novaphy.ik.solve_ik_pose_best_of_seeds

Run solve_ik_pose() from several deterministic perturbations and return the lowest-cost solution.

solve_ik_pose_best_of_seeds(
    art,
    q0,
    target_position,
    target_orientation,
    joint_limits=None,
    ee_link_index=-1,
    local_offset=None,
    n_tries=4,
    perturb_scale=0.35,
    **solve_kwargs,
)

The compatibility wrapper cycles through a fixed six-row direction table, tiles it to the coordinate count, scales it by perturb_scale, and clips seeds to explicit limits. It is deterministic and does not accept a random number generator.

solve_kwargs is forwarded to solve_ik_pose() (for example tol_pos, tol_orient, max_iter, lambda_damp, step_size, and weight_orientation). The return value is (best_q, best_converged).

For GAUSS, UNIFORM, or ROBERTS sampling in new code, construct IKSolver with IKSampler.

See also

• solve_ik_pose()
• novaphy.ik