Learning-based Initialization of Trajectory Optimization for Path-following Problems of Redundant Manipulators (ICRA23)

The contents of each package are as follows.

• pirl (path-wise IK using reinforcement learning): code related to reinforcement learning and behavior cloning to learn learning-based initializers (RL-ITG and BC-ITG)
• pirl_gazebo: Creates a 3-D occupancy grid map by randomly configuring the external environment
• pirl_msgs: consists of messages or services on ROS
• pirl_vae: variational autoencoder learning code for learning scene_encoding_vector
• torm (cloned and modified from link): TORM planner and Greedy and Linear initializer baselines
• velodyne Simulator (cloned and modified from link): lidar is used to obtain a point cloud for the external environment in the gazebo
• fetch_description (cloned from link): To use the fetch robot model
• fetch_moveit_config (cloned from link): To use the fetch robot model

You can get more information (e.g., paper and video) regarding this paper through our project page.