A Hardware-in-the-Loop Simulation Scheme for Automatic Parking Assist

This work describes the design and testing of a hardware-in-the-loop (HiL) simulation system dedicated to automatic parking system (APA). The focus is on creating a laboratory environment that is as close to the real working conditions in order to provide realistic constraints. The proposed system comprises three key components: (1) an environment and vehicle kinematics simulation platform, (2) an integrated cockpit-parking computing platform executing both perception and path-planning algorithms, with the latter running on an isolated safety island, and (3) a microcontroller for vehicle chassis control. The system incorporates actual SoC hardware with peripheral circuits, along with deployed software and algorithm modules, enabling comprehensive system-level verification of the entire hardware-software pipeline. With enhanced observability features, our solution provides valuable references for early-stage APA system design and validation. Experimental results demonstrate that the proposed APA system meets mass-production requirements in both parking accuracy (achieving <5cm positioning error) and end-to-end latency (maintaining <100ms response time).