mit-acl / deep_panther Goto Github PK

The solution was to do this before running catkin build:

pip uninstall em
pip install empy==3.3.4 [there was a newer version of empy installed but it didn't work]

When i execute "roslaunch panther simulation.launch" and use 2D Nav Goal, this problem will arise.

SolverIpopt, reading .casadi files...
Calling the student!
terminate called after throwing an instance of 'pybind11::error_already_set'
what(): ValueError: Error: Cannot determine if the observation is vectorized with the space type Box([[-1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1.
-1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1. -1.
-1. -1. -1. -1. -1. -1. -1.]], [[1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.

1. 1. 1. 1. 1. 1. 1. 1. 1. 9. 1.]], (1, 43), float32).

At:
/home/huo/installations/venvs_python/my_venv/lib/python3.8/site-packages/stable_baselines3/common/utils.py(402): is_vectorized_observation
/home/huo/installations/venvs_python/my_venv/lib/python3.8/site-packages/stable_baselines3/common/policies.py(270): obs_to_tensor
/home/huo/installations/venvs_python/my_venv/lib/python3.8/site-packages/stable_baselines3/common/policies.py(363): predict
/home/huo/deep_planner/src/deep_panther/panther_compression/compression/utils/other.py(1148): predict

[SQ01s/panther-13] process has died [pid 10899, exit code -6, cmd /home/huo/deep_planner/devel/lib/panther/panther_node ~state:=state ~who_plans:=who_plans ~term_goal:=term_goal ~trajs_predicted:=tracker_predictor_node/trajs_predicted ~traj:=traj ~trajs_zhejiang:=trajs_zhejiang ~goal:=goal ~setpoint:=setpoint __name:=panther __log:=/home/huo/.ros/log/93df689c-b9bf-11ee-9f5e-b5309e565c8a/SQ01s-panther-13.log].
log file: /home/huo/.ros/log/93df689c-b9bf-11ee-9f5e-b5309e565c8a/SQ01s-panther-13*.log

fix is to change eigen2rosvector to eigen2point

python3 policy_compression_train.py

Traceback (most recent call last):
File "policy_compression_train.py", line 27, in
from compression.policies.ExpertPolicy import ExpertPolicy
File "/home/scout/deep_panther/src/deep_panther/panther_compression/compression/policies/ExpertPolicy.py", line 5, in
from compression.utils.other import ActionManager, ObservationManager, getPANTHERparamsAsCppStruct, ExpertDidntSucceed, computeTotalTime
File "/home/scout/deep_panther/src/deep_panther/panther_compression/compression/utils/other.py", line 7, in
import py_panther
ModuleNotFoundError: No module named 'py_panther'

I don't know if this happened as a result of that my gazebo version is 9.And whether there is something wrong else or not.Could you please give me some advise?thank you very much!

When I set -DWITH_MATLAB=ON, I get this error.

See this: mit-acl/panther#10

apart from this issue I found the following issues with MATLAB R2023b:

1. In MyCasadiClampedUniformSpline.m change

obj.getPosT=casadi.Function('f',{obj.t},{obj.expression});

to

options.allow_free = true;
obj.getPosT=casadi.Function('f',{obj.t},{obj.expression}, options);

2. In main.m change

my_func = opti.to_function('my_func', par_and_init_guess_exprs, results_expresion, par_and_init_guess_names, results_names);

to

options.allow_duplicate_io_names = true;
my_func = opti.to_function('my_func', par_and_init_guess_exprs, results_expresion, par_and_init_guess_names, results_names, options);