Baxter robot dataset for learning dynamic models

This dataset contains the training and test data that were used for the recurrent neural network to learn the inverse dynamics of the Baxter robot. The data correspond to the angles and velocities of the joints when a torque is applied in order to follow a trajectory:1:7 joint angles, from s0 to w2 8:14 joint velocities, from s0 to w2 15:21 joint torques, from s0 to w2 The files in the dataset are named using the following convention: {arm}_{trajectory}_p{phi angle}_t{theta angle}.csv. arm: robot arm used for the data collection. left or right arm. trajectory: path followed by the robot's end-effector. The trajectories include circle, square, random, and spiral. phi and theta angle: These represent the angles in degrees for the center of the trajectory within a spherical coordinate system. The origin of this system is located at the base of the robot arm. This dataset was used for the article "Informed Federated Learning to Train a Robotic Arm Inverse Dynamic Model" by G. Jimenez-Perera , B. Valencia-Vidal , N. R. Luque , E. Ros and F. Barranco. It was published in IEEE Robotics and Automation Letters. 2025.