MengGuo / tiago_ltl_flexbe

Integration of TIAGo simulation environment with LTL motion planner and FlexBE real-time execution framework

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

tiago_ltl_flexbe

Integration of TIAGo simulation environment with LTL motion planner and FlexBE real-time execution framework

Description

This package is based on the following three sources:

Usage

  • The robot model as finite transition system (FTS) is specified in [robot_fts.py]:
robot_model = [robot_motion, init_pose, robot_action]

where the robot motion and action abstractions are given as FTS models.

  • While calling the [ltl_planner.py], a complex navigation task specified as LTL formulas can be fed directly:
# python ltl_planner.py '<> (r2 && <>r3)'
# python ltl_planner.py '([]<> r2) && ([]<> r3) && ([]<> r1)'
python ltl_planner.py '<> r2 && ([]<> r3) && ([]<> r1)'

  • Simulation video for different LTL tasks (with only navigation tasks)

  • Incorporating TIAGo pre-programmed actions from [play_motion], such as

close_gripper
close_gripper_half
do_weights
head_tour
home
offer_gripper
open_gripper
pick_from_floor
pregrasp_weight
reach_floor
reach_max
shake_hands
show_manip_ws
unfold_arm
wave

Thus a task can involve any regions of interest in the workspace and any [play_motion] action above,

# python ltl_planner.py '([]<> (r2 && shake_hands)) && ([]<> (r3 && head_tour))'
python ltl_planner.py '<> ((r2 && pick_from_floor) && <> (r3 && reach_max))

see video for example.

About

Integration of TIAGo simulation environment with LTL motion planner and FlexBE real-time execution framework

License:GNU General Public License v2.0


Languages

Language:Python 92.4%Language:CMake 7.6%