The point of this assignment is to get you thinking about perception and control.
Using the car model we provide, develop a controller that can drive the car around the track as fast as possible, given the limitations of the DRC (10mph max speed, friction, etc.).
Part 1 involves driving on a circular track with no obstacles on the road. You will notice that in the demo programs the visual target and a nominal steering angle are set for a particular curvature road. The goal in Assignment 2 part 1 is to automate this, so the car figures out how to steer without human input for different curvature tracks. You need to figure out how to generate steer_desired commands to follow the road. You can do this by using the vision and control framework followed in the sense() and control() subroutines in drive4 in ODE or the vision and car ROS plugins in Gazebo, or you can create your own framework.
Part 2: The track will include curved segments and 90 degree turns, and have obstacles (more oil barrels) on the road.
Starting point for Gazebo. This web page includes examples of transforming visual image into a map, which ODE folks should also look at.
Starting point for ODE. Use the drive4 tutorial.
Don't change the size of the windows in ODE or Gazebo showing you what the camera sees. That screws up the camera simulation.
Pages on car kinematics (how to model a car): Best, ok1, ok2, not as useful1, not as useful2.
drive6: An ODE example of a simple vision system that works amazingly well.
drive7: An ODE example of obstacles.