Autonomous vehicles, or self-driving cars, utilize advanced sensors, control systems, and artificial intelligence to independently navigate and operate, offering substantial improvements in road safety, traffic management, and mobility. This paper details the development and simulation of control algorithms for key functionalities in a Level 3 autonomy model of autonomous vehicles using MATLAB Simulink and IPG CarMaker. The focus spans four critical areas: Autonomous Emergency Braking (AEB), Adaptive Cruise Control (ACC), Lane Keep Assist (LKA), and object detection, integrating low-level PID controllers for precise steering, throttle, and brake actuation.
AEB employs advanced algorithms, including fine-tuned PID controllers at both high and low levels and kinematic algorithms, to achieve precise braking and stopping distances under varying conditions. ACC utilizes radar data processed through sophisticated PID controllers to maintain safe following distances while minimizing passenger discomfort. LKA utilizes Hough's algorithm for robust road edge detection, calculating precise steering angles to ensure accurate lane centering across diverse road geometries.
Additionally, the model integrates a trained neural network to identify and respond to traffic signals, signage, vehicles, and other objects with high accuracy. Robust CAN data packet transmission and reception algorithms are implemented in both high and low level controllers, facilitating reliable communication within the system. The CAN protocol ensures efficient data exchange and supports real-time decision-making.
Beyond simulation, these algorithms are integrated into a physical model with Level 3 autonomy capabilities. The onboard computer, Jetson Nano, aggregates data from a camera, radar, and IMU, supporting rigorous real-world testing and validation scenarios.
This integrated approach signifies significant strides in autonomous vehicle control, bolstering safety, comfort, and reliability in dynamic driving environments.