Research And Implementation Of A Crowdsensing-based Noise Map Platform

Smart phones have been increasingly applied to citizens’ daily life playing roles as a bridge of communication and a mental partner. Nowadays, smart phones are capable of playing another role as a “smart mobile sensing device” due to the flexible abilities of network communication, plentiful sensors, and powerful capabilities of local computing and storage. Based on this, participatory sensing or crowdsensing has been proposed by researchers that encourages the public to make contribution to sensing tasks of measuring environmental noise pollution using their own smart phones in purpose of large-scale realization of the noise map in cities with cost-effective and fast-update manners. In this dissertation, the design and implementation of the noise map platform are presented and discussed.Firstly, the practical demands for functionalities of measuring, storing, visualization, and data analysis are researched and presented on the basis of crowdsensing-based noise map platform. Furthermore, the clients for Android-based and Web-based, and the back-end services are designed, implemented, and introduced. The Android client allows users to gather sensor data using customized measuring frequency, time, and flexible data uploading approaches, as well as provides the visualization of sensor data based on the heat map. Web client provides methods of data management and presentation of scientific knowledge regarding to noise pollution. Moreover, speculations of user contexts and related sensor data evaluation are realized in the server in purpose of inferring users’ uncertain behaviors during sensing tasks. Several sensors(i.e. accelerate sensor, GPS sensor, gyroscope sensor, and proximity sensor) are applied to deduce the conditions of smart phones in order to label scores on crowdsensing-based sensor data for further cleaning and filtering works facilitating the visualization of noise map. Meanwhile, effective grading work enables to be regarded as a criteria for incentive mechanisms to encourage more citizens to contribute to their communities and cities through participating in environmental monitoring. In the last, the correctness and usability of this approaches are presented by the implementation of particular experiments.To summarize, this dissertation designed and realized a crowdsensing-based noise map platform. Moreover, classification of user contexts and evaluation of sensor data measured by users are proposed to evaluate this kind of crowdsensing-based data. Overall, the contribution of this research work proposed has theoretical and practical significance in terms of the research and development of related innovative crowdsensing-based applications through applying the great power of the public and smart phones.