An advanced framework for mobile robot trajectory tracking controller via recurrent general Type-2 fuzzy neural networks

This paper presents an advanced framework for mobile robot trajectory tracking based on a recurrent general type-2 Takagi-Sugeno-Kang fuzzy neural network (RGT2TSKFNN). The proposed framework addresses limitations of existing methods by enhancing adaptability, robustness, and precision under uncertain conditions. The trajectory tracking controller comprises two components: an adaptive kinematic controller that employs a cost function and specific constraints to guide the mobile robot along a desired reference trajectory, and a dynamic controller constructed using the novel RGT2TSKFNN architecture. Both the network structure and parameters are updated online, with parameter adaptation governed by a Lyapunov-based learning law to guarantee stability. The efficacy of the proposed approach is demonstrated through comprehensive simulations and real-world experiments under uncertainty, highlighting its ability to achieve accurate and reliable trajectory tracking. Comparative results indicate that the proposed controller outperforms existing methods, yielding lower performance indices such as mean absolute error, integral of absolute error, and integral of squared error across both simulations and experiments.