Dynamic balancing humanoid robot using completmentary filter to optimized pid controller

Abstract

Balance control on humanoid robots is one of the means to create harmonic motion in a robot movement system to make it more stable and can reduce the frequency of robot crashes. To control the position of the robot, it can use force / torque-based sensors or moment-based sensors inertia. The moment inertia sensor is a sensor that reads the initial angle changes to the linear angle of the body sensor framework. The accuracy of sensors in balancing robots with a system looking for set-point values will never be separated from the process of refining data processing, one of which is a sensor value (complementary) and continuous control (PID) filter that is very efficient in controlling feedback effects. The optimal PID value for balancing in stationary conditions is Kp = 0.60, Ki = 0.07, Kd = 0.90, but in a running condition there is no need for a steady state PID value, because when the conditions are needed it requires a fast gain response to obtain tuning while walking is Kp = 1.20, Ki = 0.09, Kd = 1.50.