Safe trajectory planning for multiple AMRs in narrow passages using Iterative LQ game with CBF
Year: 2026
Type: article
Abstract: This paper proposes a trajectory planning framework using an Iterative LQ Game (ILQ game) with Control Barrier Functions (CBF) for collision avoidance and deadlock suppression of multiple Autonomous Mobile Robots (AMRs) in narrow passages. We introduce an asymmetric weight decision rule based on the Estimated Time of Arrival, leveraging the property that weight disparity among agents effectively mitigates deadlocks. Furthermore, to reduce the quadratic approximation error associated with the input cross-terms in standard discrete-time CBF constraints, we derive a sufficient condition that eliminates these terms while rigorously ensuring forward invariance. The analysis shows that the conventional constraint introduces a cross-derivative term that is not retained in the standard ILQ approximation and whose magnitude can increase with the constraint violation and the Augmented-Lagrangian parameters. The proposed constraint eliminates this specific term. Simulation results verify collision-free trajectory generation and show fewer deadlock and divergence occurrences than the compared methods.
Authors: Keigo Okano, Toru Namerikawa
Institution: Keio University
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,235
OpenAlex ID: W7224108988