Up to now, the embodiment of bodily-kinaesthetic, perceptual and cognitive capabilities for assistive robots has been scarcely studied. This research aims to incorporate and develop the concept of robotic human science and to enable its application in a human-friendly robot for assistive purposes. In this paper, the authors describe the improvements of the velocity control of the two-wheeled inverted pendulum of the hWalk by adding a LQR as compensator for the wheel angular velocity to the existent PID controller. On the other hand, an algorithm based on computing the integral of the motor measured current was proposed in order to detect a ramp in order to cope with inclined surfaces. Experiments were carried out to verify the effectiveness of the proposed velocity control as well as verify the feasibility of the proposed ramp detection algorithm.