Zhaoquan Gu

Neighbor Discovery in Wireless Sensor Networks

Wireless sensor networks (WSNs) are widely used in many applications such as air pollution monitoring, natural disaster prevention, health care monitoring, etc. The sensor nodes are deployed in the monitored area and they form a wireless network through communication with nearby sensors. In this chapter, we introduce the fundamental process in constructing a wireless sensor network, which is called neighbor discovery, where the sensors can find nearby neighboring sensors when their distance is within a threshold. There are two main reasons for studying the neighbor discovery problem. First of all, the deployed sensors as a configuration may vary dynamically according to different reasons. For example, new sensors may be added and old ones removed; and some types of sensors have the ability to move around inside the area. Therefore, a sensor may need to find nearby neighboring sensors when they move into its communication range or some new sensors are deployed in the vicinity. The second reason is the limited power supply. In most situations, sensors are powered by battery and their energy is very limited. For example, suppose a sensor is powered by a 1200 mAh battery, and the processor consumes 2 mA under full power and the radio consumes 20 mA when it is turned on. If the sensor keeps the radio on and computing never stops, the lifetime of the sensor is a little more than two days. Therefore, sensors need some special method to save energy in order to extend their lifetime. A trivial way is to add sleeping mode, where a sensor keeps silent in a sleep state for most of the time, wakes up for work for only a small fraction of the time. Practically, in sleep mode, the processor’s power consumption drops to (2,upmu )A and the radio power drops to (1,upmu )A, and thus lifetime can be extended significantly. If the sensor wakes up for only (1%) of the time and sleeps for (99%) of the time, the estimated lifetime would be half a year. If the wake-up portion is (0.1%) of the time, the lifetime can be over five years. Therefore, we assume sensors are silent for most of the time, and they wake up mainly for data collection and communication. The neighbor discovery problem is to design the schedules of the sensors’ sleep mode and wake mode, such that two nearby sensors can be in wake mode at the same time to find each other, which is a kind of rendezvous problem. In this chapter, we first propose a motivational example in Sect. 19.1, and formulate the problem in Sect. 19.2. We introduce the trivial brute force algorithm for the neighbor discovery problem in Sect. 19.3 and another two algorithms: relaxed difference set based algorithm and co-prime algorithm in Sects. 19.4 and 19.5 respectively. Finally, we summarize the chapter in Sect. 19.6.

Oblivious Blind Rendezvous

Time is divided into slots of equal length and each user can access an available channel in each time slot. Rendezvous is achieved only when the users access the same channel in the same time slot. All the extant blind rendezvous algorithms assume they know the global parameter N and the labels of these N channels, and some works [1] also assume each user knows the number of users in the network. In this part, we introduce the oblivious blind rendezvous problem, where oblivious means the entities’ ports are labeled locally. As introduced in Part II, most blind rendezvous algorithms assume that all entities can see the same labels of the connected ports. However, this assumption is impractical in many distributed systems. For example, in cognitive radio networks, many works assume the licensed spectrum is divided into N non-overlapping channels with fixed labels ({1,2,ldots ,N}), and each user can access the channel not occupied by any nearby PUs as an available channel. However, this assumption may not align with the reality when designing blind rendezvous algorithms. Actually, all users may not see the same labels for the licensed channels. For example, the ‘TV white space’ that could be sensed by the users has operating frequencies ranging from 470–790 MHz in Europe [2, 4], but it is located in the VHF (i.e. very high frequency) (54–216 MHz) and UHF (i.e. ultra high frequency) (470–698 MHz) bands in the United States [3]. Obviously, the labeling of this space could be different and the same frequency band (channel) may be assigned different labels under different administrations. In a general distributed system, each user has N external ports and it can label these ports locally from ({1,2,ldots ,N}) in order to distinguish them. Any port k of user (u_i) may not be connected with port k of user (u_j) since both users may only use k to identify the different ports. In some special applications, the ports may be labeled according to a global rule. For example, the FTP service uses port 21 of the computers, and the default port for WWW service is 80. We study a more general situation where the users do not have a common labeling rule, and this can be used in many general applications. In this chapter, we first present the system model for the oblivious blind rendezvous problem, in Sect. 11.1; then we introduce the commonly used metrics for evaluation in Sect. 11.2. The problem definition is provided in Sect. 11.3 and we give examples of oblivious blind rendezvous for better understanding in Sect. 11.4. Finally, we summarize the chapter in Sect. 11.5.

Rendezvous Search in a Graph

The rendezvous search problem in a graph has been widely studied. The problem is defined as follows: two players are initially placed randomly in a space (mathscr {S}) which can be represented by discrete points or is continuous. Two players want to meet up, which is the so-called “rendezvous”. In a compact space, it is hard to define how exactly two players meet. Therefore, we assume they are said to meet if their distance is no larger than a given value r. This assumption is reasonable because two players can look around and find each other if someone is within the field of vision. r can be considered the detection radius of the player. The goal of rendezvous search is to minimize the time for the players to meet. In this chapter, we first introduce the hardness of rendezvous search in Sect. 18.1, where two types of symmetry are presented. In order to show the intuitive ideas of designing rendezvous search algorithms, we choose rendezvous search along a cycle as the example in Sect. 18.2. The rendezvous search algorithms are presented in Sect. 18.3, and we summarize the chapter in Sect. 18.4.

Blind Rendezvous Problem

Rendezvous is the fundamental process to establish a communication link between a pair of neighboring entities. In traditional multichannel wireless networks and cognitive radio networks, rendezvous is the prerequisite for communication, via which the users try to choose the same channel for data transmission. Here, we introduce the blind rendezvous problem, where blind means the entities or the users in the system do not know the others’ information and they have to make decisions completely locally. This definition makes a distinction away from centralized rendezvous where a central unit is used to provide the port or channel information to the users [1, 3], or some local common control channel is established and maintained to control and simplify the rendezvous process [2, 5]. Blind rendezvous draws a lot of attention from both academic and industrial areas due to its scalability, flexibility and robustness in implementing large scale distributed systems. We depict the blind rendezvous problem in Figs. 5.1 and 5.2. Consider a cognitive radio network which is composed of several secondary users (SUs) and several primary users (PUs). Because of the PUs’ occupancy on the licensed channels, the SUs can only have opportunistically a portion of the licensed spectrum. Suppose user A has three channels ({1,2,6}) that are not used by the PUs, while user B can access channels ({3,5,6}) after spectrum sensing. If they try to communicate with each other, they should choose an available channel for their communication attempt. However, neither of them knows the other SU’s information about the licensed channels, so they have to apply rendezvous strategies in a distributed “blind” way. Consider a simple algorithm: each SU accesses the available licensed channels in a round robin way, i.e. user A accesses channels by repeating the sequence ({1,2,6}): n n

Oblivious Blind Rendezvous for Multi-user Multihop CRN

begin{aligned} {1,2,6,1,2,6,1,2,6,1,2,6,ldots } end{aligned}

Oblivious Blind Rendezvous for Anonymous Users

n n(5.1) n nand user B accesses channels by repeating the sequence ({3,5,6}): n n

Asynchronous Blind Rendezvous Algorithms for Anonymous Users

begin{aligned} {3,5,6,3,5,6,3,5,6,3,5,6,ldots } end{aligned}