Design of Proportional-Integral-Derivative Controller Using Root Locus for Human Face Tracking System with Viola Jones Method

Human face tracking system is used in various fields, including smart home, security, industrial automation, biometrics, law enforcement, entertainment and social media. The main part of this human face tracking system is face detection using camera as sensor and servo motor control as actuator. Face detection with viola jones method is implemented to get bounding box coordinates for the detected face position. The servo motor control is used to drive the servo motor so that the detection results have a bounding box midpoint reference, coordinate for pan direction is 160 and coordinate for tilt direction is 120. This paper explains the design of Proportional-Integral-Derivative (PID) controller using root locus algorithm to determine the control action “reverse gain” to deal with negative effects of non-minimum phase systems on tracking system. From the motion of servo motor result, it can be inferred that PID controller using root locus can drive the servo motor to steady state conditions, reach the bounding box midpoint reference. It means, this system is successfully track the human face position with smaller overshoot (0.512%), faster settling time (1.087 ms) on pan direction and smaller overshoot (0.472%), faster settling time (1.030 ms) on tilt direction.