios where drifting emerges as a necessary component of feasible trajectory generation.