T4A.2Precise Local Positioning Radarwith implemented Extended Kalman filter

摘要

In this paper a novel solution for the tracking of arnmoving object based on Kalman-filtering is presented. The radarrnsystem applied measures the range between transponders and thernradar base station similar to a frequency modulated continuousrnwave radar with active reflectors by comparing the instantaneousrnfrequency of transmitted and received signal. It also provides thernrange rate (radial velocity) of the base station to the respectiverntransponder. To estimate a three-dimensional position from a setrnof range measurements a Trilateration of minimum three rangesrncan be calculated. Since the measured ranges and range rates arerndistorted by noise sources ill conditioned angles of cut between thernintersecting circular arcs lead to distorted position estimates.rnFurthermore multi path distorted range measurements can hardlyrnbe recognized and thus lead to heavy distortion of the estimate. Arnmodel-based filter allows taking past measurements into account.rnThis reduces the error of the estimated position. The linearrnKalman Filter (KF) is a recursive formulation of a model-basedrnfilter. Using an extended version of the Kalman filter (EKF) thernobject position can be estimated directly from ranges and rangernrates once the filter is initialized. This paper describes an EKF forrnthe estimation of a three dimensional position from ranges andrnrange rates. A gating method to filter out multi path distortedrnmeasurements is proposed. The Local Positioning Radar wasrntested in a steelwork environment and the obtained resultsrn(accuracy, deviation, repetition rate) are presented.