Electronic implementation of a neural observer in FPGA technology application to the control of electric vehicle

This paper presents a new integration approach for the control of an induction machine in the FPGA technology. The proposal is based on the design of an adaptive observer that makes it possible to generate the induction machine parameters in real time. The proposed observer is based on the use of neural networks whose their generalisation capacity allows the compensation of the induction machine parameter variations. Considering the enormous algorithmic resources requested by the neural network integration, an optimised architecture of the observer is proposed using CORDIC algorithm. It makes it possible to show the best performances in time and area on the technology SPARTAN of XILINX.