Development of fuzzy pid cascade control system for differential-drive wheeled mobile robot

Abstract

This study focuses on the control systems implementations of an intelligent autonomous mobile robot called Mobile Autonomous Robot (MAR). The robot is targeted for indoor usage, assume the role of a guide or security personnel in factories, warehouses, museums and offices. It has two driving wheels powered by two independent motors for drive and steering respectively. Each DC motor is controlled by a pulse width modulation (PWM) motor controller. A cascade control system has been implemented to control the movement of the robot, and this consists of a master controller and two slave controllers for each DC motor. The master controller is a Fuzzy Logic Controller (FLC) which computes the required speed and angular speed needed by the two motors. A vision system is used as the sensor for the robot motion. Fuzzy logic is applied to generate target trajectory movement with the information extracted from vision system such as the distance of target and the orientation of target. The landmark selected is a door because the door is a very common objects in indoor environments and the detection of a door allows a robot to do tasks such as navigation and path planning. This work presents an approach using computer vision which applies Hough Transform and Feature Matching technique after the image has been process using Canny edge detector. The two slave controllers are Proportional-Integral-Derivative (PID) controllers which ensure the desired speeds are obtained at the wheels. Two methods are used in calculating proper values of the PID. These are experimental method and model based method. The experimental method uses the Zigler-Nichols autotuning method without any knowledge about the plant to be controlled. In the model based method a mathematical model of the motors are first derived and this is used to design the PID controller. PID controller parameters were tuned according to four ranges of speeds. The PID controller parameters for both DC motors are auto-tuned using Gain Scheduling based on four ranges of speeds. Comparison was made to see the performance of the mobile robot using the PID parameters tuned by the two methods. In this study, fuzzy logic is also applied to generate obstacle avoidance techniques with the information extracted from ultrasonic sensors. Overall performance of the robot when moving to the target shows that the average relative error is less than 0.03 meter for distance and the average relative error for orientation is about 20o when arrived at the robot achieves its target. The average time taken by the robot to arrive at the target in 1 meter is approximately 23 seconds.