Self-Driving Car Engineer Nanodegree Program
Clojure version of Udacity's MPC project from term 2 of the self-driving car engineer nanodegree. This repository is intended to serve as starter code for other students who wish to complete the project in Clojure.
The most common choices for self-driving car development are C++ and Python. Clojure supports a faster development style than either of these languages (especially C++). Compared to C++, Clojure has a much simpler and more flexible syntax, clear error handling, and sophisticated dependency management. Compared to Python, Clojure is much faster (close to C++) and has excellent concurrency support.
Here's a tutorial to help you get started.
You will neeed to install Leiningen, a Clojure build tool. This is a fairly easy installation process. Just follow the instructions on the Leiningen website.
I also recommend VS Code with the Calva extension as your first Clojure text editor because it is very easy to install and use. Later, you can explore more advanced options like Cursive (IntelliJ), CIDER (Emacs), or Vim.
You'll find many TODO comments in src/mpc/core.clj indicating parts of the code that you will need to complete. The code already runs as-is, but the car will drive poorly until you make improvements.
You can run the code with the following command. You should also run Udacity's term 2 simulator at the same time and select the "MPC Control" project.
$ lein run
The idea of model predictive control is to describe the problem and desired outcome, then let an optimization library find a good solution to that problem. In this case, we use the figurer library to perform the optimization. You will need to inform figurer about what kind of outcome you want (value function), the available actions and their likelihood of each action (policy function), and the mechanics of this problem (prediction function). Start with simple versions of each of these functions, then enhance these functions to improve your car's steering. For added challenge, see how fast you can go without leaving the track. (Over 100 MPH is possible!)
You should complete this project on your own first. After you've completed the project, you can review my solution to see how it compares.
Copyright © 2018 Eric Lavigne
Distributed under the Eclipse Public License either version 1.0 or (at your option) any later version.