LRAP

Abstract

In contemporary context aware systems when an event is triggered, system responds through sense-decide-actuate cycle and carries out requisite task with corresponding processing expenditure. However, such systems are lacking the capability to establish learned association among subsequent events, furthermore, the notion of the context is embedded into applications that may pose certain processing overhead. Therefore, each event is considered as arrival of afresh situation and dealt with entirely replicated processing cycle. Such computing mechanism where each event is stimulus to complete resource utilization leads the system towards processing overwhelming. This research work proposes a LRAP: a Learned Reflex action embedded Associative context learning based processing efficient Paradigm in visual sensor networks. In which each actuation of the system serves as new context to succeeding event and aids it to evolve internally. Gradually, with exposition to multiple events system refines its context repository with introspective context extracted through processed retrospective context that serves as meta-context to upcoming events leading the system towards evolution of context addition. Such context learning through improved introspective context utilization maximizes the system internal actuation that further optimizes the independent functions of reduced sensing and improved decision with minimal resource exploitation evolving it as a cut-through processing mechanism. Furthermore, when system gains the maximum internal actuation it responds impulsively against repetition of an event through intro-spectively evolved actuation based associative learning that imitates learned reflex action through associative context learning leading the system towards exceedingly processing efficient paradigm.