Chapter 10: Steering, Throttle, and Brake Control

In this chapter, you will learn about more methods for controlling the steering, throttle, and brake using techniques from the field of control systems. If you recall Chapter 8, Behavioral Cloning, you learned how to steer a car using a neural network and camera images. While this most closely mimics how a human drives a car, it can be resource-intensive due to the computational needs of neural networks.

There are more traditional and less resource-intensive methods for controlling a vehicle. The most widely used of these is the PID (short for Proportional, Integral, Derivative) controller, which you will implement in CARLA to drive your car around the simulated town.

There is also another method that is widely used in self-driving cars, called the MPC (short for Model Predictive Controller). The MPC focuses on simulating trajectories, calculating the cost of each trajectory, and selecting the trajectory with the minimum cost...