Middleware for crowd sensing systems

Abstract

Nowadays, mobile devices are loaded with many useful sensors and have high computational capabilities. The sensors available on these devices provide valuable data that can be harnessed in many ways. Due to abundance of mobile devices, it has become a challenge to identify optimal data providers. There have been many approaches already proposed to identify a group of mobile nodes that can serve a particular request like collecting accelerometer data or wifi strength, something that makes use of the sensor data from this group of mobile nodes. But, very little work has been done to identify an optimal node within that group. An optimal node can be defined as one that has the best score based on its history to serve requests and vital stats like battery level, data connection quality etc. In my project I address this problem and try to develop a middleware to achieve the same. I also harness the middleware's capabilities to develop real world applications using the data collected by optimal nodes. This semester I was primarily working on developing applications that use the data collected using crowd sensing through the nodes assigned by the middleware.