Design and Testing of a Low-Cost Intelligent Emergency Stop System for Autonomous Ground Vehicles: Trajectory Optimization, Path Tracking, and Embedded Implementation

Abstract

Advancements in autonomous ground vehicle technology have led to widespread developments in the field. Autonomous ground vehicles require robust emergency stop systems capable of safely managing system failures without human intervention. This paper presents an intelligent emergency stop system that combines minimum-jerk trajectory optimization, linear quadratic regulator path tracking control, and friction-aware velocity planning to perform collision avoidance maneuvers while bringing the vehicle to a controlled stop. The system is designed for implementation on a low-cost single board computer operating as an inline gateway between the main autonomy system and the vehicle CAN bus, providing redundancy against primary system failures. Simulation results, including a 1000-run Monte Carlo analysis, demonstrate robustness to vehicle parameter uncertainty while maintaining tire forces within friction limits. Live testing on a Lincoln MKZ sedan at 30 mph confirmed real-time operation on a $15 Raspberry Pi Zero 2 with lateral tracking errors of less than 15 cm.