466 IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 27, NO. 4, MAY 2009 LACAS: Learning Automata-Based Congestion Avoidance Scheme for Healthcare Wireless Sensor Networks Sudip Misra, Vivek Tiwari and Mohammad S. Obaidat, Fellow, IEEE Abstract—One of the major challenges in wireless sensor network (WSN) research is to curb down congestion in the network’s traffic, without compromising with the energy of the sensor nodes. Congestion affects the continuous flow of data, loss of information, delay in the arrival of data to the destination and unwanted consumption of significant amount of the very limited amount of energy in the nodes. Obviously, in healthcare WSN applications, particularly in the ones that cater to medical emergencies or in the ones that closely monitor critically ailing patients, it is desirable in the first place to avoid congestion from occurring and even if it occurs, to reduce the loss of data due to congestion. In this work, we address the problem of congestion in the nodes of healthcare WSN using a learning automata (LA)- based approach. Our primary objective in using this approach is to adaptively make the processing rate (data packet arrival rate) in the nodes equal to the transmitting rate (packet service rate), so that the occurrence of congestion in the nodes is seamlessly avoided. We maintain that the proposed algorithm, named as Learning Automata-Based Congestion Avoidance Algorithm in Sensor Networks (LACAS), can counter the congestion problem in healthcare WSNs effectively. An important feature of LACAS is that it intelligently “learns” from the past and improves its performance significantly as time progresses. Our proposed LA- based model was evaluated using simulations representing health- care WSNs. The results obtained through the experiments with respect to performance criteria having important implications in the healthcare domain, for example, the number of collisions, the energy consumption at the nodes, the network throughput, the number of unicast packets delivered, the number of packets delivered to each node, the signals received and forwarded to the Medium Access Control (MAC) layer, and the change in energy consumption with variation in transmission range, have shown that the proposed algorithm is capable of successfully avoiding congestion in typical healthcare WSNs requiring a reliable congestion control mechanism. Index Terms—Congestion control, healthcare applications, learning automata, performance evaluation, wireless sensor net- works (WSNs). I. I NTRODUCTION D UE to the growing demand for low cost “networkable” sensors, in conjunction with the recent developments Manuscript received 30 July 2008; revised 1 January 2009. S. Misra is with the School of Information Technology, Indian Institute of Technology, Kharagpur, West Bengal, India (e-mail: sudipm@iitkgp.ac.in). V. Tiwari can be contacted at the School of Information Technology, Indian Institute of Technology, Kharagpur, West Bengal, India (e-mail: ece.vivek@gmail.com). This author is currently affiliated with the Depart- ment of Electronics and Communication Engineering, National Institute of Technology, Durgapur, West Bengal, India. M. S. Obaidat is with the Department of Computer Science, Mon- mouth University, West Long Branch, NJ 07764 USA (e-mail: obai- dat@monmouth.edu). Digital Object Identifier 10.1109/JSAC.2009.090510. of Micro-Electro Mechanical System (MEMS) and Radio Frequency (RF) technology, new sensors come with advanced functionalities for processing and communication. A WSN consists of such sensors, which have tight constraints with respect to computational power, storage and energy resources [1]. WSNs, in recent years, have advanced in leaps and bounds due to their innumerable applications in various fields including the military, civilian, mining, healthcare and scien- tific monitoring for commercial purposes. One of the popular application domains of WSNs is healthcare, specifically, the remote monitoring of the conditions of ailing patients [2], [3], [4], [5], [6]. In general, in healthcare applications, the sensor nodes are, typically, deployed over a region in space and, based on the specific healthcare related task they are targeted for, they generate data (such as the pulse rate of a critically sick patient), which are eventually delivered to the control center for analysis by medical personnel. As such, WSNs are charac- terized by several features, out of which their unique network topology, diverse applications, distinct traffic characteristics and message size are of concern. Additionally, the hard energy consumption constraints imposed on the nodes in WSNs make the different protocols and mechanisms devised for these networks consider the energy consumption parameters with prime importance. Quite intuitively, superposing the healthcare criticalities, such as fast response to medical emergencies, high reliability of transmission of data from the source nodes to the sink node and their prompt delivery, bring further challenges in healthcare WSNs. In this paper, we focus only on the issue of congestion in healthcare WSNs. In particular, we focus on large-scale medical disaster response applications. As will be discussed elaborately in Section II-A, congestion is a severe problem in such situations. Obviously, it is unwanted to have packets containing life-critical information being queued-up in the intermediate nodes in the multihop paths, thereby delaying the transmission of information, or such packets being dropped because of occurrences of congestion in the intermediate nodes. In short, the issue of occurrences of congestion in WSNs is linked to the following. Congestion leads to ei- ther dropping the packets at the intermediate nodes or the formation of queues in them, which, again, in effect, would lead to packet delay. These challenges are to be mitigated in a very effective and efficient manner so that fairness, in terms of the distribution of the packets amongst the nodes, is guaranteed without unwantedly withering away much of the nodes’ energy levels. There are two terminologies relating to how congestion could be handled – one is congestion 0733-8716/09/$25.00 c  2009 IEEE Authorized licensed use limited to: University of Nebraska - Lincoln. Downloaded on May 12, 2009 at 23:27 from IEEE Xplore. Restrictions apply.