Abstract : This paper proposes a rule-based self-tuning fuzzy controller based on neural networks. An online fuzzy inference algorithm is designed, and the fuzzy control rules are adjusted using neural networks. This algorithm is then applied to the control of an AC servo system. Simulation results show that the controller has fast response, strong robustness, and the system using this controller has good dynamic and static performance and anti-interference capability. Keywords : Neural network; Rule-based self-tuning fuzzy control; AC servo system 1 Introduction With the development of production and technology, AC servo systems have been increasingly widely used due to their simple structure and ease of maintenance. AC servo systems consist of AC motors, and their mathematical models are nonlinear, coupled, and time-varying. Traditional object-model-based control methods cannot achieve satisfactory results when controlling them. Introducing fuzzy control into the control of AC servo systems utilizes the characteristics of fuzzy control being independent of object models and having strong robustness. This can effectively overcome the influence of coupling, nonlinearity, and parameter variations in AC servo systems. The control performance of fuzzy control depends on its control rules, but control rules summarized from human experience often have certain limitations. Once the control rules of a conventional fuzzy controller are determined, they cannot be changed, which inevitably makes it difficult to adapt well to changes in the dynamic characteristics of the system or the influence of random disturbances, thus affecting the effectiveness of fuzzy control. Improving the accuracy and anti-interference capabilities of fuzzy controllers by endowing them with adaptive and self-organizing abilities while retaining their robustness can effectively enhance their performance. In recent years, scholars both domestically and internationally have conducted extensive research on fuzzy adaptive control, combining fuzzy control with various control techniques and optimization methods such as neural networks, evolutionary algorithms, sliding mode variable structure, and robust control. This has led to the proposal of many adaptive and self-organizing fuzzy learning control methods, some of which have been applied in industrial fields, including AC servo systems, with good results. This paper proposes a rule-based self-tuning fuzzy controller based on neural networks. It employs a real-time fuzzy inference method, utilizing neural networks to adjust the influence of system errors and error change rates on the output of the fuzzy controller online, thereby achieving the goal of adjusting the fuzzy control rules. This significantly improves the effectiveness of fuzzy control and enhances system performance. [b][align=center]For more details, please click: Rule-based Self-tuning Fuzzy Controller Based on Neural Networks and Its Application in AC Servo Systems[/align][/b]