Abstract: [Objective] In order to achieve precise control of clamping force, a fuzzy PID control algorithm is proposed for horizontal translational joints of the climbing robot.[Method] Firstly, a Butterworth two-order low pass filter is designed to filter the output of the force sensor for weakening the influence of the interference on the force sensor. Then, the error and the variation rate of force are input to the system, and fuzzification is done by choosing the trigonometric function as the membership function. Followly, the fuzzy inference rules of PID are established according to the PID rules, and defuzzication of output is achieved using coefficient weighted average method. Finally, the effect of the force control test of the fuzzy PID algorithm and the PID coupling algorithm is analyzed comparatively.[Result] In the contact, the fluctuation of force is weakened obviously by the filter. In the clamping, comparing to the coupling PD algorithm, the maximum precision is increased by 51.8%, the maximum force error between the relative joints is improved by 99.8%, and the time of dynamic adjustment is also optimized in different degrees. In addition, the overshoot of the clamping force is eliminated as well.[Conclusion] The filter can effectively filter the interference in the signals of the force sensors, and improve the reliability of the contact detection. The effectiveness of fuzzy PID clamping force control algorithm is also verified better than that of the coupling PD clamping force control algorithm, which can not only improve the reliability of the climbing robot, but also guarantee the safety of the wooden column.

Key words: climbing robot, wooden column, fuzzy PID, force control