add ball velocity into observation space

fancy_gym/envs/__init__.py

@@ -578,7 +578,7 @@ for _v in _versions:
     kwargs_dict_tt_prodmp['basis_generator_kwargs']['pre_compute_length_factor'] = 5
     kwargs_dict_tt_prodmp['phase_generator_kwargs']['alpha_phase'] = 3
     # kwargs_dict_tt_prodmp['black_box_kwargs']['duration'] = 4.
-    kwargs_dict_tt_prodmp['black_box_kwargs']['max_planning_times'] = 4
+    kwargs_dict_tt_prodmp['black_box_kwargs']['max_planning_times'] = 3
     kwargs_dict_tt_prodmp['black_box_kwargs']['replanning_schedule'] = lambda pos, vel, obs, action, t : t % 50 == 0
     register(
         id=_env_id,

fancy_gym/envs/mujoco/table_tennis/mp_wrapper.py

@@ -16,6 +16,7 @@ class MPWrapper(RawInterfaceWrapper):
             [False] * 7,  # joints velocity
             [True] * 2,  # position ball x, y
             [False] * 1,  # position ball z
+            [True] * 3,    # velocity ball x, y, z
             [True] * 2,  # target landing position
             # [True] * 1,  # time
         ])

fancy_gym/envs/mujoco/table_tennis/table_tennis_env.py

@@ -154,7 +154,7 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
         mujoco.mj_forward(self.model, self.data)
 
         if self._enable_wind:
-            self._wind_vel[1] = self.np_random.uniform(low=-10, high=10, size=1)
+            self._wind_vel[1] = self.np_random.uniform(low=-5, high=5, size=1)
             self.model.opt.wind[:3] = self._wind_vel
 
         self._hit_ball = False
@@ -180,6 +180,9 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
             self.data.joint("tar_x").qpos.copy(),
             self.data.joint("tar_y").qpos.copy(),
             self.data.joint("tar_z").qpos.copy(),
+            self.data.joint("tar_x").qvel.copy(),
+            self.data.joint("tar_y").qvel.copy(),
+            self.data.joint("tar_z").qvel.copy(),
             # self.data.body("target_ball").xvel.copy(),
             self._goal_pos.copy(),
         ])