The Research Of Cross-Layer Networking Protocol For Asynchronous Wireless Sensor Networks

Wireless Sensor Networks (WSN) is a collection of plenty of smart moteswith sensing, computing, and storing capabilities, which are deployed over themonitoring area. WSN has extensive application values since it doesn't needthe support of communication infrastructures, can be deployed quickly, and cansustain long-term and large-scale monitoring. However, due to its high complexity,there have not been really e?ective networking technologies to bring long-term andlarge-scale monitoring into practice. It is very important to encourage informationshare and cooperation among di?erent layers to improve the e?ciency of WSNwhose all kinds of resources are severely restricted.Limited energy has been one of the most important constraints of long-termand large-scale monitoring. Sleep/active scheduling helps decrease idle listeningand reduce energy consumption e?ectively. According to whether clock synchro-nization is needed, MAC protocols with sleep/active scheduling can be classi?edinto the asynchronous and the synchronized. Without periodical synchronizationcosts, asynchronous MAC outperforms the synchronized in low tra?c networks.However, due to the lack of schedule knowledge of the receiver, the sender has towait for the receiver to wake up before data transmission, which introduces ad-ditional waiting cost and may a?ect the performance of routing protocol greatly.In this paper, by employing cross-layer design, we develop a more energy-e?cientasynchronous MAC and two e?cient stateless routing protocols for asynchronouswireless sensor networks. The main contributions and innovations are fourfold:1. Propose an opportunistic-cooperation based asynchronous MAC protocol,OC-MAC:If data packet is small in size and duty cycle is low, waiting for the re-ceiver may even consume more energy than transmitting data packet. InOC-MAC, neighboring senders that are all waiting for receivers make useof routing information to launch cooperation communications. After dele-gating its data to a selected agent node, a sender can go to sleep before thereceiver wakes up. Cooperation communication not only can reduce energyconsumption and delay, but also helps decrease the interferences and colli-sions. Though cooperation communication works in the contending and cas- cading scenarios, itself hardly brings additional overhead. Extensive resultsfrom test-bed measurements and simulations show OC-MAC that makes useof routing information to launch cooperation communications outperformsexisting asynchronous MAC protocols.2. Build a more practical virtual coordinate system for asynchronous wirelesssensor networks, including energy and delay coordinates, and propose anenergy-e?cient cross-layer routing protocol based on newly built energy co-ordinates (EEMR: Energy-E?cient Integrated MAC and Routing Protocol):With the help of geographic information or hop count, traditional statelessrouting protocols aim to ?nd the shortest path from the source to the desti-nation through local greedy forwarding strategies. However, their shortestpath is not necessarily the path with the lowest energy consumption orend-to-end delay, since they ignored the in?uences of actual irregular prop-agation of wireless channel and sleep/active duty cycles. In this paper, wedivide link cost into transmission cost and waiting cost, and assign each nodewith a virtual coordinate, which can re?ect the actual path costs, includingenergy and delay costs. A data-driven and hop-by-hop mechanism is usedto update coordinates against time-varying link costs network topologies.Then, an energy-e?cient cross-layer routing protocol which is based on en-ergy coordinates is proposed. In EEMR, each sender dynamically selects theoptimal relay and minimizes the path energy consumption by taking boththe energy coordinates and wakeup time of neighbors into account jointly.Simulation results show that EEMR is more energy-e?cient than traditionalgeographic routing in asynchronous wireless sensor networks.3. Propose a delay-constrained and energy-e?cient routing protocol (DCEER)for time-sensitive and asynchronous wireless networks. Compared withEEMR, DCEER has the following improvements:Firstly, DCEER maximizes the energy e?ciency under the constraint thatthe delay will not exceed the allowable upper bound. Next, EEMR assumesa sender is aware of the current coordinates of all neighbors, but these co-ordinates may become outdated if the data tra?c is relatively low, whicha?ects the energy e?ciency of forwarding path greatly. Instead, a senderin DCEER only estimates the coordinate distribution of neighbors, models the dynamic relay selection as a strong Markov process and employs optimalstopping theory to control the relay selection process and minimize the pathenergy. Thus, DCEER adapts to dynamic coordinates better than EEMR.Finally, the sleep intervals of all nodes are optimized according to the datatra?c to balance energy consumption, settle the"hot spot"problem whichuniversally exists in multihop networks, and prolong the network lifetimefurther.4. Propose an adaptive and distributed clustering scheme (ADCS):Clustering is a kind of synchronized energy saving protocol, which is designedfor periodical WSN in the early research, and refers to both MAC androuting protocols. In clustering protocols, all nodes are divided into clustersand each cluster elects a head to form backbone. Plain node communicateswith the cluster head and operates duty cycle according to the slots allocatedby the cluster head. ADCS improves the algorithm of electing cluster headand forming the cluster. Both residual energy and neighbor node degreeof a candidate are taken into account jointly during the cluster election.Then, both the communication distance and cluster overload are consideredwhen a plain node decides to join which cluster. ADCS adjusts to all kindsof network distributions adaptively, generates evenly distributed clusters,which helps balance energy consumption and prolong the network lifetime.Simulation results show that ADCS outperforms LEACH signi?cantly byprolonging the network lifetime over 40% in uniform networks and 75% innon-uniform networks, respectively.