In phase 1, a controller is implemented and the quadrotor can track a pre-defined trajectory. Phase 2 will focus on trajectory generator. A carefully designed path planner can enable the quadrotor to operate aggressively and precisely.

Download the code.

Your tasks include:

Trajectory Generator

A natural way to command the quadrotor is to set waypoints that it needs to pass by. What the path planner needs to do is to generate a trajectory that:

connects all waypoints (including start and goal points).

meets smoothness criterion.

Two sets of waypoints are provided in test_trajectory.m. You need to design two more sets of waypoints (with at least 6 waypoints). You can choose either 5th order polynomial trajectory or minimum snap trajectory [1]. The later one is our recommendation and will win you bonus points.

Tutorial

You can use the naive trajectory generation method in your lecture (only smoothness and connection of waypoints is required), or you can try the optimization-based method (We encourage you generate the trajectory using this method). If you prefer the latter one, you have two ways to implement it:

You can use the method in the slides of Lecture4 to map the original constrained quadratic program (QP) to an unconstrained QP, and then obtain the closed form solution of the unconstrained QP directly.

You can use quadprog function in Matlab to solve the constrained QP. This function is originated in Matlab, and you can type in “help quadprog” in the command window for more detail.

Sturcture of the Simulator

The simulation code is almost the same with proj1phase1 but a trajectory_generator.m as the main entry point.

+ code
  + readonly #supposed to be read only
    - quadModel_readonly.m #parameters of a 500g quadrotor
    - quadEOM_readonly.m #dynamics model of quadrotor.
    - run_trajectory_readonly.m #solve the equation of motion, receive desired trajectory,run your controller, iteratively. visualization is also included.
  + utils #useful functions.
  - controller.m #You have alredy had a good controller in proj1phase1
  - trajectory_generator.m #What you need to work with. Design the trajectory for quadrotor given the path. And calculate desired state given current time.
  - test_trajectory.m #main entry.
  

[1]: D. Mellinger and V. Kumar, “Minimum snap trajectory generation and control for quadrotors,” in Proc. of the IEEE Int. Conf. on Robotics and Automation, Shanghai, China, May 2011.