Data Fusion Architecture - An FPGA Implementation

Abstract

Architecture for multisensor data fusion based on adaptive Kalman filter is described. The architecture uses several sensors that measure same quantity and each is fed to Kalman filter. For each Kalman filter a correlation coefficient between the measured data and predicted output was used as an indication of the quality of the performance of the Kalman filter. Based on the values of the correlation coefficient an adjustment to the measurement noise covariance matrix (R) was made using fuzzy logic technique. Predicted outputs obtained from Kalman filters were fused together based on weighting coefficient, which was also obtained from the correlation coefficient. Results of fusing data of several sensors showed better results than using individual sensor. Matrix-matrix multiplication using FPGA also presented