Scalable data and sensor fusion via multiple agent hybrid systems

From MaRDI portal





The aim of the paper is to investigate the problem of synchronizing estimates of plant states, and to establish for this purpose a new framework called MAHEA (Multiple Agent Hybrid Estimation Architecture). MAHEA is designed to model fusion of heterogeneous sensor data, to model coordination of heterogeneous sensors for distributed plants, and it is based on the authors' previous framework MAHCA (Multiple Agent Hybrid Control Architecture), oriented to extracting digital control programs for a variety of processes. The MAHEA framework allows heterogeneous data to flow between individual agents in the network to improve their individual estimates of the current plant state. Each agent of MAHEA formulates a relaxed variational optimization problem whose successful resolution produces an estimate of the plant state, and operates as a real-time theorem prover in the domain of relaxed variational theory. The MAHEA architecture is based on the construction of estimation Lagrangians on a manifold designed for the problem at hand. In the dual Hamilton-Jacobi formulation, the problem is a relaxed dynamic problem on the manifold for computing geodesic fields. A variety of differential geometric techniques was developed for solving such problems using connections, covariant derivatives, and Cartan derivatives associated with the Hessian form for the corresponding Lagrangian. A basic mathematical model for the sensor fusion is provided. MAHEA agent network proposes an efficient mechanism for state estimation from sensor readings that is extensible, robust, scaleable, allows cross-checking, and supports heterogeneous sensing of heterogeneous processes. For the entire collection see [Zbl 0909.00062].











This page was built for publication: Scalable data and sensor fusion via multiple agent hybrid systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4246816)