Buy4Me: A Delivery System via Mobility Prediction Based on Mobile Crowd Sensing

With the prevalent development of rich sensor-equipped smart phones, mobile crowd sensing becomes a new sensing paradigm beyond traditional sensor networks. In this paper we propose and develop Buy4Me, a delivery system based on mobile crowd sensing. It can be used in different scales, such as a university campus or a city, to make shopping and delivering more convenient. When you need something not near, Buy4Me can recommend a friend to bring it for you without extra work. Users' mobility histories are analyzed to predict where a user will visit, and whether she will encounter with the others. Then Buy4Me ranks friends by the probability of whether they completing the delivering task, and returns top-k friends. We evaluate the method with the dataset contributed by Microsoft Research Asia, and find that it can achieve acceptable recommendation accuracy.