FPGA-Based Hybrid Neuro-FCM and ANFIS Adaptive Controller for High-Gain POSLRLC Converters in EV Charging Networks

This paper presents a real-time hybrid intelligent control framework for a Positive Output Super-Lift Re-Lift Luo Converter (POSLRLC), developed to enhance voltage regulation, efficiency and power quality in electric vehicle (EV) charging networks. The proposed architecture integrates a Neuro-Fuzzy C-Means (Neuro-FCM) clustering mechanism with an Adaptive Neuro-Fuzzy Inference System (ANFIS) for dynamic control optimization. The Neuro-FCM module continuously performs online fuzzy clustering on the error [E(k)] and change of error [[Formula: see text](k)], generating adaptive Gaussian membership functions, while the ANFIS layer refines rule consequents using hybrid learning combining least-squares estimation and gradient descent. This dual-layer adaptation provides self-tuning control action without manual gain adjustment, overcoming the limitations of static fuzzy and PI-type controllers. The proposed model was implemented and verified on a Xilinx Zynq UltraScale [Formula: see text] MPSoC platform under a bare-metal C environment, with real-time fixed-point inference executed on the FPGA fabric. Simulation and hardware validation were conducted under varying PV input (10–20[Formula: see text]V) and dynamic load (10–50 [Formula: see text]) conditions. Results demonstrate significant performance improvements compared to conventional PID, MPC, SMC and AGA-FLC controllers. The proposed system achieves a rise time of 6.7[Formula: see text]ms, settling time of 7.2[Formula: see text]ms and steady-state error of 0.12%, with converter efficiency maintained at 96.7–97.2% and total harmonic distortion (THD) reduced to 2.48%. The FPGA implementation yields sub-microsecond computation latency with [Formula: see text]50[Formula: see text]ns jitter, enabling deterministic high-speed control. Overall, the hybrid Neuro-FCM [Formula: see text] ANFIS control system demonstrates superior dynamic response, efficiency, stability and harmonic suppression, offering a scalable and intelligent solution for next-generation EV power conversion and renewable energy integration.