Direct adaptive control using self recurrent wavelet neural network via adaptive learning rates for stable path tracking of mobile robots

This paper proposes a direct adaptive control method for stable path tracking of mobile robots using self recurrent wavelet neural network (SRWNN). As the proposed SRWNN is a modified model of the wavelet neural network (WNN), the SRWNN includes the basic ability of the WNN such as fast convergence. Besides the SRWNN has a property, unlike the WNN, that the SRWNN can store the past information of the network because a mother wavelet layer of the SRWNN is composed of self-feedback neurons. Accordingly, the SRWNN can easily cope with the unexpected change of the system. For the control problem, two SRWNNs are used as each direct adaptive controller for generating two control inputs, the translational and rotational displacement of the mobile robot. Specially, the gradient-descent method with adaptive learning rates (ALRs) is applied to train the parameters of the SRWNN controllers. The ALRs are derived from the discrete Lyapunov stability theorem out of consideration for the model of mobile robots, which are used to guarantee the stable path tracking of mobile robots. Finally, through computer simulations, we demonstrate the effectiveness and stability of the proposed controller.