Estimation of 3D Device Position by Analyzing Ultrasonic Reflection Signals Esko Dijk 1,2 , Kees van Berkel 1,2 , Ronald Aarts 2 , Evert van Loenen 2 1 Eindhoven University of Technology, Postbus 513, 5600 MB Eindhoven, The Netherlands E-mail: esko@ieee.org 2 Philips Research Laboratories Eindhoven, Prof. Holstlaan 4, 5656 AA Eindhoven, The Netherlands E-mail: evert.van.loenen@philips.com Abstract— In future domestic context aware applications the location of mobile devices is often required. Ultrasound technology enables high resolution indoor position measure- ments. A disadvantage of state-of-the-art ultrasonic systems is that several base stations are required to estimate 3D posi- tion. Since fewer base stations would lead to lower cost and easier setup, a novel method is presented that requires just one base station. The method uses information from acous- tic reflections in a room, and estimates 3D positions using an acoustic room-model. The method has been implemented, and verified within an empty room. It can be concluded that ultrasonic reflection data provides useful clues about the 3D position of a device. Keywords— Location awareness, ultrasonic location sys- tems, signal processing I. I NTRODUCTION In future computing systems and devices, context aware- ness will play an increasingly important role. A device or system that is context aware [4] makes use of information, that characterizes the context or physical situation that the device or its user is currently in. Often, the physical lo- cation of mobile devices is important context information. In such cases the term location aware systems can be used. For example, a user carrying a context-aware museum au- dio guide could be informed automatically about the ex- hibits at the current location. Within the PHENOM project [14], several application scenarios were developed that bring location awareness into the area of domestic consumer electronics. The goal of location awareness in these applications can be func- tional, e.g. to improve the ease-of-use of consumer de- vices. But it may also enable new applications and ex- periences that are attractive to users. An example is the PHENOM portable screen for photo-browsing, that detects nearby displays, using them to display content. Location-aware applications may need either absolute or relative location information. Some of our application scenarios require the absolute 3D position of devices within a room. The required position accuracy (typically ≤ 1 m) can not be delivered by wide-area systems like GPS. Therefore, a specialized indoor location system is required. Such systems exist [8] for several context-aware applica- tions. These systems may use radio waves (RF), magnetic fields, ultrasonic waves, or combinations thereof. We fo- cus on ultrasonic location systems, because of their proven track record in low cost accurate indoor position estima- tion. Existing state-of-the-art ultrasonic location systems cal- culate a set of distances, using ultrasound time-of-flight measurements between fixed base stations (BSs) and a mo- bile device (MD). They subsequently use trilateration al- gorithms [8, 11] to calculate a 2D or 3D position of the MD. Existing systems are e.g. the Bat [1], Constellation and others from InterSense [7], Cricket [12], and the sys- tem by Randell and Muller [13]. A disadvantage of all these systems is that several units of infrastructure are re- quired at fixed known positions in a room, e.g. attached to the ceiling. Generally four BSs are required in a non- collinear setup to estimate 3D position of MDs. In special cases like ceiling-mounted BSs, three is sufficient. The required infrastructure and installation effort make these systems unsuitable for domestic deployment. Impor- tant requirements in the domestic domain are that a loca- tion system should be robust, safe, easy to install, minimal in its infrastructure, and low cost. These requirements led to our current research direction of a single base station positioning system. A single BS unit is the minimum of infrastructure (apart from no infrastructure at all), is easier to install than multiple units, and lowers system cost if BS units are mass-produced. Two methods were developed to realise such a single-BS system. The first method, pre- sented in this paper, uses ultrasonic reflections in a room for estimation of 3D positions of devices. The second method [6] employs an acoustic array within the BS, that