A note on "book review tracking and data fusion: A handbook of algorithms" [Authors' reply]

Abstract

This book, which is the revised version of the 1995 text MULTITARGET-MULTISENSOR TRACKING: PRINCIPLES AND TECHNIQUES, at double the length, is the most comprehensive state of the art compilation of practical algorithms for the estimation of the states of targets in surveillance systems operating in a multitarget environment using data fusion. This problem is characterized by measurement origin uncertainty, typical for low observables. The tools for design of algorithms for the association of measurements and tracking are presented. Explicit consideration is given for measurements obtained from different sensors under realistic assumptions -lack of synchronicity and different detection and accuracy characteristics. Several real-data examples are given to illustrate the techniques discussed. The modeling accounts for target maneuvers, non-unity detection probability, false alarms, interference from other targets and the finite resolution capability of sensors. The problems of track initiation, maintenance and multisensor data fusion are considered. The optimization of certain signal processing parameters based on tracking performance is also discussed. The latest results on measurement extraction for unresolved targets, sensor management and data fusion are included.Many of these techniques have applications to state estimation when using multiple sensors in control systems, autonomous vehicle navigation, robotics and wireless communication. An extensive index is provided with all the indexed terms highlighted in the text for the convenience of the reader.

Radiolocation and tracking of automatic identification system signals for maritime situational awareness

“…The demonstrated average performance of the proposed radiolocation and tracking technique based on EKF confirms that the approach could be used to perform automatic validation of declared GNSS positions encoded in the broadcast AIS messages. This could be performed using a single step gate validation [27] or a more effective binary hypothesis testing over a moving time window [28]. The former approach may be sufficient if the EKF estimate is close to the true vessel position but is intrinsically prone to a large number of false positives.…”

Multiple Model Adaptive Nonlinear Filter for Range Estimation Using Bearing Only Tracking

“…EKF uses the Taylor's series approximation to linearize the nonlinear measurement model [29]. The predicted state estimate (̂| −1, ) and covariance ( | −1, ) is given by,…”

Partial observable update for subjective logic and its application for trust estimation

“…In essence, the true observations are hidden similar to states in a hidden Markov model [42]. The likelihood function is very natural in state estimation/track filtering [43] and in pattern classification [44]. It represents abductive reasoning where the partial observation that is measured x is the consequent, the full observation z is the antecedent, and z is inferred through the likelihoods f (x|z = i) that serve as the conditionals.…”