Unity integration

README.md

@@ -1,11 +1,6 @@
 # Arena Rosnav Task Generator
 
-This is the task generator package designed to work with:
-
-- [arena-rosnav](https://github.com/Arena-Rosnav/arena-rosnav)
-- [arena-rosnav-3d]()
-
-The task generator is especially designed for single robot simulations.
+This is the task generator package designed to work with the [Arena Benchmark](https://github.com/Arena-Rosnav/arena-rosnav) infrastructure.
 
 ## Task Modes
 
@@ -15,13 +10,6 @@ Our task generator package offers four task modes. We define a task as the proce
 
 Creates random static and dynamic obstacles when a new task is started. When starting the task a random goal and start position is selected. After the robot reaching the goal a new task is started.
 
-### Manual Task
-
-Manual task mode is the same as the random task except that a goal must be set manually via the rviz *2D Nav Goal* button.
-
-![image](https://arena-rosnav-wiki.readthedocs.io/en/latest/images/manual_task.png)
-
-
 ### Staged Task
 
 The staged task mode is designed for the trainings process of arena-rosnav. In general, it behaves like the random task mode but there are multiple stages between one can switch. Between the stages, the amount of static and dynamic obstacles changes. The amount of obstacles is defined in a curriculum file, the path to said file is a key in the `paths` parameter.
@@ -51,10 +39,10 @@ robot are defined.
 The scenario declaration file can be created with [arena-tools](https://github.com/Arena-Rosnav/arena-tools) and has to follow
 the specified file schema.
 
-## Environment Factory
+## Simulator Factory
 
-To be able to use the task generator module in all our environments without changes, a unified interface between environment and task generator is needed. The interface contains a lot of functions to spawn, publish or move robots or obstacles, and a lot more.
+To be able to use the task generator module in all our simulators without changes, a unified interface between simulator and task generator is needed. The interface contains a lot of functions to spawn, publish or move robots or obstacles, and a lot more.
 
-At the moment we provide environment interfaces for **Flatland** and **Gazebo**. In order to add a new environment, in which the task generator should be used, a new environment interface has to be created in `/taks_generator/environments/` and has to be registrated in the environment factory.
+At the moment we provide simulator interfaces for **Flatland** and **Gazebo**. In order to add a new simulator, in which the task generator should be used, a new simulator interface has to be created in `/taks_generator/simulators/` and has to be registrated in the simulator factory.
 
-Your newly created environment interface should derive the **BaseEnvironment** located [here](TODO) and implement all functions. A detailed description of the functions is contained in the **BaseEnvironment** itself.
+Your newly created simulator interface should derive the **BaseSimulator** located [here](TODO) and implement all functions. A detailed description of the functions is contained in the **BaseSimulator** itself.

robot_setup/marl_map_empty.yaml

@@ -3,14 +3,18 @@ robots:
   #   planner: rosnav
   #   amount: 6
   #   agent: rto_tlabs_marl
-  - model: agvota
-    planner: teb
-    amount: 1
-    agent: new_jackal
-  - model: burger
-    planner: teb
-    amount: 1
-    agent: new_jackal
+  - model: jackal
+    planner: rosnav
+    amount: 2
+    agent: jackal_marl
+  - model: rto
+    planner: rosnav
+    amount: 2
+    agent: rto_marl
+  - model: ridgeback
+    planner: rosnav
+    amount: 2
+    agent: ridgeback_marl
   # - model: cob4
   #   planner: teb
   #   amount: 1

robot_setup/rosnav_agents_marl.yaml

@@ -1,15 +1,15 @@
 robots:
   # - model: ridgeback
-  #   planner: rosnav
-  #   amount: 1
+  #   planner: teb
+  #   amount: 6
   #   agent: ridgeback_marl
   - model: rto
-    planner: rosnav
-    amount: 3
+    planner: teb
+    amount: 6
     agent: rto_tlabs_marl
   # - model: jackal
-  #   planner: rosnav
-  #   amount: 1
+  #   planner: teb
+  #   amount: 6
   #   agent: jackal_marl
   # - model: jackal
   #   planner: rosnav

