Dissertation > Excellent graduate degree dissertation topics show

Security and Energy Performance Optimization in Wireless Sensor Networks

Author: Maan Younis Abdullah
Tutor: GuiWeiHua
School: Central South University
Course: Applied Computer Technology
Keywords: Wireless sensor networks New key change function protocol (Novel Re-keying Function Protocol, NRFP) Role-based sleep scheduling and allocation unit of energy management (SRDC-LEACH) Switch evolutionary algorithm (RE)
CLC: TP212.9
Type: PhD thesis
Year: 2010
Downloads: 278
Quote: 0
Read: Download Dissertation

Abstract

The rapid growth of human needs and the pursuit of excellence has led to sophisticated technology chip, communications and the rapid development of sensor technology, and wireless sensor network (WSN) has become the new century, the core part of the sensing technology. In WSN, a tiny sensor nodes by the sensing components, data processing components and wireless communications components, these tiny nodes wide, non-selective, a large number of distributed open, and can be in a small way constitute a network of self-organization . Based on the circuit board sensors can detect different WSN adjacent areas, and effectively and in a timely manner to pass information to the sink node. WSN performance and reliability of remote sensing has been significant development. WSN can be applied to large-scale military and security, environmental monitoring, traffic control, medical care, monitoring and architecture, and even anti-terrorism. Secure communication in a computer network is very important. However, the urgent needs of pro-sensing technology have some small, low power processors and storage space low WSN into use, and these characteristics result in a WSN in the class can not be used security systems, and does not support some of the rules and complex security protocols. Over time, the class WSN has a lot of problems. Therefore, the need to find appropriate solutions and approaches to improve the safety performance of WSN. Meanwhile, the energy consumption of sensor nodes limits the lifetime of WSN, WSN can not meet the needs of different applications. Therefore, energy issues are also the basis for WSN research problems. This thesis focuses on wireless sensor network security and energy issues in-depth studies, which mainly includes three parts: The first part of the study defense issues, including the design and implementation issues, and key management technology, proposed a new security methods to achieve prevent malicious attacks on the network and reduce energy consumption two goals; second part describes the energy-saving method in wireless sensor networks; third part is to optimize the technology in the WSN application. This contribution and innovations include the following four points. A. Analysis of the security threats in wireless sensor networks and sources of energy consumption, as well as security and energy relations. Secure communication is very important in a computer network, security certification is one of the best pretreatment method, however, since the wireless sensor network nodes restricted and is not suitable arrangement in the network need to focus on certified public key infrastructure, can not be in the network using common security authentication mechanism. In WSN, the sensor nodes need automatic grouping to accomplish a specific task. Despite the global security mechanism that may already exist, but among the members of each team can safely communicate with each other is still very necessary. Wireless sensor network coverage area of ??energy waste from the same node duplication of tasks and functions performed, and the same message is sent repeatedly. When the expansion of the network, local area caused by energy consumption over a large part of the General Assembly node dies, and further lead to paralysis of the entire network. Cluster head distribution affects the communication energy efficiency is an important factor, however, on the optimal distribution of cluster heads few studies, therefore, it is necessary to find a distributed cluster head selection method to form a reasonable clusters, so that clusters head can complete tasks more efficiently forwarded. In order to improve the operability of the network, some methods should be used to ensure that some of the services offered by the network feasibility, but also need a way to distinguish between legitimate and malicious requests to prevent execution of the tasks may lead to the collapse of the network, which is the necessity of security. These problems can be divided into information security, denial of service (DoS) attacks and other attacks. Penetration in the security and energy consumption in these two areas of energy efficiency throughout clue. By strengthening the hardware protection and strengthen computing power sensor nodes can effectively enhance security, while the actual WSN security is the efficient use of limited resources and relentlessly seek the highest levels of protection balance. (2) proposed a new key change function protocol (Novel Re-keying Function Protocol, NRFP), for key management in WSN, the sensor network security can be described as the most important issue minimal network resource consumption and maximum contradiction between safety performance. WSN in space openness allow an attacker to easily eavesdrop, intercept, tampering, replay packets. These four attacks can be divided into two main categories: active attacks and passive attacks. Passive attack occurs when an attacker a protocol through observation and analysis of data, a data communications eavesdropping without interference information resources. The active attack common way is by capturing the attacker a node in the network, and using the damaged node on the network for active attacks. Active attacks on network security in which the impact is particularly severe. Since the network node energy is limited, when the WSN node receives a malicious active attacks, it is easy because of excessive loss of energy, resulting node death. When a node mortality rate is too high, the entire network will face the danger of paralysis. Key management is self-organizing