Optimizing ZigBee for data streaming in body-area bio-feedback applications Marco Benocci – Elisabetta Farella – Luca Benini DEIS University of Bologna Bologna, Italy ሼ ȁ ȁ ሽ Laura Vanzago Subsystem Products Group ST Microelectronics Agrate Brianza (MI), Italy laura.vanzagost.com Abstract— Wireless Body-area Sensor Networks (WBSN) are a key component of e-Health solutions. Many different physiological parameters can be monitored and collected continuously by wearable wireless sensors with high accuracy, low power consumption and limited cost. In this work we focus on WBSN for real-time biofeedback applications. In this domain the real-time nature of I/O data streams is of critical importance for providing an useful and efficient sensorial feedback for the user while system lifetime should be maximized. Thus, bandwidth, throughput and energy efficiency of the communication protocol must be carefully optimized. The research community is still divided on the choice of the protocol to use (proprietary, ZigBee, Bluetooth). In this article we investigate how ZigBee meets WBSN requirements showing a higher communication efficiency and lower power consumption, with respect to Bluetooth Serial Port Profile, (SPP) based solution. Keywords: WBSN; streaming; ZigBee; accelerometer; balance. I. INTRODUCTION The recent advances in wireless communications and miniaturized sensors are revolutionizing the healthcare domain, by providing relatively low-cost and low power wearable solutions to enable applications of ubiquitous computing and embedded systems for diagnostic, monitoring, rehabilitation, and training purposes [1]. The particular domain of rehabilitation, where bio-feedback provisioning is required, demands for real-time performance, unobtrusiveness and adequate lifetime, while at the same time guaranteeing bandwidth and throughput to extract from sensors the features needed for the expected feedback. Thus, the choice of the wireless protocol is crucial. Several studies prefers proprietary solutions [2] which enable to better control protocol characteristics. However, the use of standard protocols is highly desirable to ensure inter-operability and improved integration into large-scale pervasive healthcare environments. Many recent works show the efficiency of both Bluetooth (BT) [4-6] and ZigBee [7-11] solutions in the wireless sensor domain and compare them to investigate their complementary and competitive characteristics [12- 15]. ZigBee typically is considered for environmental monitoring applications that require very low duty cycles and tenth/hundreds of nodes. The easy way in which BT can be connected to a PC or a PDA through the serial port, make it one of the preferred solution for WBSNs. However, recently, several new components have been announced to support seamless interfacing ZigBee wireless nodes with PCs or PDAs by means of ZigBee USB dongles supporting serial communication. Hence, we can consider ZigBee as a competitive solution in the WBSN arena, and explore its use also when data streaming is required. In this paper we initially compare our two BT and ZigBee end-nodes prototypes (based, respectively, on the Bluegiga iWrap stack BT v2.0 compliant and the Ember ZigBee stack, EmberZNet v3.0) to highlight the benefits of these protocols for WBSN applications. Then, we present our approach to enhance the real-time data streaming (throughput) while reducing power consumption within a ZigBee network. To achieve this objective we minimize the packet overhead and tune the internal Finite State Machine that manages the transition between the different states of the protocol. Our intention is to reduce the overall transmission time, thus both increasing throughput and reducing power consumption. An simple analytical expression is proposed to help designers chose the optimal power consumption-performance tradeoff point depending on application requirement. The reminder of this paper is organized as follows: background information on Bluetooth and ZigBee radio protocol are provided in section 2. WBSN streaming optimization and the performance vs. power analytical model are described in section 3. Section 4 describes results. Section 5 concludes the paper. II. LOW RATE PROTOCOL COMPARISON Several wireless networking standards have been proposed in recent years: BT, ZigBee, Wi-Fi and UWB are probably the best-known. Each standard reflects its own unique strengths making the point in terms of coverage area, throughput, network topology, security, power consumption. BT and ZigBee, in particular, are extensively used in low-data rate sensor networks for their low energy consumptions and low cost. In this work, we focus on short-range body area sensor networks (WBSNs). A. Coverage area Both BT and ZigBee are compatible with WPAN applications. The typical coverage area spans from 10m indoor to 100m outdoor. Further coverage can be obtained through multi-hop networks, which is not critical for WBSNs. B. Throughput Bluetooth and ZigBee provide lower data rate with respect to other IEEE standard (UWB and Wi-Fi). Authorized licensed use limited to: Universita degli Studi di Bologna. Downloaded on June 22,2010 at 14:23:27 UTC from IEEE Xplore. Restrictions apply.