Control-Informed Constraint Adaptation in Minimum-Time Trajectory Planning for Autonomous Racing

Noosaga Placements

• Robot Motion Planningsubfield95%
  The paper focuses on trajectory planning for autonomous racing, specifically adapting spatial constraints in a motion planner based on execution errors.
  we introduce a control-informed online trajectory planning frameworkdynamically adapt spatial track constraints and iteratively expand the free-space planning areaThe planner remains time-optimal while compensating for accumulated execution errors
• Trajectory Optimizationframework95%
  The paper uses an online time-optimal MPC planner that solves a receding-horizon optimal control problem for trajectory generation.
  The online planner solves a receding-horizon MLTP formulated (similar to [4])we extend an online time-optimal MPC trajectory planner
• Optimal Control and Model Predictive Controlframework90%
  The planner is based on Model Predictive Control, solving an optimal control problem for minimum-time trajectory generation.
  feeding these updated constraints into a Model Predictive Control (MPC) plannerThe baseline trajectory planner solves online an optimization-based MLTP
• Robot Controlsubfield80%
  The framework integrates with a downstream tracking controller and uses feedback from control execution to inform planning, bridging planning and control.
  establishes a direct feedback loop from the control module back to the plannerthe observation model evaluates execution deviations and the adaptation model iteratively updates the spatial constraints of the OCPutilizing an MPC tracking controller downstream of the planner
• Roboticssubfield60%
  The work is situated in the broader field of autonomous vehicle robotics, with a focus on planning and control integration.
  Autonomous racecars operate at the limits of vehicle dynamicsintegrated into a modular autonomous racing software stack
• Self-Supervised Learning from Physical Interactionframework50%
  The framework learns from its own execution errors during runtime, which is conceptually similar to self-supervised learning from physical interaction, though it uses heuristics rather than learned models.
  learns from its own execution errorsBy measuring systematic tracking deviations during runtime, we dynamically adapt spatial track constraints

Abstract

Autonomous racecars operate at the limits of vehicle dynamics, where small control errors translate into safety-critical behavior and lost performance. Trajectory planners assume perfect tracking and remain blind to execution errors. To guarantee safety, trajectory planners therefore restrict themselves to conservative spatial margins, leaving usable track space untapped. To overcome these issues, we introduce a control-informed online trajectory planning framework that learns from its own execution errors. By measuring systematic tracking deviations during runtime, we dynamically adapt spatial track constraints and iteratively expand the free-space planning area. The planner remains time-optimal while compensating for accumulated execution errors. This method was analyzed in a high-fidelity closed-loop simulation environment with autonomous racecars. The results demonstrate that our approach reduces lap time by 1.8\,s without increasing computational burden, maintaining a median runtime of 25 ms. Our finding indicates that feeding control-induced deviations back into the planning layer unlocks performance previously inaccessible to modular architectures and enables autonomous vehicles to exploit track limits systematically.

Paper Context