direction = data['targets'][i][direction_position]
# print direction
speed = data['targets'][i + 2][speed_position]
steer_pred1_vec.append(steer_pred1_order[1] * 1.22)
steer_pred2_vec.append(steer_pred2_order[1] * 1.22)
steer_gt_vec.append(steer_gt_order[1] * 1.22)
# print actions[j].steer
screen.plot3camrcnoise(images[1], actions_noise[1].steer, actions[1].steer,
0, [0, 0])
#figure_plot(steer_pred1_vec, steer_pred2_vec, steer_gt_vec, count)
count += 1
# for j in range(sensors['depth']):
# #print j
# screen.plot_camera(depths[j] ,[j,2])
# save_gta_surface(gta_surface)
actions_noise[int(data['targets'][i + 2][26])] = action_3
direction = data['targets'][i + 2][22]
speed = data['targets'][i + 2][10]
time_use = 1.0
car_lenght = 6
actions[0].steer += min(4 * (math.atan((0.26 * car_lenght) / (time_use * speed + 0.05))) / 3.1415, 0.2)
actions[2].steer -= min(4 * (math.atan((0.26 * car_lenght) / (time_use * speed + 0.05))) / 3.1415, 0.2)
print(" Steer Left MIDDLE Right ")
print(actions[0].steer)
print(actions[1].steer)
print(actions[2].steer)
for j in range(1):
screen.plot3camrcnoise(images[j], \
actions[j].steer, -actions[j].steer + actions_noise[j].steer,
actions_noise[j].steer, j) #
time.sleep(0.0)
image_result = Image.fromarray(images[0])
image_result.save('temp/image' + str(i) + '.png')
speed_list.append((actions[0].steer))
speed_list_noise.append((actions_noise[0].steer))
just_noise.append((-actions[0].steer + actions_noise[0].steer))
ts.append((data['targets'][i][20] - 93532.0) / 1000.0)
# reimg = np.array(redata['images_center'][i])
# recontrol_input = np.array(redata['control'][i][1])
# print img
# img = img*255