def environment(X_old, car_id, action_id, t_step, params, dist_id, Level_ratio):
# the selected car moves based on the action
# 1:maintian 2:turn left 3:turn right 4:accelerate 5:decelerate 6:hard # brake
X_new = motion_update.motion_update(X_old, car_id, action_id, t_step, params)
# compute reward
R, params = reward.reward(X_new, car_id, action_id, params, dist_id, Level_ratio)
return X_new, R
def environment_multi(X_old, action_id, t_step, params):
X_new = X_old.copy()
# the selected car moves based on the action
# 1:maintian 2:turn left 3:turn right 4:accelerate 5:decelerate 6:hard brake
X_sub = np.zeros((len(action_id[0]), X_new.shape[0], X_new.shape[1]))
for step in range(0, len(action_id[0])):
size_action_cell = len(action_id)
for car_id in range(0, size_action_cell):
X_new = motion_update.motion_update(X_new, car_id,
action_id[car_id][step],
t_step, params)
X_sub[step, :, :] = X_new.copy()
return X_sub
def Environment_Multi_Sim(X_old1, action_id, t_step, params):
# the selected car moves based on the action
# 1:maintian 2:turn left 3:turn right 4:accelerate 5:decelerate 6:hard brake
X_new = np.zeros((X_old1.shape[0], X_old1.shape[1]))
R = np.zeros((len(action_id), 1))
for step in range(0, 1):
size_action_cell = len(action_id)
for car_id in range(0, size_action_cell):
X_old1 = motion_update.motion_update(X_old1, car_id,
action_id[car_id], t_step,
params)
X_new = X_old1.copy()
for car_id in range(0, size_action_cell):
R[car_id,
step], params = reward_sim.reward_sim(X_new, car_id,
action_id[car_id], params)
return X_new, R