fixes and requirements.txt file

carla_gym/__init__.py

@@ -3,9 +3,7 @@
 import os
 import sys
 
-from carla_gym.carla_api import PythonAPI
-from carla_gym.multi_env import env, parallel_env
-from gym.envs.registration import register
+from gymnasium.envs.registration import register
 
 os.environ["PYGAME_HIDE_SUPPORT_PROMPT"] = "hide"
 LOG_DIR = os.path.join(os.getcwd(), "logs")

carla_gym/carla_api/PythonAPI/agents/navigation/local_planner.py

@@ -80,9 +80,9 @@ def __init__(self, vehicle, opt_dict=None):
         self.init_controller(opt_dict)
 
     def __del__(self):
-        if self._vehicle:
+        if self._vehicle and self._vehicle.is_alive:
             self._vehicle.destroy()
-        print("Destroying ego-vehicle!")
+            print("Destroying ego-vehicle!")
 
     def reset_vehicle(self):
         self._vehicle = None

carla_gym/core/controllers/pedestrian_manager.py

@@ -24,7 +24,7 @@ def __init__(self, pedestrian_config, pedestrian_object, planner, destination):
         self._destination = destination
 
         # Spawn collision and lane sensors if necessary
-        self._collision_sensor = CollisionSensor(self._pedestrian, 0) if pedestrian_config.collision_sensor else None
+        self._collision_sensor = CollisionSensor(self._pedestrian) if pedestrian_config.collision_sensor else None
         if self._config.auto_control:
             self._pedestrian.controller.start()
             self._pedestrian.controller.go_to_location(carla.Location(*self._destination[:3]))

carla_gym/core/controllers/props.py

@@ -121,6 +121,7 @@ def apply_npc(world, traffic_manager, num_vehicles, num_pedestrians, safe=False)
             controller = world.try_spawn_actor(pedestrian_controller_bp, carla.Transform(), pedestrian)
             world.tick()
             if controller is not None:
+                controller.start()  # start walker
                 controller.go_to_location(world.get_random_location_from_navigation())
                 controller.set_max_speed(float(speed))
                 pedestrian.controller = controller
@@ -135,9 +136,6 @@ def apply_npc(world, traffic_manager, num_vehicles, num_pedestrians, safe=False)
 
     # Initialize each controller and set target to walk
     world.set_pedestrians_cross_factor(percentagePedestriansCrossing)
-    for pedestrian in pedestrians_list:
-        pedestrian.controller.start()  # start walker
-
     traffic_manager.global_percentage_speed_difference(30.0)
 
     return vehicles_list, pedestrians_list

carla_gym/core/controllers/vehicle_manager.py

@@ -44,8 +44,8 @@ def __init__(self, vehicle_config, vehicle_object, traffic_manager, path_tracker
         self._path_tracker = path_tracker
 
         # Spawn collision and lane sensors if necessary
-        self._collision_sensor = CollisionSensor(self._vehicle, 0) if vehicle_config.collision_sensor else None
-        self._lane_invasion_sensor = LaneInvasionSensor(self._vehicle, 0) if vehicle_config.lane_sensor else None
+        self._collision_sensor = CollisionSensor(self._vehicle) if vehicle_config.collision_sensor else None
+        self._lane_invasion_sensor = LaneInvasionSensor(self._vehicle) if vehicle_config.lane_sensor else None
 
         self._vehicle.set_autopilot(self._config.auto_control, self._traffic_manager.get_port())
 

carla_gym/multi_env.py

@@ -24,7 +24,7 @@
 import carla
 import cv2
 import GPUtil
-import gym
+import gymnasium as gym
 import numpy as np
 import psutil
 from carla_gym.carla_api.PythonAPI.agents.navigation.global_route_planner import GlobalRoutePlanner
@@ -42,7 +42,7 @@
 from carla_gym.core.utils.utils import collided_done, preprocess_image
 from core.controllers.pedestrian_manager import PedestrianManager
 from core.controllers.vehicle_manager import DISCRETE_ACTIONS, VehicleManager
-from gym.spaces import Box, Dict, Discrete, Tuple
+from gymnasium.spaces import Box, Dict, Discrete, Tuple
 from gymnasium.utils import EzPickle
 from path import Path
 from pettingzoo import AECEnv
@@ -1041,7 +1041,7 @@ def _step(self, actor_id: str, action):
         py_measurements["total_reward"] = self._total_rewards[actor_id]
 
         # Compute output observation
-        obs = (self._encode_obs(actor_id, py_measurements),)  # TODO this is a tuple?
+        obs = self._encode_obs(actor_id, py_measurements)
 
         # End iteration updating parameters and logging
         self._previous_actions[actor_id] = action
@@ -1114,7 +1114,7 @@ def step(self, actions: dict):
                     if done:
                         self._active_actors.discard(actor_id)
                 info_dict[actor_id] = info
-            self._dones["__all__"] = sum(self._dones.values()) >= len(self._scenario_objects)
+            self._dones["__all__"] = sum(self._dones.values()) >= self.max_num_agents
             self.render()
 
             return obs_dict, reward_dict, self._dones, info_dict