Adaptive hybrid control for linear piezoelectric ceramic motor drive using diagonal recurrent CMAC network

Area of Science:

• Robotics and Control Systems
• Artificial Intelligence
• Materials Science

Background:

• Linear piezoelectric ceramic motors (LPCMs) exhibit complex nonlinear and time-varying dynamics.
• Precise position control of LPCMs is challenging due to these inherent characteristics.
• Existing control systems may not adequately address the dynamic uncertainties in LPCMs.

Purpose of the Study:

• To develop an adaptive hybrid control system for high-precision position control of LPCMs.
• To introduce a novel control architecture based on a diagonal recurrent cerebellar-model-articulation-computer (DRCMAC) network.
• To validate the effectiveness of the proposed system against uncertainties and compare it with traditional control methods.

Main Methods:

• Designed an adaptive hybrid control system incorporating a DRCMAC network.
• Modified the cerebellar-model-articulation-computer (CMAC) network by integrating diagonal recurrent neural network (DRNN) concepts.
• Implemented a two-part control strategy: a DRCMAC controller for unknown dynamics and a compensated controller for approximation errors.
• Utilized a two-inductance two-capacitance (LLCC) resonant inverter for driving the LPCM.

Main Results:

• The DRCMAC network effectively captured system dynamics, converting a static CMAC into a dynamic one.
• The hybrid control system demonstrated robust performance in hardware experiments despite uncertainties.
• The proposed control scheme showed advantages over traditional integral-proportional (IP) position control systems.

Conclusions:

• The proposed adaptive hybrid control system, utilizing DRCMAC, achieves high-precision position control for LPCMs.
• The integration of DRNN concepts enhances the control system's ability to handle nonlinear and time-varying dynamics.
• The system offers a superior alternative to conventional control methods for LPCM applications.