3146 IEICE TRANS. COMMUN., VOL.E90–B, NO.11 NOVEMBER 2007 PAPER A Practical Routing and MAC Framework for Maximum Lifetime Sensor Telemetry Ozgur ERCETIN † , Nonmember, Ozgur GURBUZ †a) , Member, Kerem BULBUL †† , Ertugrul CIFTCIOGLU † , and Aylin AKSU † , Nonmembers SUMMARY The recent progress in sensor and wireless communica- tion technologies has enabled the design and implementation of new ap- plications such as sensor telemetry which is the use of wireless sensors to gather fine-grained information from products, people and places. In this work, we consider a realistic telemetry application in which an area is pe- riodically monitored by a sensor network which gathers data from equally spaced sample points. The objective is to maximize the lifetime of the net- work by jointly selecting the sensing nodes, the node transmission powers and the route to the base station from each sensing node. We develop an optimization-based algorithm OPT-RE and a low complexity algorithm SP- RE for this purpose and analyze their dynamics through extensive numeri- cal studies. Our results indicate that SP-RE is a promising algorithm which has comparable performance to that of the more computationally intensive OPT-RE algorithm. The energy consumption is significantly affected by the channel access method, and in this paper, we also compare the effects of the collision free TDMA and contention based CSMA/CA methods. We propose practical enhancements to CSMA/CA so that the energy consump- tion due to collisions is reduced. Our simulation results indicate that with the proposed enhancements contention based channel access can provide comparable performance to that of the collision free methods. key words: cross-layer design, optimization, algorithms, energy efficient routing, sensor networks, network lifetime 1. Introduction The recent progress in sensor and wireless communication technologies has enabled the design and implementation of new applications such as sensor telemetry which is the use of wireless sensors to gather fine-grained information from products, people and places. For example, sensor teleme- try could help in monitoring wind, water, soil and air tem- peratures or humidity in large farms or vineyards [18] or in tracking the movement of a glacier to acquire vital informa- tion that can help to predict the extent and immediate con- sequences of global warming [19]. Knowing the state of an object such as a product or a piece of equipment and using that information for creating business value is where sen- sor telemetry has high potential. For instance, determining when a pump is under pressure and might need replacing, or when fresh food is about to go out of the safe temperature Manuscript received July 28, 2006. Manuscript revised April 20, 2007. † The authors are with Electronics Engineering, Faculty of Engineering and Natural Sciences, Sabanci University, Istanbul, 34956, Turkey. †† The author is with Manufacturing Systems and Industrial En- gineering, Faculty of Engineering and Natural Sciences, Sabanci University, Istanbul, 34956, Turkey. a) E-mail: ogurbuz@sabanciuniv.edu DOI: 10.1093/ietcom/e90–b.11.3146 range can provide real-time data to new services and may improve future products. In this paper, we focus on energy-efficient, periodic sampling based sensor telemetry and assume that a monitor- ing application periodically takes measurements from sam- ple points that are equally apart in a given region. The qual- ity of service (QoS) of the monitoring application depends on the number of sample points from which the measure- ments are taken, i.e., the precision level. In real-world appli- cations, such as agricultural or structural health monitoring, this relationship between the application QoS and precision level is usually nonlinear. A low precision level results in a low QoS, because for a large portion of the area the ambi- ent conditions may remain unknown. However, a high pre- cision level may result in over-sampling (and high energy consumption) because the ambient conditions do not usu- ally change within short distances, e.g., the soil temperature on a farm. Thus, for efficient operation, it is essential to choose precision levels that do not result in under- or over- sampling. In this article, we propose algorithms that jointly se- lect a set of sensing nodes, the path from each sensing node to the monitoring station (or base station) and the associ- ated transmission power levels for all nodes in the network for a given precision level in order to maximize the sen- sor network lifetime. The network lifetime is defined as the period of time until the first node in the network exhausts its battery. We first consider the ideal time division multi- plexed case, in which the nodes transmit during dedicated time slots without interfering with each other. Assuming that nodes are awake in their dedicated time slots, we de- velop a multi-objective optimization model for the selection of sensing nodes and routing of sensor data from those nodes to a base station. Based on this optimization model, we design a near-optimal algorithm. The proposed algorithm is computationally intensive, so we also design a low com- plexity algorithm based on a shortest path procedure. In this algorithm, links are labeled with two new different link met- rics appropriate for the maximization of the network life- time. We demonstrate by simulations that the low complex- ity algorithm achieves a performance close to that of the op- timal solution, providing a significant lifetime extension as compared to the traditional but non-optimal methods such as Tiny-DB [11]. We also investigate the dynamics of the low complexity algorithm by considering various telemetry pa- rameters and network scenarios. Finally, we employ a prac- Copyright c  2007 The Institute of Electronics, Information and Communication Engineers