Observer design in systems with discrete event output and applications to association memory.

摘要

from user data is in great demand in a lot of fields today. An algorithm or method is needed to be able to get the profile of a user online and realtime. Trying to solve the problem, we study the biological structure and psychological behavior of human cognitive system. Parameter estimation is the key but estimation in this domain is challenging. In order to allow the parameters to be tracked automatically, we develop a new technique using sliding mode control observer design. Our approach is to trace the output switching time, model internal dynamics in a persistent way while driving the estimation error to approach zero. The performance of the proposed observer has robust reaction to disturbance. It is easy and efficient to implement since the information it requires is easy to access. It can facilitate solving profile identification problems in a variety of domains. Using sliding mode observer, it makes possible the construction of a system that observes closely human interest individually. Its unique cognition oriented paradigm, proven theories, and supporting architecture allows for continuously adaptive monitoring and updating the current information.