Passive, device-free recognition on your mobile phone: tools, features and a case study Stephan Sigg 1 , Mario Hock 2 , Markus Scholz 3 , Gerhard Troester 4 , Lars Wolf 1 , Yusheng Ji 5 , and Michael Beigl 3 1 TU Braunschweig, Braunschweig, Germany, [sigg,wolf]@ibr.cs.tu-bs.de, 2 Karlsruhe Institute of Technology, Karlsruhe, Germany mario.hock@student.kit.edu 3 Chair for Pervasive Computing Systems (TecO), KIT, Germany [scholz,michael]@teco.edu, 4 Electronics Laboratory, ETH Zurich, Switzerland troester@ife.ee.ethz.ch, 5 National Institute of Informatics, Tokyo, Japan kei@nii.ac.jp Abstract. We investigate the detection of activities and presence in the proximity of a mobile phone via the WiFi-RSSI at the phone. This is the first study to utilise RSSI in received packets at a mobile phone for the classification of activities. We discuss challenges that hinder the utili- sation of WiFi PHY-layer information, recapitulate lessons learned and describe the hardware and software employed. Also, we discuss features for activity recognition (AR) based on RSSI and present two case studies. We make available our implemented tools for AR based on RSSI. Key words: Activity recognition, Passive device-free recognition 1 Introduction In urban areas, a mobile device is constantly exposed to (possibly encrypted) communication. In contrast to traditional communication, where such data is considered as noise or congestion, we exploit implicit information on environ- mental situations carried by the signal strength fluctuation of overheard packets. Fluctuation on WiFi RSSI might indicate presence, the number of people around or even activities conducted in proximity (cf. figure 1). Localisation of objects and individuals [11] as well as the classification of activities [13] or crowd counting [15] has been considered recently. We distin- guish between systems utilising software-defined-radio (SDR) devices to sample the fluctuation in a received signal and systems leveraging information avail- able from received data packets. In the former case, due to the higher sampling frequency and additional information available, the recognition accuracy is gen- erally higher. However, these systems require specialised SDR devices. We recognise activities on a single device (a mobile phone) from packets broadcast by a transmitter not under our control (a WiFi access point (AP)).