Developing a Framework for Image-based Integrity

摘要

The concept of navigation system integrity has been well- developed for the Global Positioning System(GPS).This concept and associated integrity monitoring processes fo- cus on identifying,and ideally excluding,faulty measure- ments before they can corrupt the GPS position solution. The definition of an assured protection limit is also a GPS integrity requirement.It is an unfortunate fact that GPS signal availability cannot be guaranteed under all environ- mental or operational conditions.Image-based navigation has been shown to be a potential alternative to navigation in the absence of GPS.However,the concept of integrity in an image-based context has not been fully explored. rnImage-based measurements from a single camera are pro- vided in pixel pairs for each target tracked,vice a single pseudorange measurement for each observed satellite,re- sulting in a different dimensionality in the measurement model equations.In addition,an image-based system is not always supported by a fixed constellation of measurement sources(e.g.available satellites),so there is the potential for a greatly varying number of overall measurements and possible measurement source geometries in a vision-based system.While an increase in the number of redundant mea- surements may aid integrity monitoring,it also may pro- duce a higher probability that a measurement is faulty,in- creasing the importance of having a robust monitoring pro- cess. rnThis paper investigates the extension of fundamental GPS autonomous integrity monitoring methodologies to develop a framework for integrity in image-based navigation.A re- lationship is derived between the horizontal position error and an associated test statistic in order to predict the least observable error source with the greatest influence on po- sition error.Results are presented using a minimal subset of measurements in different geometries to demonstrate the influence of the two-dimensional image measurement error on the position error.