A Research Of Power Optimization On Mobile Embedded Device Based On Context Awareness

Recent years, embedded industry is developing rapidly, mobile handheld devices such as mobile phone and tablet PC are more and more popular. However, when feeling the positive experience brought by science and technology we also face problems like battery running out outdoors. At present most embedded devices’power is supplied by battery whose capacity is limited because of their size and weight. Therefore, the contradiction between the embedded devices’rapid update rate and battery industry’s slow technical revolution leads to that embedded low-power becomes one of the research hotspots.The research results of embedded low-power so far mainly focus on optimizing and improving hardware technology and software development process at home and abroad. This paper combines embedded low-power optimization with context awareness, propose the thought of mobile embedded power optimization on account of context awareness that choosing smart phone as the platform and adjusting the performance of the devices according to device context information perceived by sensors to decrease the power dissipation. This paper’s main work is as follows:1. Introduce and discuss the research status of the embedded low-power domain, expound the research background and significance of the embedded power optimization based on context awareness, and introduce related basic theory of power optimization.2. Aimed at the WiFi network environment in user’s context information, analyze the power consumption during network handover. Introduce the handover method in802.11protocol used now and analyze its shortcomings. Study the handover strategies in WiFi network raised by other researchers. Pick the fast handover algorithm as the basis and propose our power preference handover algorithm of WiFi network after optimizing.3. Aimed at the user behavior in context information, analyze the feature of several daily user behaviors and the user habits of embedded devices. Propose power consumption optimization strategy based on user behavior recognition to reduce the power consumption accordingly in different user behaviors. Classify user behavior using machine learning and apply relevant power optimization strategy to different classes.4. Complete the experiment of power preference handover in WiFi network and get the result that the power preference handover algorithm reduces power consumption during handover among WiFi access points. Complete the experiment of user behavior recognition. Get data from sensors using tools developed by ourselves and recognize user behavior using machine learning. The accuracy rate of the experiment reaches a high level. Apply relevant strategy to different behavior and achieve the target to bring down power dissipation. In the end, the design of the power optimization system on account of context-awareness is proposed.