jo1990 / carnd-mpc-project

CarND Term 2 Model Predictive Control (MPC) Project

CarND-Controls-MPC

Self-Driving Car Engineer Nanodegree Program

---

MPC Controll

MPC -> Model Predictive Control

MPC is an advanced method of process control that is used to control a process while satisfying a set of constraints.

Model Implementation :

Goal :

The goal of Model Predictive Control is to optimize the control inputs: [δ,a]. An optimizer will tune these inputs until a low cost vector of control inputs is found.

State Vector :

x, y : Car's position.

psi : Car's heading direction.

v : Car's velocity cte : Cross-track error.

epsi : Orientation error

Latency Handling :

There was a latency of 100 ms introduced , to handle this , the actual data were shifted 100 ms before passing it to the MPC module "state vector " to help reduce the effect of the latency.

Equations :

MPC attempts to approximate a continuous reference trajectory by means of discrete paths between actuations. Larger values of dt result in less frequent actuations, which makes it harder to accurately approximate a continuous reference trajectory. This is sometimes called "discretization error".

N, dt, and T are hyperparameters you will need to tune for each model predictive controller.

dt is set to small value because it define the time between each step , so it need to predict accurate.

Larger N will cause the simulator to run slower .

At the end duration of trajectory T is defined by the dt and N.

Note : The cost function parameters were tuned by try-and-error .