Development framework for hardware implementation of digital neural network used in smart sensors [articol]

Abstract

This paper presents a development platform for the implementation of digital neural network used in Sensor System for an Artificial Hand. The use of neural networks to add learning and adaptive behavior to smart sensors is essential and the FPGA implementation is an easy an attractive way for hardware implementation. This platform was developed in order to provide a fast prototyping environment using reconfigurable devices (FPGA) and a microcontroller. The microcontroller is used to implement the Data Acquisition System and to adapt signal sensors to neural network input requirements. The reconfigurable device (XC2S50 Xilinx) is used to implement the neural networks and other logic blocks of the same application. The System Generator tool for Simulink allow the easy generation of hardware Description Language (HDL) code that can be synthesized for implementation in the Xilinx family of FPGA devices. The developed framework allows device communication with a PC in order, to perform the offline learning task or, to transfer data for analysis. Software is designed to manage the communication protocol with Matlab via parallel port.