Control of Mobile Robot Using Prediction-based FNN

A prediction model-based fuzzy neural network (PFNN) approach is proposed, in which a basic FNN is created at first to predict the relative position of the trajectory. Then a FNN is used independently to get the control values of the variables for motor motion according to those variables including trajectory position both from those measured and predicted values, and those speed variables. At last membership functions and network weights of the second FNN are also trained with a BP algorithm. Meanwhile, the measured values of the trajectory are memorized so as to compare them with the memorized values to confirm if the motion is moving in cycles. If it is moving in cycles, a decision making unit would cease the prediction unit. The emulated experiments show that the performance of the proposed approach is higher, the process to train the network is relatively easy, and the control strategy is simple.

I. Introduction

One of the challenging tasks for mobile robot navigation is to ensure that the robot follows a certain trajectory and avoid any obstacles placed along the trajectory. The fuzzy logic-based approach was selected initially since fuzzy logic is able to provide human reasoning capabilities to deal with uncertainties about the navigation in which the robot's actions directly depend on the perception of the world by means of its sensors.