Fixes handling of time-out signal in RSL-RL and RL-Games wrapper (#375)

# Description

On termination of an episode, three conditions arise:

1. **bad** terminations (terminated dones): the agent gets a termination
penalty
2. **timeout** terminations (truncated dones):
* infinite-horizon: bootstrapping by the agent based on terminal state
    * finite-horizon: no penalty or bootstrapping

Currently, we have not handled the last case, which leads to issues when
training RL tasks with a finite horizon (for instance, Nikita's agile
locomotion work).

This MR adds a flag to the RLTaskEnvCfg called `is_finite_horizon` that
helps deal with this case. The flag is consumed by the env wrappers to
decide how they want to specifically handle the finite horizon problem.

## Type of change

- Bug fix (non-breaking change which fixes an issue)
- New feature (non-breaking change which adds functionality)

## Checklist

- [x] I have run the [`pre-commit` checks](https://pre-commit.com/) with
`./orbit.sh --format`
- [x] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [x] I have run all the tests with `./orbit.sh --test` and they pass
- [x] I have updated the changelog and the corresponding version in the
extension's `config/extension.toml` file
- [x] I have added my name to the `CONTRIBUTORS.md` or my name already
exists there

source/extensions/omni.isaac.orbit/omni/isaac/orbit/envs/rl_task_env_cfg.py

@@ -21,6 +21,28 @@ class RLTaskEnvCfg(BaseEnvCfg):
     ui_window_class_type: type | None = RLTaskEnvWindow
 
     # general settings
+    is_finite_horizon: bool = False
+    """Whether the learning task is treated as a finite or infinite horizon problem for the agent.
+    Defaults to False, which means the task is treated as an infinite horizon problem.
+
+    This flag handles the subtleties of finite and infinite horizon tasks:
+
+    * **Finite horizon**: no penalty or bootstrapping value is required by the the agent for
+      running out of time. However, the environment still needs to terminate the episode after the
+      time limit is reached.
+    * **Infinite horizon**: the agent needs to bootstrap the value of the state at the end of the episode.
+      This is done by sending a time-limit (or truncated) done signal to the agent, which triggers this
+      bootstrapping calculation.
+
+    If True, then the environment is treated as a finite horizon problem and no time-out (or truncated) done signal
+    is sent to the agent. If False, then the environment is treated as an infinite horizon problem and a time-out
+    (or truncated) done signal is sent to the agent.
+
+    Note:
+        The base :class:`RLTaskEnv` class does not use this flag directly. It is used by the environment
+        wrappers to determine what type of done signal to send to the corresponding learning agent.
+    """
+
     episode_length_s: float = MISSING
     """Duration of an episode (in seconds)."""
 

source/extensions/omni.isaac.orbit_tasks/config/extension.toml

@@ -1,7 +1,7 @@
 [package]
 
 # Note: Semantic Versioning is used: https://semver.org/
-version = "0.5.3"
+version = "0.5.4"
 
 # Description
 title = "ORBIT Environments"

source/extensions/omni.isaac.orbit_tasks/docs/CHANGELOG.rst

@@ -1,6 +1,18 @@
 Changelog
 ---------
 
+0.5.4 (2024-02-06)
+~~~~~~~~~~~~~~~~~~
+
+Added
+^^^^^
+
+* Added a check for the flag :attr:`omni.isaac.orbit.envs.RLTaskEnvCfg.is_finite_horizon`
+  in the RSL-RL and RL-Games wrappers to handle the finite horizon tasks properly. Earlier,
+  the wrappers were always assuming the tasks to be infinite horizon tasks and returning a
+  time-out signals when the episode length was reached.
+
+
 0.5.3 (2023-11-16)
 ~~~~~~~~~~~~~~~~~~
 

source/extensions/omni.isaac.orbit_tasks/omni/isaac/orbit_tasks/utils/wrappers/rl_games.py

@@ -240,8 +240,10 @@ def step(self, actions):  # noqa: D102
         obs_dict, rew, terminated, truncated, extras = self.env.step(actions)
 
         # move time out information to the extras dict
+        # this is only needed for infinite horizon tasks
         # note: only useful when `value_bootstrap` is True in the agent configuration
-        extras["time_outs"] = truncated.to(device=self._rl_device)
+        if not self.unwrapped.cfg.is_finite_horizon:
+            extras["time_outs"] = truncated.to(device=self._rl_device)
         # process observations and states
         obs_and_states = self._process_obs(obs_dict)
         # move buffers to rl-device

source/extensions/omni.isaac.orbit_tasks/omni/isaac/orbit_tasks/utils/wrappers/rsl_rl/vecenv_wrapper.py

@@ -165,7 +165,9 @@ def step(self, actions: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor, torch
         obs = obs_dict["policy"]
         extras["observations"] = obs_dict
         # move time out information to the extras dict
-        extras["time_outs"] = truncated
+        # this is only needed for infinite horizon tasks
+        if not self.unwrapped.cfg.is_finite_horizon:
+            extras["time_outs"] = truncated
 
         # return the step information
         return obs, rew, dones, extras

source/extensions/omni.isaac.orbit_tasks/test/wrappers/test_rsl_rl_wrapper.py

@@ -92,6 +92,41 @@ def test_random_actions(self):
             print(f">>> Closing environment: {task_name}")
             env.close()
 
+    def test_no_time_outs(self):
+        """Check that environments with finite horizon do not send time-out signals."""
+        for task_name in self.registered_tasks[0:5]:
+            print(f">>> Running test for environment: {task_name}")
+            # create a new stage
+            omni.usd.get_context().new_stage()
+            # parse configuration
+            env_cfg: RLTaskEnvCfg = parse_env_cfg(task_name, use_gpu=self.use_gpu, num_envs=self.num_envs)
+            # change to finite horizon
+            env_cfg.is_finite_horizon = True
+
+            # create environment
+            env = gym.make(task_name, cfg=env_cfg)
+            # wrap environment
+            env = RslRlVecEnvWrapper(env)
+
+            # reset environment
+            _, extras = env.reset()
+            # check signal
+            self.assertNotIn("time_outs", extras, msg="Time-out signal found in finite horizon environment.")
+
+            # simulate environment for 10 steps
+            with torch.inference_mode():
+                for _ in range(10):
+                    # sample actions from -1 to 1
+                    actions = 2 * torch.rand(env.action_space.shape, device=env.unwrapped.device) - 1
+                    # apply actions
+                    extras = env.step(actions)[-1]
+                    # check signals
+                    self.assertNotIn("time_outs", extras, msg="Time-out signal found in finite horizon environment.")
+
+            # close the environment
+            print(f">>> Closing environment: {task_name}")
+            env.close()
+
     """
     Helper functions.
     """