A distributed data fusion approach for mobile ad hoc networks

Abstract

This paper establishes a distributed data fusion method for ad hoc networks, enabling each device or agent to operate autonomously and collaboratively. In such a network, no infrastructure exists for centralized processing, and a lack offixed network membership creates ambiguous data fusion and reasoning architectures, communications patterns, and communications timing. The resulting data fusion method is shown to be consistent with other methods, including known architectures with closed-form solutions. The paper also applies the method to a short analysis of distributed data fusion with non-deterministic communications duty cycle where existing methods are impractical.

Fast Consistent Chernoff Fusion of Gaussian Mixtures for Ad Hoc Sensor Networks

“…'Optimal' exact Bayesian DDF thus requires explicit tracking and removal of the common information pdf from the product of and . However, this requires heavy computational expense in networks without tree communication topologies [1], [2], [4]. As such, exact Bayesian DDF is generally infeasible in dynamic ad hoc sensor networks.…”

“…A key practical issue for DDF is the avoidance of rumor propagation, which arises from double-counting of common information across multiple sensor nodes. Exact Bayesian data fusion techniques can be applied to track common information in sensor networks with static hierarchical and tree-based communication topologies [1], [2]. However, such methods are much more computationally expensive and infeasible to implement with dynamic ad hoc communication topologies.…”

Analytical and Computational Evaluation of Scalable Distributed Fusion Algorithms

“…The theoretic fundamentals of distributed information fusion are well documented and have been studied in depth [7][8][9][10][11]. It is noted, however, that practical applications of these theoretical results to nondeterministic information flow have remained a challenge.…”

“…The difficulty is due to the need to recognize correlated information resulting from past fusion events and know the values of their data sets. The information graph (IG) technique presented in [7][8][9] provides an analytical tool for identifying duplicate information in distributed estimation systems. The approach is a symbolic representation of the collection, propagation, and fusing of data among a set of fusion agents.…”

Agent-based information fusion

“…Martin and Chang [20] developed a tree based distributed data fusion method for ad hoc networks, where a collection of agents share and fuse data in an ad hoc manner for estimation and decision making.…”