The Novel Method for Calculating Control Coefficients of Legless Piezo Capsubot

In this paper, a new method is presented to calculate control coefficients online for a legless piezo capsubot. The capsubot robot is a nonlinear underactuated system. This robot consists of three parts: internal mass, shell mass and piezoelectric actuator. The motion of the system is generated by friction that acts between the internal mass and the shell. Here, to control the internal mass, a partial feedback linearization controller is used in which control coefficients are calculated online based on a function of the system states. This function is created from a database of genetic algorithms. To evaluate the efficiency of the proposed control method, it has been compared with two methods of optimal control and MPC. In the proposed method, energy consumption was 49% lower than that of the MPC method and 23% more than the optimal control method. This is while its error is 60% and 450% less than the optimal control and MPC method, respectively.