How imitation learning and human factors can be combined in a model predictive control algorithm for adaptive motion planning and control