Alex Boulanger, Alex Dixon, William Jen
The vector field potential is a path-planning algorithm. In other words, given a map of all the obstacles, create a path that will get the robot from point A to point B. This is done by giving the end point B a "pull" and each obstacle a "repelling" force.
- Simulation:
sudo python3 VectorField.py sim. - Actual hardware:
sudo python3 VectorField.py
You may also omit the "sim" argument if you are not on the RPi - the program will detect if the hardware libraries are not available and run the simulation instead.
This vector field algorithm has multiple parameters with which to change the behavior. They are specified below.
- Robot dimensions (
WHEEL_WIDTHandBODY_LENGTH) - Timestep (
DELTA_T) - Workspace dimensions (
WORKSPACE_WIDTHandWORKSPACE_LENGTH) - Vector field resolution (
NODE_GRID_LENGTHandNODE_GRID_WIDTH) - Vector field scaling factors (
ENDPOINT_SCALING_FACTOR,REPULSION_SCALING_FACTOR): These control the magnitude of the obstacle repulsion and the endpoint pull vector. - Minimum vector field magnitude (
MINIMUM_VECTOR_MAGNITUDE): Some vectors may be very small due to the obstacle and end point vectors cancelling to some degree. This value forces those small vectors to have the specified minimum magnitude. - Initial robot pose (
START_X,START_Y,START_THETA). Note thatSTART_THETAis referenced off the +y axis (i.e. 0 is straight up) - End robot position (
END_X,END_Y) - Robot speeds (
PIVOT_ANGULAR_SPEED,LINEAR_SPEED) - End point tolerance (
MINIMUM_DISTANCE_THRESHOLD) - Robot headling alignment tolerance (
CORRECT_HEADING_THRESHOLD): To reduce motion jerkiness, this angle tells the robot when to pivot to correct its heading - Obstacle list (
OBSTACLES)