fix bug in generating target_frame_body_names

Signed-off-by: Tao <[email protected]>

source/extensions/omni.isaac.orbit/omni/isaac/orbit/sensors/frame_transformer/frame_transformer.py

@@ -5,6 +5,7 @@
 
 from __future__ import annotations
 
+import re
 import torch
 from typing import TYPE_CHECKING, Sequence
 
@@ -230,7 +231,7 @@ def _initialize_impl(self):
         first_env_prim_paths = all_prim_paths[0 : self._num_target_body_frames + 1]
         first_env_body_names = [first_env_prim_path.split("/")[-1] for first_env_prim_path in first_env_prim_paths]
 
-        target_frame_body_names = first_env_body_names[1:]
+        target_frame_body_names = [x for x in first_env_body_names if not re.match(self._tracked_body_names[0], x)]
 
         # The position and rotation components of target frame offsets
         target_frame_offset_pos = []