README

How to use

Run main.py to control the robot using a game controller
Run joystick_test.py to test the game controller and obtain button index
Run Yuna.py to watch yuna controlled by a series of commands Run Yuna_TrajPlanner.py to visualise thetrajectory of one foot

Yuna configuration

   ^^ Front ^^   
    1 ----- 2       +x  
        |           ^  
    3 ----- 4       |  <-╮+yaw   
        |     +y <--o    
    5 ----- 6       +z     

Functions for yuna maneuvering

• step()

  • parameters:

    • step_len : float
      Stride length (in metre) that robot will move every step, but the robot will not always move the exact distance, it also depends on the previous commands, here are some exceptions:

      1. When robot is in initial position, the robot will move stride/2 to start walking
      2. When stride==0, and robot is in the process of walking, the robot will step another stride/2 to recover to initial position The value is clipped in [-0.2, 0.2] to avoid collision, but the positive value is encouraged for a more intuitive control

    • course : float
      Moving direction (in degree) the robot will move to

    • rotation : float Angle (in degree) that robot body will rotate every step, positive for counter-clock wise and negetive for clock wise, but the robot will not always rotate the exact angle, it also depends on the previous commands, here are some exceptions:

      1. When robot is in initial position, the robot will rotate rotation/2 to start turning
      2. When rotation==0, and the robot is in the process of turning, the robot will turn another rotation/2 angle to recover to initial position The value is clipped in [-20, 20] to avoid collision

    • step : int
      Steps robot will move, the step for starting is counted. If the robot directly changes the maneuvering mode (e.g. turning -> walking) without executing the stop() funtion, the robot will automatically insert the stop() in between and this step is not counted

• stop()
  By executing this, the robot will do an extra step to recover to the initial standing pose, if robot is already standing still, the robot will do nothing

Understanding of Pose

The pose refers to the tripod position and orientation with respect to the robot body frame. The frame that is rigidly attached to the centre of the tripod can be called a Task Coordinate. For the tripod gait, 3 legs share the same task coordinate, but for other gait patterns, each leg may have different task coordinates.
The overall pose matrix for the robot is a 4 by 6 numpy array, with each column stores the relative pose of each leg's task coordination, and 4 rows stands for relative $x$, $y$, $z$ and $\theta$ with respect to the body frame. At robot initial pose, all the task coordinations of each coincide with the body frame, thus the initial pose is np.zeros((4,6)). In the process of robot motion, the origin of the task coordination does the translational movement and the frame does the rotational movement about its z-axis