fix(terminations): index before reduce in joint_pos_out_of_limit

Root cause: reduced over joints to [N] then indexed with [:, joint_ids] → IndexError.
Change: compare to limits → [N,J], slice columns with joint_ids → [N,K], then .any(dim=1).
Result: returns [N] without error for both [1,J,2] and [N,J,2] limit tensors; joint_ids=None still uses all joints.

Author: moribots

source/isaaclab/isaaclab/envs/mdp/terminations.py

@@ -84,10 +84,19 @@ def joint_pos_out_of_limit(env: ManagerBasedRLEnv, asset_cfg: SceneEntityCfg = S
     if asset_cfg.joint_ids is None:
         asset_cfg.joint_ids = slice(None)
 
-    limits = asset.data.soft_joint_pos_limits[:, asset_cfg.joint_ids]
-    out_of_upper_limits = torch.any(asset.data.joint_pos[:, asset_cfg.joint_ids] > limits[..., 1], dim=1)
-    out_of_lower_limits = torch.any(asset.data.joint_pos[:, asset_cfg.joint_ids] < limits[..., 0], dim=1)
-    return torch.logical_or(out_of_upper_limits, out_of_lower_limits)
+    # compute any per-joint violations (avoid reducing before indexing)
+    out_of_upper_limits = asset.data.joint_pos > asset.data.soft_joint_pos_limits[..., 1]  # [N, J]
+    out_of_lower_limits = asset.data.joint_pos < asset.data.soft_joint_pos_limits[..., 0]  # [N, J]
+
+    # truncate above output to just the joints we care about
+    out_of_upper_limits = out_of_upper_limits[:, asset_cfg.joint_ids]  # [N, K]
+    out_of_lower_limits = out_of_lower_limits[:, asset_cfg.joint_ids]  # [N, K]
+
+    # reduce over selected joints
+    out_of_upper_limits = torch.any(out_of_upper_limits, dim=1)  # [N]
+    out_of_lower_limits = torch.any(out_of_lower_limits, dim=1)  # [N]
+
+    return torch.logical_or(out_of_upper_limits, out_of_lower_limits)  # [N]
 
 
 def joint_pos_out_of_manual_limit(