Industry-Academia Collaboration Report: Adaptive Cruise Control (ACC) System Based on CarSim-MATLAB Co-Simulation

Industry-Academia Collaboration Report: Adaptive Cruise Control (ACC) System Based on CarSim-MATLAB Co-Simulation

Adaptive Cruise Control (ACC) is a key functionality in modern intelligent driving technology, designed to maintain a safe following distance from preceding vehicles through automatic speed adjustment. This industry-academia collaboration project implements ACC control under nonlinear vehicle dynamics using a CarSim-MATLAB co-simulation platform, employing input-state linearization algorithms to optimize control performance.

System Framework Co-simulation Architecture: CarSim provides high-fidelity vehicle dynamics models simulating real-world physical responses, while MATLAB/Simulink designs control algorithms for sensor data processing, decision logic, and longitudinal control implementation. Nonlinear Model Handling: Addressing nonlinear vehicle dynamics (e.g., aerodynamic drag, tire friction) through input-state linearization techniques that transform the system into linear form, simplifying controller design while preserving original model accuracy. Control Strategy: The ACC system processes preceding vehicle data from radar/camera sensors, combined with ego-vehicle states (velocity, acceleration), utilizing PID or Model Predictive Control (MPC) to generate throttle/brake commands for safe car-following operations.

Technical Highlights Gaussian-Laplacian Pyramid Implementation: In visual perception modules, Laplacian pyramid decomposition extracts multi-scale image features to enhance environmental perception capabilities (e.g., obstacle detection), though detailed enhancement is not yet incorporated in this project. Algorithm Robustness: Input-state linearization effectively addresses nonlinear control challenges, maintaining ACC stability under complex scenarios (e.g., grade variations, emergency braking) through mathematical transformation of system dynamics.

Industry-Academia Significance The CarSim-MATLAB co-simulation approach validates algorithm reliability in virtual environments while accelerating technology transfer from laboratory to industrial application, providing theoretical foundation and engineering experience for practical autonomous driving system deployment.