In this repository, the implementation of forward and inverse kinematics by redundancy resolution is presented for KUKA on linear axis 7-DOF robot. The Redundancy Resolution includes three methods, which are Jacobian-based (Damped Least Square and Weighted Pseudoinverse), Null Space, and Task Augmentation.

Kinematic redundancy resolution for replicating natural posture of human arm by performing constrained muscle-effort minimization (considering wrist pose and range of joint motion) with a human arm musculoskeletal model and OpenSIM C++ API.

ROS/MoveIt! extension for redundancy resolution with dynamic programming. Please cite: Ferrentino, E.; Salvioli, F.; Chiacchio, P. Globally Optimal Redundancy Resolution with Dynamic Programming for Robot Planning: A ROS Implementation. Robotics 2021, 10, 42. https://doi.org/10.3390/robotics10010042

One reach target, many postures. Behavioral diversity for a redundant 7-DOF arm (Unitree G1), from massively parallel RL — no human demos.

Muscle-activation control of a planar two-link, six-muscle human arm. Classical physics-based controllers on main; learning-based controllers (MotorNet, synergies, ANFIS, BC, MPC) with trained models on the learning branch.