MEAM620 Project Motion planning for AIBO dog in scoring goals Wei Jiao
Motivation oInvolving term working in shooting ball oShooting ball more effectually oSimulation tool for robot soccer match
Approach Three zones in the field initially Ball generation zone, Offensive robot generation zone, Defensive robots generation zone
Approach Offensive robots are indicated by red triangles Defensive robots are indicated by blue triangles (the size of defensive dog is bigger, increase the collision probability, more suitable ball moving trajectory need to be generated in order to score goals) Goal is plot in cyan color.
Approach Coordination of the robot the coordination for each robot: [ x ; y ; rotation ] Coordination for the ball [center_x; center_y ] Rotation angle Center
Approach Collision checking: Compute the collision boundary or C_obj Step: Compute the slope for each edge of the triangle Find the perpendicular slope to the edges’ slope Compute the boundary collision edge vertices the distance between the vertices and the triangle edge is the radius of the ball Three curvature boundary collision edge is plot by a circle, radius is the radius of the ball and center is the vertices of the triangle
Approach Collision checking: Check if the center of the ball outside the blue convex shape Check if the center of the ball outside the green circle f=(x_center-a)^2+(y_center-b)^2-radius^2 (if f 0, outside, f=0, on the ball) The collision is determined by OR operation of the above two results.
Progress examples MATLAB example (shooting directly) No Collision
Progress examples MATLAB example (shooting directly) Collision
Goal The attacking dog moves itself to the best position to shooting the ball avoiding the collision between the ball and the defender. the attacking supporter moves it itself to the best position to score goals
Goal Attacker passes the ball to supporter to shooting balls Hybrid system supervised framework to assign the role in term attacking, scoring goal by term working.