scenarios/marl_map_empty_obs0.json

@@ -33,8 +33,60 @@
       "goal": [29.94602317350325, 20.795265461388748, 0]
     },
     {
-      "start": [3.0220509108329354, 20.442857447479454, 0],
-      "goal": [30.58035759853998, 7.262797727271735, 0]
+      "start": [2.778715176849392, 3.053770554072754, 0],
+      "goal": [30.146818089333973, 23.16670627326246, 0]
+    },
+    {
+      "start": [5.237410725688214, 3.3359159449231095, 0],
+      "goal": [28.010574415752707, 23.16670627326246, 0]
+    },
+    {
+      "start": [8.05886463419178, 3.0134640696655604, 0],
+      "goal": [24.826362147584394, 23.126399788855267, 0]
+    },
+    {
+      "start": [2.980247598885362, 4.545110477138927, 0],
+      "goal": [31.114173715106627, 21.35291447493874, 0]
+    },
+    {
+      "start": [2.778715176849392, 8.092081104971982, 0],
+      "goal": [30.469269964591522, 18.53146056643517, 0]
+    },
+    {
+      "start": [2.9399411144781684, 11.27629337314029, 0],
+      "goal": [30.42896348018433, 16.516136346075484, 0]
+    },
+    {
+      "start": [2.4965697859990357, 14.379892672494213, 0],
+      "goal": [30.18712457374117, 12.727326811799266, 0]
+    },
+    {
+      "start": [3.141473536514136, 17.080427127776197, 0],
+      "goal": [30.348350511369944, 8.93851727752305, 0]
+    },
+    {
+      "start": [3.101167052106943, 19.619735645429405, 0],
+      "goal": [30.388656995777136, 6.802273603941779, 0]
+    },
+    {
+      "start": [2.8190216612565857, 22.803947913597714, 0],
+      "goal": [30.711108871034686, 2.932851100851173, 0]
+    },
+    {
+      "start": [5.277717210095408, 22.965173851226492, 0],
+      "goal": [28.292719806603063, 3.0134640696655604, 0]
+    },
+    {
+      "start": [8.099171118598974, 22.683028460376136, 0],
+      "goal": [25.672798320135467, 3.1343835228871413, 0]
+    },
+    {
+      "start": [10.033882370144276, 22.683028460376136, 0],
+      "goal": [25.14881402284195, 3.6180613357734708, 0]
+    },
+    {
+      "start": [10.598173151844989, 22.72333494478333, 0],
+      "goal": [23.77839355299736, 3.0134640696655604, 0]
     }
   ],
   "map": "map_empty",

scripts/task_generator_node.py

@@ -13,9 +13,9 @@
 from task_generator.constants import Constants, TaskMode
 
 from task_generator.tasks.utils import get_predefined_task
-from task_generator.environments.environment_factory import EnvironmentFactory
-from task_generator.environments.gazebo_environment import GazeboEnvironment
-from task_generator.environments.flatland_environment import FlatlandRandomModel
+from task_generator.simulators.simulator_factory import SimulatorFactory
+from task_generator.simulators.gazebo_simulator import GazeboSimulator
+from task_generator.simulators.flatland_simulator import FlatlandRandomModel
 
 
 class TaskGenerator:
@@ -37,7 +37,7 @@ def __init__(self) -> None:
         rospy.Service("reset_task", Empty, self.reset_task_srv_callback)
 
         ## Vars
-        self.env_wrapper = EnvironmentFactory.instantiate(Utils.get_environment())("")
+        self.env_wrapper = SimulatorFactory.instantiate(Utils.get_simulator())("")
 
         rospy.loginfo(f"Launching task mode: {self.task_mode}")
 
@@ -60,6 +60,7 @@ def __init__(self) -> None:
         self.number_of_resets = 0
 
         self.reset_task()
+
         ## Timers
         rospy.Timer(rospy.Duration(0.5), self.check_task_status)
 

task_generator/constants.py

@@ -1,6 +1,6 @@
 class Constants:
     GOAL_REACHED_TOLERANCE = 1.0
-    TIMEOUT = 3.0 * 60 ## 3 min
+    TIMEOUT = 60 * 3 ## 3 min
     WAIT_FOR_SERVICE_TIMEOUT = 5 # 5 secs
     MAX_RESET_FAIL_TIMES = 3
 
