Open dujinyu opened 4 years ago
In this repo, it offers two methods to set the init_traj_distr, which are init_pd, init_lqr.
when dX < 2 * dU, errors occur. how to fix this bug?
# init_pd K = -config['pos_gains'] * np.tile( np.hstack([ np.eye(dU) * Kp, np.eye(dU) * Kv, np.zeros((dU, dX - dU*2)) ]), [T, 1, 1] ) # init_lqr Fd = np.vstack([ np.hstack([ np.eye(dU), dt * np.eye(dU), np.zeros((dU, dX - dU*2)), dt ** 2 * np.diag(gains) ]), # joint angles, joint_velocities, end-effector angles and velocities, action np.hstack([ np.zeros((dU, dU)), np.eye(dU), np.zeros((dU, dX - dU*2)), dt * np.diag(gains) ]), np.zeros((dX - dU*2, dX+dU)) ]) fc = np.hstack([acc * dt ** 2, acc * dt, np.zeros((dX - dU*2))])
In this repo, it offers two methods to set the init_traj_distr, which are init_pd, init_lqr.
when dX < 2 * dU, errors occur. how to fix this bug?