FPGA‐Based Realization of Intelligent Escalator Controller Using Artificial Neural Network

In this work, a proposed intelligent controller has been designed and implemented for a prototype of four stair‐step escalator. The required task of this controller is to manage the supplied power of driving motor of the escalator according to the applied load on the stair‐steps, which is represented by the number of persons standing on the stair‐steps. The proposed intelligent controller must realize the following objectives: adaptive consumption power, fast processing, high reliability, and low cost, which the contribution of this work. The backpropagation neural network (BPNN) was used in designing the software of the proposed controller for the following reasons: fast learning, fast processing, and capability of finding the optimal solution. The proposed neural network was designed using MATLAB software package; it involves three layers, they are input, hidden, and output layers; the input layer involves four neurons, while the hidden layer involves 10 neurons, and the output layer involves three neurons. After executing and testing the proposed controller, it was observed that the mean square error (MSE) value reached 8.68 × 10 −18 , the gradient reached 3.41 × 10 −9 , and there was fitting 100% between the desired and actual outputs, which indicates the high reliability and accuracy of the proposed controller. Finally, the software of the proposed controller was downloaded in field‐programmable gate array (FPGA) using ISE Design Suit software, as known, the main characteristics of FPGA are fast processing, small size, and low cost.