@@ -15,9 +15,10 @@ class ObstacleManager:
     class RobotManager:
         SPAWN_ROBOT_SAFE_DIST = 0.4
 
-    class Environment:
+    class Simulator:
         FLATLAND = "flatland"
         GAZEBO = "gazebo"
+        UNITY = "unity"
 
     class ArenaType:
         TRAINING = "training"
@@ -30,10 +31,10 @@ class TaskMode:
     SCENARIO = "scenario"
 
     class Random:
-        MIN_DYNAMIC_OBS = 1
-        MAX_DYNAMIC_OBS = 3
-        MIN_STATIC_OBS = 1
-        MAX_STATIC_OBS = 3
+        MIN_DYNAMIC_OBS = 10
+        MAX_DYNAMIC_OBS = 15
+        MIN_STATIC_OBS = 10
+        MAX_STATIC_OBS = 20
 
     class Scenario:
         RESETS_DEFAULT = 5
@@ -60,11 +61,7 @@ class FlatlandRandomModel:
         "body": "base_link"
     }
     LINEAR_VEL = 0.2
-    ANGLUAR_VEL_MAX = 0.2
-<<<<<<< HEAD
-
-=======
->>>>>>> master
+    ANGLUAR_VEL_MAX = 0.1
 
 class Pedsim:
     START_UP_MODE = "default"
@@ -89,4 +86,7 @@ class Pedsim:
     FORCE_FACTOR_OBSTACLE = 1.0
     FORCE_FACTOR_SOCIAL = 5.0
     FORCE_FACTOR_ROBOT = 0.0
-    WAYPOINT_MODE = 0
+    WAYPOINT_MODE = 0
+
+class UnitySimulatorConstants:
+    UNITY_ROS_NAVIGATION = "UNITY_ROS_NAVIGATION"

task_generator/manager/obstacle_manager.py

@@ -2,21 +2,21 @@
 
 
 class ObstacleManager:
-    def __init__(self, namespace, map_manager, environment):
+    def __init__(self, namespace, map_manager, simulator):
         self.map_manager = map_manager
         self.namespace = namespace
-        self.environment = environment
+        self.simulator = simulator
 
     def start_scenario(self, scenario):
-        self.environment.spawn_pedsim_agents(scenario["obstacles"]["dynamic"])
+        self.simulator.spawn_pedsim_agents(scenario["obstacles"]["dynamic"])
 
     def reset_scenario(self, scenario):
-        self.environment.reset_pedsim_agents()
+        self.simulator.reset_pedsim_agents()
 
-        self.environment.remove_all_obstacles()
+        self.simulator.remove_all_obstacles()
 
         for obstacle in scenario["obstacles"]["static"]:
-            self.environment.spawn_obstacle(
+            self.simulator.spawn_obstacle(
                 [*obstacle["pos"], 0],
                 yaml_path=obstacle["yaml_path"],
             )
@@ -27,20 +27,20 @@ def reset_random(
             static_obstacles=Constants.ObstacleManager.STATIC_OBSTACLES,
             forbidden_zones=[]
         ):
-        self.environment.remove_all_obstacles()
+        self.simulator.remove_all_obstacles()
 
         for _ in range(dynamic_obstacles):
             position = self.map_manager.get_random_pos_on_map(
                 safe_dist=Constants.ObstacleManager.OBSTACLE_MAX_RADIUS, 
                 forbidden_zones=forbidden_zones
             )
 
-            self.environment.spawn_random_dynamic_obstacle(position=position)
+            self.simulator.spawn_random_dynamic_obstacle(position=position)
 
         for _ in range(static_obstacles):
             position = self.map_manager.get_random_pos_on_map(
                 safe_dist=Constants.ObstacleManager.OBSTACLE_MAX_RADIUS, 
                 forbidden_zones=forbidden_zones
             )
 
-            self.environment.spawn_random_static_obstacle(position=position)
+            self.simulator.spawn_random_static_obstacle(position=position)