Performance of ANN and fuzzy logic in rounded shape platform with ultrasonic range finder in DWMR

Autonomous navigation of a mobile robot usually includes path planning and path tracking so that it can move toward the target in an unknown environment. Moreover, autonomous navigation is mainly relied on the sensors attached on the robot to obtain the information of its surrounding. In real time application, the sensor readings are not very consistent, which will mislead the mobile robot from its goal. This research study is mainly focused on increasing the accuracy of the mobile robot's decision although there are some abnormalities and inconsistent in the sensor readings. This objective will be achieved with the application of artificial neural network (ANN) and fuzzy logic (FL) on the sensors readings, and then the performances of these two algorithms are compared. In order to test the algorithms, a differential drive wheeled mobile robot (DWMR) platform with five ultrasonic range finders is designed. The experimental results obtained show that the ANN algorithm works better in term of increasing the action accuracy while the FL algorithm has the advantage of flexibility in term of output. Therefore, a better result can be achieved by combining the ANN for the decision making process and FL to control the output.