safemotions 0.2.5

Learning Collision-free and Torque-limited Robot Trajectories based on Alternative Safe Behaviors.

Learning Collision-free and Torque-limited Robot Trajectories based on Alternative Safe Behaviors

This python package provides the code to learn torque-limited and collision-free robot trajectories without exceeding limits on the position, velocity, acceleration and jerk of each robot joint.

Installation

The package can be installed by running

pip install safemotions  # for safe trajectory generation only or
pip install safemotions[train]  # to include dependencies required to train and evaluate neural networks.

Trajectory generation

To generate a random trajectory with a single robot run

python -m safemotions.random_agent --use_gui --check_braking_trajectory_torque_limits --torque_limit_factor=0.6 --plot_trajectory

For a demonstration scenario with two robots run

python -m safemotions.random_agent --use_gui --robot_scene=1

Collision-free trajectories for three robots can be generated by running

python -m safemotions.random_agent --use_gui --robot_scene=2

Pretrained networks

Pretrained networks for various reaching tasks are provided.
To generate and plot trajectories for a reaching task with a single robot run

python -m safemotions.evaluate --checkpoint=one_robot/P_CT_S_5_J_A --use_gui --plot_trajectory --plot_actual_torques

Trajectories for two and three robots with alternating target points can be generated by running

python -m safemotions.evaluate --checkpoint=two_robots/P_C_S_1_J_A_D_5_T_A --use_gui 

and

python -m safemotions.evaluate --checkpoint=three_robots/P_C_S_1_J_A_D_5_T_A --use_gui 

Training

Networks can also be trained from scratch. For instance, a reaching task with a single robot can be learned by running

python -m safemotions.train --logdir=safemotions_training --name=One_robot_P_CT_S_5_J_A --robot_scene=0 --online_trajectory_time_step=0.1 --online_trajectory_duration=8.0 --use_target_points --target_point_cartesian_range_scene=0 --target_link_offset="[0, 0, 0.126]" --target_point_radius=0.065 --obs_add_target_point_pos --obs_add_target_point_relative_pos --obstacle_scene=3 --obstacle_use_computed_actual_values --use_braking_trajectory_method --closest_point_safety_distance=0.05 --check_braking_trajectory_torque_limits --acc_limit_factor_braking=0.75 --jerk_limit_factor_braking=0.75 --punish_action --action_punishment_min_threshold=0.95 --action_max_punishment=0.4  --target_point_reached_reward_bonus=5  --pos_limit_factor=1.0 --vel_limit_factor=1.0 --acc_limit_factor=1.0 --jerk_limit_factor=1.0 --torque_limit_factor=1.0 --iterations_per_checkpoint=100 --time=216

Publication

The corresponding publication is available at https://arxiv.org/abs/2103.03793.

Disclaimer

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.