Udacity's Self-Driving Car Engineer Nanodegree Program, a project for second term
This project implements a PID (Proportional, Integral and Derivative) controller in C++ to maneuver a vehicle in a simulated environment on a track. The project involves two steps:
-
Manual step to initialize PID parameters
-
Parameter optimization using Gradient Decent (Twiddle) algorithm
Using PID controller, for the input signal the controller tries to estimate the output by minimizing the error compared to the desired value by tweaking the three control terms.
-
Proportional term: This term influences the output by generating a value in proportion to the error, i.e. if the error is small it outputs a smaller value, however if the error is large then it tries generate larger value to compensate for the larger error.
-
Integral term: This term tries to minimize the error by accounting for past errors after the proportional term is applied. As the error in the system is reduced this term tends to reduce as well.
-
Differential term: This term provides an estimate to reduce the error in the system by anticipating the future error based on the rate of error change. It exerts a higher value to growing error change vs a smaller value to the reducing error.
For more technical details visit Wikipedia.
As a first step, I played with each of the PID terms and observed it's impact on the vehicle movement when used indvidually or in combination of them. Here are few trials that I did and capture videos in some cases:
-
P modified, I & D set to zero: Executed as
./pid 1.0 0.0 0.0 0.02
While the vehicle progresses straight forward and as the road curvature comes into play, the controller is trying to generate steering angle to bring the vehicle back to center position. During this process, the error keeps accumulating and the controller trying to catch up with the error resulting into the vehicle swaying a lot as shown
.
-
I modified, P & D set to zero: Executed as
./pid 0.0 1.0 0.0 0.02
In order to handle the systemic bias, when integral controller is applied alone I observed that the steering engaged often and abruptly resulting in an unstable ride as it tries to account various errors from previous runs.
is a video to demonstrate that.
-
D modified, P & I set to zero: Executed as
./pid 0.0 0.0. 1.0 0.02
To compensate for the error in future steps the D parameter is used. As we are not applying any P value to correct the course the input D values makes no difference in vehicle maneuver. Following
shows the behavior.
-
P & D modified, I set to zero: Executed as
./pid 0.5 0.0 2.5 0.02
By setting P and D value the vehicle moves within road lanes with few oscillations eventually going it out of control. The
shows the vehicle going farthest though it can be much more smooth which indicates the parameters needs further tuning.
-
Finally, P, I and D all are modified: Executed as
./pid 0.28 0.00015 2.12 0.02
Using the trial and error method in adjusting the parameters, the vehicle ran successfully within a track as shown in the following
.
Though it was fun to play and observe the vehicle maneuvering erratically, it was much more fun to implement the gradient descent algorithm covered in the twiddle lessons and optimize these parameters.
Twiddle is enabled in main.cpp
by setting corresponding flag to true bool RUN_TWIDDLE = true;
. The optimization is executed as ./pid 0.3 0.0005 2.8 0.01
that initialized the PID parameters for further fine tuning.
Twiddle algorithm is iterated on each of PID parameter for short segment (400 frames) and error is measured to adjust the parameter. Following shows the first few iterations. These iterations are executed until the predefined tolerance value (0.01 specified as the last argument) is achieved. Finally it generated PID parameters as Kp: 4.41076 Ki: 0.0172157 Kd: 48.1967. I scaled down those numbers by the factor of 1/18 to get the numbers in line with the parameters used in manual testing. The final parameters obtained as below:
- Kp: 0.245042222
- Ki: 0.000956428
- Kd: 2.677594444
Following shows the vehicle maneuvering using these parameters.
- cmake >= 3.5
- All OSes: click here for installation instructions
- make >= 4.1(mac, linux), 3.81(Windows)
- Linux: make is installed by default on most Linux distros
- Mac: install Xcode command line tools to get make
- Windows: Click here for installation instructions
- gcc/g++ >= 5.4
- Linux: gcc / g++ is installed by default on most Linux distros
- Mac: same deal as make - [install Xcode command line tools]((https://developer.apple.com/xcode/features/)
- Windows: recommend using MinGW
- uWebSockets
- Run either
./install-mac.sh
or./install-ubuntu.sh
. - If you install from source, checkout to commit
e94b6e1
, i.e.Some function signatures have changed in v0.14.x. See this PR for more details.git clone https://github.com/uWebSockets/uWebSockets cd uWebSockets git checkout e94b6e1
- Run either
- Simulator. You can download these from the project intro page in the classroom.
There's an experimental patch for windows in this PR
- Clone this repo.
- Make a build directory:
mkdir build && cd build
- Compile:
cmake .. && make
- Run it:
./pid
.
Tips for setting up your environment can be found here
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)
Please (do your best to) stick to Google's C++ style guide.
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.
- 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.
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./
A well written README file can enhance your project and portfolio. Develop your abilities to create professional README files by completing this free course.