ELSA: A MULTISENSOR INTEGRATION ARCHITECTURE FOR INDUSTRIAL GRADING TASKS

摘要: Abstract This paper presents the topology of Extended Logical Sensor Architecture (ELSA) for multisensor integration and methodology constructing industrial sensor systems based on this architecture. ELSA has been developed applications, particularly, on-line grading classification non-uniform food products. It addresses a number issues specific to inspection. The system must be modular scalable accommodate new processes changing customer demands. easy understand so that non-expert users can construct, modify, maintain system. Furthermore, data representation scheme which allows quantification product deviations from an ideal model is required. To address these needs, sensors are encapsulated by logical model, providing robustness flexibility. construction upon object represents classifications through combinations primary features weighted fuzzy membership functions. guide selection processing routines; determine rulebase used inference engine process decisions. Although inspection focus work, it intended applicable variety automation tasks may benefit multiple perception