If you consider all the data streams in the data a self-driving car operates on the fastest time scale. The frequency is slightly lower than that of Sensor Fusion.
Just lower than you have localization and you will learn more about trajectory planning in the next lesson. Next is the forecast you just learned about.
Then at the top of this graph is the behavioral plan with the lowest update rate. The input of behavior planning comes from the prediction module and the localization module.
Both get input from sensor fusion. The output of the behavior module is sent directly to the trajectory planner. This also requires prediction and input
Localized so that it can send trajectories to the motion controller. Everything in this box is the point
In this lesson, you'll learn about the specifics that take place over this relatively long period of time. But intuitively, this long-term span comes from the fact that the behavioral plan must contain a lot
The data makes a decision for quite a long time frame in 10 seconds or more. We'll start with a short introduction where you'll learn more about typing and typing
Output to behavior planners and understand what problems the behavior planning module is expected to solve .
Next, we'll talk about finite state machines as a technique for implementing behavioral plans.
Moving on to a deeper dive, we'll use the cost function to actually make behavioral level decisions.
But first let's spend some time looking at the inputs we'll work with and produce our outputs.