Edson Hiroshi Aoki

On the “near-universal proxy” argument for theoretical justification of information-driven sensor management

In sensor management applications, sometimes it may be difficult to find a goal function that meaningfully represents the desired qualities of the estimate, such as when we do not have a clear performance metric or when the computation cost of the goal function is prohibitive. An alternative is to use goal functions based on information theory, such as the Rényi divergence (also called α-divergence). One strong argument in favor of information-driven sensor management is that the Rényi divergence is a “near-universal” proxy for arbitrary task-driven risk functions, implying that these could be replaced by a Rényi divergence-based criterion, and this would usually result in satisfactory performance. In this paper, we present a rebuttal to that argument, which implies that finding theoretical justification for information-driven sensor management still seems to be an open problem.

A general approach for altitude estimation and mitigation of slant range errors on target tracking using 2D radars

When target tracking using polar (azimuth and slant range only) measurements is performed, the most usual approach is to simply ignore slant range errors and perform target position estimation on a 2D plane. In reality, slant range errors are very significant and can seriously impair tracking. 3D target tracking can mitigate the effect of slant range errors, and, in some cases, even allow altitude to be estimated. This paper analyzes previous approaches on 3D target tracking using 2D radars and their drawbacks, and proposes efficient methods (based on EKF, UKF and AMM-EKF) that can be used on realistic scenarios, without significant increase on computational cost. A robust filter initialization technique is also proposed for these methods.

Distributed registration of a network of asynchronous sensors

Registration of multiple sensors through common targets of opportunity is an extensively studied problem. The majority of proposed methods for computationally efficient estimation of sensor biases considered only the case of synchronous sensors. The relatively recent EXX method, however, allows exact estimation (under certain conditions) of sensor biases of asynchronous sensors. Unfortunately, the EXX method requires all measurements (or pseudomea-surements) originated by the targets of opportunity, which implies in high communication costs for large networks of sensors. In this paper, we formulate an extension of the EXX method that can be used for distributed bias estimation, i.e. obtains exact joint bias estimates for the entire network of sensors from joint bias estimates from subsets of these sensors. The proposed method can also be hierarchized in any manner, and can work with dissimilar sensors and different forms of sensor biases, thus being highly suited for todays demands of distributed data fusion.