A Review on System Architectures for Sensor Fusion Applications

Abstract

Abstract. In the literature there exist many proposed architectures for sensor fusion applications. This paper briefly reviews some of the most common approaches, i. e., the JDL fusion architecture, the Waterfall model, the Intelligence cycle, the Boyd loop, the LAAS architecture, the Omnibus model, Mr. Fusion, the DFuse framework, and the Time-Triggered Sensor Fusion Model, and categorizes them into abstract models, generic and rigid architectures. While an abstract model does not guide the designer in the concrete implementation, the generic architectures provide a generic design but leave open several design decisions regarding operating system, hardware, communication system, or database system. Rigid architectures specify at least some of these aspects and therefore provide existing hardware designs, tools, and source code at the cost of flexibility.

Data fusion architectures: A survey and comparison

“…This architecture was introduced by LAAS (Laboratory Analysis Architecture Systems) in 1998 [20][21][22]. It consisted of the following levels:…”

Emulating the perceptual system of the brain for the purpose of sensor fusion

“…There does not yet exist a model for sensor fusion that is generally accepted. Many researchers even point out that it is very unlikely that one technique or architecture will provide a uniformly, superior solution [9]. In [21], it is pointed out that it generally accepted that sensor fusion in the perceptual system of the human brain is of far superior quality than sensor fusion achieved with existing mathematical methods.…”

Neurosymbolic Alerting Rules

“…However, future applications will target to "understand" the human users of a building and thus make it a safer, more secure, more comfortable, and more (energy-)efficient place [1], [2]. To do so, buildings will have to be equipped with a large number of diverse sensors, whose information has to be merged in order to get a robust representation of the environment, as studied in sensor fusion [3]. However, this research field is still quite recent and existing approaches are challenged by this abundance of data and the ways in which it should be analyzed and responded to [4], [5].…”