def map_action(self, action): ## convert to the format used by the simulator O_i = self.R_man[0] state = State_Function(self.R_tgt, O_i, action, showRender=False) state_point = state.run() #print action #print state_point return state_point
def run(self):
granluarity = 5 # start with granularity 1
while (granluarity <= self.k):
R_tgt_rect_sg = self.R_tgt_rect_g
impulse_sg = self.impulse_g
solution = self.solve(R_tgt_rect_sg, self.initial_R_id, impulse_sg, granluarity)
# test solution, if works, break, else increase granularity
if solution != -1:
print 'solution is found', solution
action = ((np.random.uniform(solution[0][0], solution[0][1]),np.random.uniform(solution[1][0], solution[1][1])), (1, np.random.uniform(solution[2][0], solution[2][1])))
state = State_Function(R_tgt_rect_sg, self.O_man_rect, action, showRender=True)
state_point = state.run()
print "resulting state: ", state_point
else:
print 'solution is not found at granluarity ', granluarity
granluarity = granluarity + 1
def __init__(self):
# do noting.
self.O_man_rect = Rect(45, 29, 1, 5)
# An initial target region (position, (hwidth, hheight))
#R_tgt_rect = Rect(75, 29, 1, 5)
R_tgt_rect = Rect(25, 2, 1, 5)
self.R_tgt_rect_g = R_tgt_rect # currently just use the entire region
self.impulse_g = ((500,1000), (500, 1000))
self.k = 5
initial_state = State_Function(R_tgt_rect)
self.objects_distance_matrix = initial_state.objects_distance_matrix()
self.initial_R_id = initial_state.initial_R_tgt_id()
self.used_objects = []
#print action #print state_point return state_point #def __sample_pts_box(box): ## sample each surface if __name__=="__main__": impulse_sg = ((500, 1000),(500,1000)) O_i_rect = Rect(55,29,1,5) O_i = (O_i_rect, O_i_rect.left_edge()) #R_tgt_rect = Rect(75,29,1,5) R_tgt_rect = Rect(45,29,1,5) #R_tgt = (R_tgt_rect, R_tgt_rect.left_edge()) R_tgt = R_tgt_rect test = subsolver(impulse_sg, R_tgt, O_i, 2) local_solution = test.run() action = ((np.random.uniform(local_solution[0][0], local_solution[0][1]),np.random.uniform(local_solution[1][0], local_solution[1][1])), (1, np.random.uniform(local_solution[2][0], local_solution[2][1]))) #action = ((0, -628), (-1, 2)) state = State_Function(R_tgt_rect, O_i_rect, action, showRender=True) state_point = state.run() print local_solution
