Backstepping fuzzy-neural-network control design for hybrid maglev transportation system

Area of Science:

• Control Systems Engineering
• Artificial Intelligence
• Transportation Technology

Background:

• Hybrid magnetic levitation (maglev) systems require advanced control for stable balancing and precise positioning.
• Traditional backstepping control (BSC) faces challenges with complex transformations and control signal chatter.
• Uncertainties in system dynamics necessitate robust control strategies for reliable operation.

Purpose of the Study:

• To design a backstepping fuzzy-neural-network control (BFNNC) for hybrid maglev transportation systems.
• To address the limitations of conventional BSC, including complex control transformations and chattering.
• To ensure system stability and reliable performance despite uncertainties, without strict system information.

Main Methods:

• A dynamic model of the hybrid maglev system, incorporating levitated electromagnets and a linear induction motor, was constructed.
• A fuzzy neural network (FNN) was employed as the core control component, mimicking BSC strategies.
• Adaptation laws based on projection algorithms and Lyapunov stability theorem were derived for network parameter convergence.

Main Results:

• The proposed BFNNC scheme demonstrated effective online levitated balancing and propulsive positioning.
• Experimental results verified the control strategy's effectiveness for the hybrid maglev transportation system.
• The BFNNC scheme showed superiority compared to BSC and backstepping particle-swarm-optimization control.

Conclusions:

• The BFNNC offers a robust and effective control solution for hybrid maglev transportation systems.
• This approach mitigates issues like control chattering and reduces reliance on precise system models.
• The study validates the BFNNC's practical applicability and enhanced performance through experimental evidence.