[{"type":"paper-conference","title":"Partial Likelihood for Unbiased Extended Object Tracking","issued":{"date-parts":[["2015"]]},"page":"1022-1029","container-title":"Proceedings of the 18th International Conference on Information Fusion (Fusion 2015), 6-9 July 2015, Washington, DC, USA","author":[{"family":"Faion","given":"Florian"},{"family":"Zea","given":"A."},{"family":"Baum","given":"M."},{"family":"Hanebeck","given":"U. D."}],"publisher":"IEEE","publisher-place":"Piscataway (NJ)","ISBN":"978-0-9824-4386-6","abstract":"An extended object gives rise to several measurements that originate \r\nfrom unknown measurement sources on the object. In this paper, we \r\nconsider the tracking and parameter estimation of extended objects that \r\nare modeled as a curve in 2D such as a circle or an ellipse. A standard \r\nmodel for such extended objects is to assume that the unknown \r\nmeasurement sources are uniformly distributed on the curve. We argue \r\nthat the uniform distribution may not be the best choice in scenarios \r\nwhere the true distribution of the measurements significantly differs \r\nfrom a uniform distribution. Based on results from curve fitting and \r\nerrors-in-variables models, we develop a partial likelihood that ignores \r\nthe distribution of measurement sources and can be shown to outperform \r\nthe likelihood for a uniform distribution in these scenarios. If the \r\ntrue measurement sources are in fact uniformly distributed, our new \r\nlikelihood results in a slightly slower convergence but has the same \r\nasymptotic behavior.","kit-publication-id":"1000051021"}]