Exploring the Role of Agent and Multi-Agent in Data Fusion Systems

Data fusion methods reduce the effort of data processing systems and facilitate achieving optimum results for complex problems. Recently, various complex and distributed systems such as Internet of Things (IoT) systems are implemented where many sensors and data sources are allocated to collect heterogeneous data. These systems demand innovative and advanced data fusion methods. This paper presents an overview of agent and multi-agent roles in data fusion methods. Basically, agents provide features that enable complex and distributed systems to achieve deep data fusion in a collaborative multi-agent environment. Interaction among agents allows disseminating global overviews to a localized set of data fusion modules. There are different architectures to establish communication and authority among agents, such as (1) centralized, (2) decentralized, and (3) hybrid. Each of these approaches affords essential attributes of multi-agent systems (MAS). As a result, agents have the ability to improve communication and decision-making among data fusion modules and ensure producing better features or values during the fusion processes.