fix fix_commands bug in play.py

scripts/rsl_rl/play.py

@@ -74,6 +74,7 @@ from isaaclab.envs import (
     ManagerBasedRLEnvCfg,
     multi_agent_to_single_agent,
 )
+from isaaclab.envs.mdp import UniformVelocityCommandCfg
 from isaaclab.utils.assets import retrieve_file_path
 from isaaclab.utils.dict import print_dict
 # from isaaclab.utils.pretrained_checkpoint import get_published_pretrained_checkpoint
@@ -83,12 +84,35 @@ from isaaclab_tasks.utils.hydra import hydra_task_config
 import robot_lab.tasks  # noqa: F401
 
 def fix_commands(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg):
-    from isaaclab.managers import ObservationTermCfg as ObsTerm
-    num_envs = env_cfg.scene.num_envs
-    def fix(env):
-        return torch.tensor([1.0, 0.0, 0.0], device=env_cfg.sim.device).repeat(num_envs, 1)
-    env_cfg.observations.policy.velocity_commands = ObsTerm(func=fix)
-    env_cfg.observations.critic.velocity_commands = ObsTerm(func=fix)
+    """Fix commanded velocity to a constant target.
+
+    Prefer locking the command generator to keep policy observations consistent
+    with environment internal command state.
+    """
+    fixed_lin_x, fixed_lin_y, fixed_ang_z = 1.0, 0.0, 0.0
+
+    base_velocity_cfg = getattr(getattr(env_cfg, "commands", None), "base_velocity", None)
+    if base_velocity_cfg is None:
+        return
+
+    fixed_cfg = UniformVelocityCommandCfg(
+        asset_name=getattr(base_velocity_cfg, "asset_name", "robot"),
+        heading_command=False,
+        rel_standing_envs=0.0,
+        rel_heading_envs=0.0,
+        resampling_time_range=(5.0, 5.0),
+        ranges=UniformVelocityCommandCfg.Ranges(
+            lin_vel_x=(fixed_lin_x, fixed_lin_x),
+            lin_vel_y=(fixed_lin_y, fixed_lin_y),
+            ang_vel_z=(fixed_ang_z, fixed_ang_z),
+            heading=None,
+        ),
+    )
+    env_cfg.commands.base_velocity = fixed_cfg
+
+    # terrain_levels_vel_gym expects custom Go2RLGymCommand fields.
+    if hasattr(env_cfg, "curriculum") and hasattr(env_cfg.curriculum, "terrain_levels"):
+        env_cfg.curriculum.terrain_levels = None
 
 @hydra_task_config(args_cli.task, args_cli.agent)
 def main(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg, agent_cfg: RslRlBaseRunnerCfg):
@@ -109,7 +133,7 @@ def main(env_cfg: ManagerBasedRLEnvCfg | DirectRLEnvCfg | DirectMARLEnvCfg, agen
     env_cfg.observations.policy.enable_corruption = False
     # remove random pushing
     env_cfg.events.randomize_apply_external_force_torque = None
-    env_cfg.events.push_robot = None
+    env_cfg.events.randomize_push_robot = None
     env_cfg.curriculum.command_levels_lin_vel = None
     env_cfg.curriculum.command_levels_ang_vel = None