Design and Implementation of an On-body Placement-aware Smartphone Kaori Fujinami, Satoshi Kouchi, and Yuan Xue Department of Computer and Information Sciences Tokyo University of Agriculture and Technology Tokyo, Japan Email: fujinami@cc.tuat.ac.jp Abstract—A mobile phone is getting smarter by employing a sensor and awareness of various contexts about a user and the terminal itself. In this paper, we deal with five storing positions of a smartphone on our body as a context: 1) front pocket of trousers, 2) back pocket of trousers, 3) jacket pocket (side), 4) chest pocket, and 5) around the neck (hanging). We propose a method for identifying the five positions with 29 features that characterize specific movements of a terminal at the position during walking. The result of offline experiment shows that an overall accuracy was 72.3% in a strict condition where datasets for a test were obtained from different people whose datasets were utilized for training a classifier. We also present a working prototype system with software framework for Android OS. An event of positional change is delivered to applications that conforms to APIs as well as to existing applications via preference settings. As a proof-of- concept application, a placement-aware heatstroke alerter was developed that tells a user about possible over (under)-estimate of the potential risk based on a storing position. I. I NTRODUCTION Recent advancement of technologies such as Micro Elec- tro Mechanical Systems (MEMS), high performance and low power computation has allowed a mobile phone to be augmented with various sensors and to extract contex- tual information of a user, a device and/or environment [1][7][9][11]. These sensors are (or will) not only utilized for explicit usage of mobile phone’s functionalities like user authentication [18], display orientation change and backlight intensity control, but also for monitoring user’s activities [11], indoor location [1], the state of a device [7], [9], pedestrian identification [22], environment, society, and more [14]. In this paper, we focus on the position of a smartphone on the human body as a context. The position is not an exact 3D coordinate, but parts of our body or clothes such as “hanging from the neck” and “inside a chest pocket”. A recent study of phone carrying [4] shows that 17% of people determine the position of storing a smartphone based on contextual restrictions, e.g. no pocket in the T-shirt, too large phone size for a pants pocket, comfort for an ongoing activity. These factors are variable throughout the day, and thus smartphone users would change the locations in a day. This suggests that a context, on-body placement, has a great potential to provide services from various aspects: facilitating human-human communication, playing important roles in sensor-dependent services and in improving usability of the terminal, and contributing to reduce unnecessary energy consumption. Starting from a pioneering work of Kunze et al. [12][13], on-body position sensing is getting attention [6][21][23]. Vahdatpour et al. recently proposed a method to identify 6 regions on the body, e.g. head, upper arm, for health and medical monitoring systems [23], where a sensor was attached directly on the skin or on the clothes and the identification process were conducted in an offline manner. A preliminary work by Shi et al. seeks a method of on- body positioning of a mobile device, e.g. mobile phone, into typical containers such as a trousers’ pocket. In [6], we presented an early study of tracking the placement of a mobile phone in typical containers like Shi et al. The tracking allows a system to identify the position even in the change of a position while a person is standing still. However, the processing was actually done on a remote PC while data were collected from a Bluetooth-based ac- celerometer attached on a mock-up of a mobile phone. In this paper, we redesign the placement detection scheme in a stationary motion, e.g. walking, and implemented a working prototype on a commercial smartphone with an embedded accelerometer. Also, we provide a software framework for an Android platform that supports the development of on- body placement-aware applications. The rest of the paper is organized as follows: application scenarios are presented in section II to clarify the concept and the benefit of identifying device position on the body. Section III describes algorithm design and implementation. The performance of the algorithm is evaluated in section IV. Section V and VI show a software framework for application developers and an application for reliable heatstroke alert, respectively. Finally, Section VII concludes the paper. II. POSSIBLE APPLICATIONS To emphasize the relevance of placement-awareness of a smartphone (or a mobile device augmented with a sensor or an effector), possible applications are presented, which is classified into two categories: placement-aware functionality control and annotation for sensor readings. 2012 32nd International Conference on Distributed Computing Systems Workshops 1545-0678/12 $26.00 © 2012 IEEE DOI 10.1109/ICDCSW.2012.52 69