The foot-ground contact force is one of the important factors affecting the gait behaviour, balance and safety of a bipedal robotic mechanism. This force is constrained by the coefficient of friction of the foot-ground contact. This coefficient of friction is thus of high importance, as non-sliding contact and/or the knowledge of the coefficient of friction of the foot-ground contact are common assumptions among bipedal robotics literature. Failure to keep the ratio of the fric-tion force to normal force of the foot-ground contact below the coefficient of friction causes the foot to slip, resulting in a large external disturbance being applied to the biped and causing the bipedal robot to lose balance and fall. In this work an online estimation technique for estimating the coefficient of friction of the foot-ground contact is developed. It is expected that the estimation algorithm converges to the value of the coefficient of friction of the contact before the ratio of the friction and normal forces exceeds the coefficient of friction, thus allowing the robot control strategy to avoid slipping altogether by altering the forces exerted by the robot actuation. The proposed estimator is validated experimentally using real-world materials. The estimator is shown to be able to determine the coefficient of friction of the materials ahead of the applied force ratio exceeding the coefficient of friction. Finally a 6-link bipedal robot is modelled and a slip prevention algorithm based on the estimator is implemented to illustrate how a bipedal robot would be able to avoid slipping while walking on a surface of unknown coefficient of friction. The bipedal gait is achieved using a previously developed control algorithm and walks on two flat surfaces of unknown coefficients of friction under ideal simulated conditions. The resulting simulations are compared to demonstrate the effectiveness of the proposed approach. The bipedal robot is shown to be able to walk on a surface of coefficient of friction of 0.4 without slipping and without prior knowledge of this value.

(далее…)

Категория: Научные статьи | Нет комментариев »