This software implements a PID controller in C++ to maneuver a vehicle around a track in a simulator. The simulator will provide the cross track error (CTE) and the velocity (mph) in order to compute the appropriate steering angles. The following clip shows the vehicle in the simulator controlled by the PID controller.

This is only for demonstration. The coefficients can be tuned to follow the intended trajectory more accurately resulting in a more comfortable and less shaky ride.

A large number of different controllers exist to move robots and vehicles. The most fundamental is called PID (Proportional-Integral-Differential). It enables complex, changing accelerations and steering angles. The PID controller is an algorithm that calculates a value (for example a steering wheel angle) according to an error called CTE (see next section for an explanation).

PID is a controller with three coefficients. But first, let's have a look on two important terms before explaining the three components.

• distance between the vehicle´s actual trajectory and the intended trajectory

• Systematic bias is a problem that often occurs in robotics. The problem is due to mechanics. Buying for example a vehicle, we think the steerable front wheels are 100% aligned. But in reality the wheels can be aligned a little bit at an angle. The systematic bias manifests itself in a steering drift. For humans this is not a big deal because we just countersteer intuitionally but in robotics this is a fact to be considered.

• sets the steering angle in proportion to CTE (the coefficient tau_p is called "response strength"):

steering angle = -tau_p * cte

• steers the vehicle towards the trajectory
• using only P-Controller leads to overshooting and oscillating which is demonstrated in the following graph:

• steering angle is not just proportional to CTE but also to the time derivative of CTE:

steering angle = -tau_p * cte - tau_d * diff_cte

• the derivative component countersteers the vehicle and helps not to overshoot the trajectoy and not to oscillate
• The following graph demonstrates the behaviour of P- and PD-controller combined:

• this controller finally solves the problem of overshooting and systematic bias by adding one more term. Using a PID-Controller the steering angle is additionally proportional to the integral of CTE. It considers the systematic error and compensates it.

steering angle = -tau_p * cte - tau_d * diff_cte - tau_i * int_cte

Now it seems like PID is doing worse than PD but this is only because the graphs above assumed an ideal robot with no bias.

Below you can see the corresponding P-, PD- and PID-graphs using a robot with systematic bias. Now you can see that with a real robot with bias the PD-controller is actually not doing better than the PID-Controller.

The dashed line is the goal PID-Controller. We achieve this with a so-called twiddle PID controller (green line) with tuned parameters. Now the PID controller outshines the PD controller. Also, with twiddle the PID controller converges faster but we overshoot drastically at first. This overshoot can be reduced by tuning the twiddle parameters.

Manual tuning of PID coefficients for steering values. Please refer to the code lines 40 to 46 in main.cpp to see the different, tested values for the PID coefficients and retrace the approach to good values. In the following you can see what it would look like to drive with only P- or PD-Controller and the final result with a PID controller.

The vehicle drives along the trajectory with oscillations:

The vehicle follows the trajectory with relatively low oscillations:

The vehicle follows the trajectory quite well.

• 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.

    git clone https://github.com/uWebSockets/uWebSockets 
    cd uWebSockets
    git checkout e94b6e1
    

    Some function signatures have changed in v0.14.x. See this PR for more details.
• Simulator. You can download these from the project intro page in the classroom.

Fellow students have put together a guide to Windows set-up for the project here if the environment you have set up for the Sensor Fusion projects does not work for this project. There's also an experimental patch for windows in this PR.

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

Tips for setting up your environment can be found here