(Stable) virtual landmarks: Spatial dropbox to enhance retail experience

Landmarks are signatures of our surroundings which help us to uniquely identify a location. Recent studies show that like us humans, it may be possible by the sensors on mobile devices to identify landmarks [1], [2]. This can open up the possibility of a lot of applications in the domain of augmented reality, gaming, retail etc. However, to make such applications a reality, a particular landmark need to be stable across mobile phones, persons carrying the mobile phones etc. This paper specifically builds up a framework to discover such stable landmarks and demonstrates its utility in the development of next generation apps. In order to identify such virtual landmarks, we employ a clustering algorithm to perform non-intuitive feature combination of sensors like Accelerometer, Gyroscope, Magnetometer, Light, Sound, Wi-Fi, GSM signal strength etc. Further, we rigorously test the clusters to ensure that landmarks are stable across different devices, people, and time. According to our results, change in device affects the stability of a landmark most. Finally as a proof of concept, we develop a prototype system RetailGuide using landmarks to facilitate smart retail analytics cum recommendation service.