Advanced Deep Learning with Python

Path planning

Path planning (or driving policy) is the process of calculating the vehicle trajectory and speed. Although we might have an accurate map and exact location of the vehicle, we still need to keep in mind the dynamics of the environment. The car is surrounded by other moving vehicles, pedestrians, traffic lights, and so on. What happens if the vehicle in front stops suddenly? Or if it's moving too slow? Our AV has to make the decision to overtake and then execute the maneuver. This is an area where ML and DL in particular can be especially useful, and we'll discuss two ways to implement these in this chapter. More specifically, we'll discuss using an imitation driving policy in an end-to-end learning system, as well as a driving ...