A survey on pervasive computing over context-aware system

Pervasive computing (PerC) systems are now being integrated into everyday life which is deployed in homes, offices, hospitals, universities. The sensor data can be integrated with different range of sources in pervasive systems which also offers an extensible, open and portfolio services. Due to remarkable progress in various domains such as smart phones, computing power, sensor, network and embedded devices, wireless communications which are combined with social networking paradigms, cloud computing and data mining techniques, that enabled the users for creating PerC systems with global accessibility. The major challenge of PerC is to provide the suitable consistent adaptive behaviors and context-aware systems in a vast amount of sensor data for these services which need to improve the accuracy, precision and dynamism. This research work provides an inclusive analysis of characteristics of data, then the complexities of the existing technique are reviewed which are mostly used in inferring situations from sensor data. The extensive experiments are carried out on benchmark dataset to validate the efficiency of existing techniques namely multi-context, mechanisms of user-side publish/subscribe by using the metrics such as accuracy, f-measure, precision, recall and communication overhead. Many open research opportunities are identified in this area by comparing and contrasting the existing techniques, which are discussed in this research work.