MINING USER GOALS FOR INDOOR LOCATION-BASED SERVICES WITH LOW ENERGY AND HIGH QoS

Location-based services (LBSs) play a very important role in pervasive computing environment, and QoS (quality of service) is one of the key evaluations for LBS. To maintain high QoS, the traditional approaches rely on accurate and continuous localization. However, the energy consumption of the mobile device under this situation is often too high for practical applications. Thus, it seems that the energy consumption and QoS become two conflicting factors in LBS systems. In this article, a new adaptive goal-aware computing framework (Adaware) is proposed to solve this contradiction. We show that the QoS of LBS can be evaluated by recognizing user goals. We design new algorithms to mine user goals from discontinuous location data to reduce the energy consumption while keeping a high QoS at the same time. More specifically, Adaware employs an accelerometer to implement motion-based localization, which greatly reduces the unnecessary energy consumption on Wi-Fi scanning compared to the original continuous localization methods. Then based on the estimated discontiguous critical point traces which have been postprocessed by our proposed Localization Confident Coefficient filter method, a novel N-gram goal inference algorithm is used to predict the accurate goal. The experimental results in real-world wireless network environments validate the effectiveness of our framework. We can get 80% QoS under 70% location estimation accuracy within 10 meters and 30% energy saving compared to continuous Wi-Fi scanning.