A conditioned reflex action embedded associative context learning-based energy efficient paradigm in smart city milieu

An intelligent video surveillance system is crucial to enhance public safety, crime prevention, traffic, and crowd management in a smart city milieu. Situational awareness is an essential aspect of these surveillance systems and it is inferred through underlying context aware frameworks. However, these systems may not possess the ability to proactively disseminate the real-time context among its sensor nodes. Moreover, in the specific conditions of occurrence of related or repeated events, these systems may also perform inefficiently through afresh context processing and disseminate cycles, without learning from the relevant context that has already been occurred and processed by the system. It leads to deteriorated performance, especially delay in reaction, overwhelmed processing, and energy expenditures. Therefore, to counter such issues, this research work proposes an energy efficient situational aware framework deployed in visual sensors network that is incorporated with context associative learning. System observes currently occurring context at each instance of an event. Overtime, context is refined and stored in context database. Such mechanism empowers the system to learn from previous experiences and develop relationship among the subsequent events that is embedded through this associative (adaptive) learning. Eventually, each event is processed through intelligent resource allocation, supported through mechanism of context learning that further illustrates the independent functions of reduced processing and improved (rapid) decision making resulting in evolution of energy efficient computing paradigm. Ultimately, the capability of learned reflex-action is induced through introspectively evolved context of the system in entirety and against specific condition of recurred situation depicting minimum energy expenditure.