/Term2-P5-CarND-MPC-Control

Primary LanguageC++

CarND-Controls-MPC

Self-Driving Car Engineer Project 10- Term2 project 5

Introduction

This repository contains my solution to the Udacity self-driving car MPC Project.

The goal of this project is to navigate a track in a Udacity-provided simulator, which communicates telemetry and track waypoint data via websocket, by sending steering and acceleration commands back to the simulator. The solution must be robust to 100ms latency, as one may encounter in real-world application.

This solution, as the Nanodegree lessons suggest, makes use of the IPOPT and CPPAD libraries to calculate an optimal trajectory and its associated actuation commands in order to minimize error with a third-degree polynomial fit to the given waypoints. The optimization considers only a short duration's worth of waypoints, and produces a trajectory for that duration based upon a model of the vehicle's kinematics and a cost function based mostly on the vehicle's cross-track error (roughly the distance from the track waypoints) and orientation angle error, with other cost factors included to improve performance.

Rubric Points

  • The Model: Student describes their model in detail. This includes the state, actuators and update equations.
  • The model includes three main parts.

The first one is the state which is consisted of (x,y,psi,v,cte,epsi). The second part is the actuators and they are steer value and throttle/brake. The last part is the cost function and constrains. we also use the equation below to predict the state at timestep t+1 according to the state and actuation at timestep t.

  • x_[t+1] = x[t] + v[t] * cos(psi[t]) * dt

  • y_[t+1] = y[t] + v[t] * sin(psi[t]) * dt

  • psi_[t+1] = psi[t] + v[t] / Lf * delta[t] * dt

  • v_[t+1] = v[t] + a[t] * dt

  • cte[t+1] = f(x[t]) - y[t] + v[t] * sin(epsi[t]) * dt

  • epsi[t+1] = psi[t] - psides[t] + v[t] * delta[t] / Lf * dt

  • Timestep Length and Elapsed Duration (N & dt): Student discusses the reasoning behind the chosen N (timestep length) and dt (elapsed duration between timesteps) values. Additionally the student details the previous values tried.

I choose the N and dt to be (10,0.1), as 100.1=1s which delta_t is small enough and Ndt is large enough. Also the larger the N is, the more parameters the IPOPT will be tested to find the best set. So 10 timesteps seems to be reasonable. Of course, I tried other n and dt but not get the best tradeoff between the performance and the speed of solve calculation.

  • Polynomial Fitting and MPC Preprocessing: A polynomial is fitted to waypoints. If the student preprocesses waypoints, the vehicle state, and/or actuators prior to the MPC procedure it is described.

I transformed the way point from the global coordinates to car coordinates in the main function. By doing so, I get the x,y,phi init state with the value of zero, which makes things eaiser.

  • Model Predictive Control with Latency: The student implements Model Predictive Control that handles a 100 millisecond latency. Student provides details on how they deal with latency.

As the latency is a constant, I predict the vehicle state in 100ms first before puting it into the update equation.

//take 100ms simulator's latency into account. Predict state after latency and pass that to the solver

double dt = 0.1;

double x_new = x + v*cos(psi)*dt;

double y_new = y + v*sin(psi)*dt;

double psi_new = psi + v/Lf * last_delta * dt;

double v_new = v + last_a * dt;

double cte_new = cte + v * sin(epsi) * dt;

double epsi_new = epsi + v/Lf * last_delta * dt;

PS

The most gorgous thing I want to mention is that I tune the hyperparameters of the weights of cost events.

It makes sense that the velocity error and cte should not be the same important. I give every cost event a fine-tuned weight.

// Both the reference cross track and orientation errors are 0.

// The reference velocity is set to 70 mph.

double ref_cte = 0;

double ref_epsi = 0;

double ref_v = 70;//40;

//Multipliers(AKA weights) for the cost function. Tuned based on performance in the simulator

//The certain weight is larger, then we put more penalty on it.

int cost_cte = 3000;//1500;

int cost_eps = 3000; //500

int cost_v = 10; // 1 stop on the track, 100 almost go outside the track around the corner.

int cost_current_delta = 50; //50

int cost_current_a = 25; //25

int cost_v_product_delta =1000; //the smaller the velocity is, the larger the delta is. 700

int cost_diff_delta = 200; //300

int cost_diff_a = 10; //125

Dependencies

Basic Build Instructions

  1. Clone this repo.
  2. Make a build directory: mkdir build && cd build
  3. Compile: cmake .. && make
  4. Run it: ./mpc.

Tips

  1. It's recommended to test the MPC on basic examples to see if your implementation behaves as desired. One possible example is the vehicle starting offset of a straight line (reference). If the MPC implementation is correct, after some number of timesteps (not too many) it should find and track the reference line.
  2. The lake_track_waypoints.csv file has the waypoints of the lake track. You could use this to fit polynomials and points and see of how well your model tracks curve. NOTE: This file might be not completely in sync with the simulator so your solution should NOT depend on it.
  3. For visualization this C++ matplotlib wrapper could be helpful.)
  4. Tips for setting up your environment are available here
  5. VM Latency: Some students have reported differences in behavior using VM's ostensibly a result of latency. Please let us know if issues arise as a result of a VM environment.

Editor Settings

We've purposefully kept editor configuration files out of this repo in order to keep it as simple and environment agnostic as possible. However, we recommend using the following settings:

  • indent using spaces
  • set tab width to 2 spaces (keeps the matrices in source code aligned)

Code Style

Please (do your best to) stick to Google's C++ style guide.

Project Instructions and Rubric

Note: regardless of the changes you make, your project must be buildable using cmake and make!

More information is only accessible by people who are already enrolled in Term 2 of CarND. If you are enrolled, see the project page for instructions and the project rubric.

Hints!

  • You don't have to follow this directory structure, but if you do, your work will span all of the .cpp files here. Keep an eye out for TODOs.

Call for IDE Profiles Pull Requests

Help your fellow students!

We decided to create Makefiles with cmake to keep this project as platform agnostic as possible. Similarly, we omitted IDE profiles in order to we ensure that students don't feel pressured to use one IDE or another.

However! I'd love to help people get up and running with their IDEs of choice. If you've created a profile for an IDE that you think other students would appreciate, we'd love to have you add the requisite profile files and instructions to ide_profiles/. For example if you wanted to add a VS Code profile, you'd add:

  • /ide_profiles/vscode/.vscode
  • /ide_profiles/vscode/README.md

The README should explain what the profile does, how to take advantage of it, and how to install it.

Frankly, I've never been involved in a project with multiple IDE profiles before. I believe the best way to handle this would be to keep them out of the repo root to avoid clutter. My expectation is that most profiles will include instructions to copy files to a new location to get picked up by the IDE, but that's just a guess.

One last note here: regardless of the IDE used, every submitted project must still be compilable with cmake and make./

How to write a README

A well written README file can enhance your project and portfolio. Develop your abilities to create professional README files by completing this free course.