Hardware acceleration design and implementation for inertial navigation applications

The paper describes a data preparation technique of inertial sensors for navigation purpose. Data preparation consists of calibrating and denoising low cost accelerometer used for pose determination of vehicles in case of low rated velocities and accelerations. A hardware acceleration technique is presented in this article using Xilinx ZYNQ SoC (System on Chip). The aim of the designed hardware is to accelerate the process of DSP (Digital Signal Processing) by means of XADC (Xilinx Analog to Digital Converter) acquisition and DRP (Dynamic Reconfiguration Port) interfacing between XADC and the PL (Programmable Logic) part of the ZYNQ. The filter has first been modeled and simulated on MATLAB then two IPs (Intellectual Properties) has been implemented using XSG (Xilinx System Generator). A statistical data analysis of the accelerometer noise has been performed before and after filtering. Simulation and experimental results of the denoising process applied for analog accelerometer are presented.