Simulation Design of an Intelligent System for Automotive Transmission Gearbox Based on FPGA

In this paper, a simulation of artificial intelligent system has been realized and then implemented in FPGA (Field Programmable Gate Array). This intelligent system acts as a controller for electrical types of Automotive Transmission Gearbox ATG, which is used to control five steps (1,2,3,4,5) of speed ratios of a vehicle. The first speed ratio (step 1) is characterized by highest torque and lowest velocity, while the fifth speed ratio (step 5) is characterized by lowest torque and highest velocity. The Back-propagation neural network has been used as an intelligent system for this work. The proposed neural network involves eight neurons in the input layer, five neurons in the hidden layer, and five neurons in the output layer. In this work, accurate results have been obtained due to the following reasons: 1) The MATLAB linear activation functions, 'Satlins' and 'Satlin' have been used for the hidden and output layers respectively, 2) The MATLAB training function 'Trainlm' (Levenberg-Marqurdt training) has been used as a learning method for the proposed neural network, which is powerful algorithm. The Mean-Square-Error MSE of this work has reached to 3.94x 10-25 in 15 iterations, which is acceptable value with fast training process. After implementing the proposed system in FPGA, the execution time has reached to 10 nsec, which is also an acceptable value. The proposed simulation system has been designed using MATLAB software package, and has been implemented in FPGA using ISE Design Suit software package.