wireless sensor nodes in a new key communication methods to ensure secure communications nodes in the network, and can effectively solve the WSN security risks. NRFP algorithm is proposed in this paper belong to the scope of the algorithm LEACH algorithm has been improved, effectively improve the security of WSN. Using LEACH protocol WSN, its cluster head once attacked by malicious nodes and death, where the cluster head cluster will not work. WSN cluster head in the number of attacks increases, the number of clusters can not work also increases. At the same time the death of nodes increases, the network lifetime will be shortened, seriously affect the security of the network. NRFP algorithm can effectively reduce the network against malicious attacks at the same time when the node mortality and prolong the survival time of the network, improve network security. NRFP algorithm defines three key sensor node: all the nodes in the network shared master key (MK), with a base station (BS) a local shared key (LK) and the other sensor nodes shared session secret key (SK). SK is a global shared key is used to encrypt the message from the base station, and the message is propagated to the entire cluster. LK algorithm is the key change function in the basic parameters for the node secure communications between the base station, each node has a separate base station shared with LK. SK is shared between a node and its neighbor keys. In NRFP model, SK is used for authentication or privacy origin authentication required for secure communication. Which, HMAC as a model of security authentication protocols. Each sensor node has a local administrator function (LAFs), this function consists of three components: Administrator function, password change functionality and export functions. LAFs is loaded according to the MK or LK cluster responsible for generating the session key and coordinate HMAC. Administrator functions and export functions by asking the news is from the base station or a cluster head to generate a new key value. Prior to deployment in the key, each node inject initial LK, key change function in each cycle the LK configure a new key value. Cluster head periodically refreshed in response to LK, and to inform its members of the secret of this new change. NRFP algorithm developed between the sensor function key followed by four basic principles: key deployment, key establishment, increasing nodes and node recovery. In the key deployment, each node has a unique imprint ID, MK and LK, MK and LK administrator functions generated according to a separate shared keys with other nodes. The communication process in the node, the printed key (ID, MK, and LK) is not exchanged, the only exchange the SK. In the key establishment process, all nodes in the network use the same mechanism secure communication between each other. In the cluster head after the completion of the deployment and performance of the cluster head generates SK, and sends a message to the cluster members, encouraging members of the node generates SK. Communication between nodes in the cluster to support the following cycle: between two nodes want to establish communication, they use the same secret SK establish communication and start data exchange. In this process, SK will remain consistent for all nodes use the same export function. Add nodes in the network node must be followed to increase the principle. When a new node wants to join the network, it must first join the network in any one cluster. If the node is a cluster head receives the beacon transmitted, between the cluster key is constantly updated, and the node will generate its own SK. If a node does not receive any cluster head beacon transmitted, the node itself constitutes a new cluster and the cluster as a cluster head, and then run LAFs and generate your own key. When any node in the cluster for some reason (energy consumption, the node migration, node capture, etc.) leaving in their area, the node processing node recovery principle must be followed. Node recovery is divided into two cases: child node cluster head recycling and recovery. When the cluster head node does not receive a child's HELLO message; cluster head sends HELLO message to the node and wait for a response. If within a certain time, the cluster head does not receive a response message, the cluster head will send the message to inform all members of the ID of the node removed from the neighbor list. This situation is called a child node recovery. When an initiative to leave the cluster where the cluster head, the cluster head must send a message to inform all members will leave the cluster, and the cluster members must elect a new cluster head. To be elected out of the cluster head in the cluster must have the most neighbors or have maximum energy. If the cluster head secretly left, then a period of time will not receive cluster members to the cluster head beacon. Cluster member must be based on the basic information of the cluster rebuild a new cluster, and the election of a new cluster of cluster head. In both cases referred to the cluster head recycling. In order to improve safety communications using NRFP for key changes in the process of establishment of two models: the base station model and cluster-head model. Base model and the cluster model are used passive cluster heads selection scheme, which is divided into two parts: first, the establishment of clusters; the second step, the cluster head election. Each sensor node before the embedding object region in a certain time broadcasts its own ID, and receives the ID of the neighbor messages. Node in the routing table to add the received neighbor ID information, and accumulates the received ID number. Each node obtained by this method he can reach the number of neighbors (NBR), and nodes can be connected with the establishment of these clusters or groups. Sensor nodes transmit their ID and mutual NBR information, and based on this information the election of a new cluster head. When a sensor node NBR maximum value, the node will be elected as the cluster head. The neighbor called the cluster head cluster head of the \Because the cluster heads and their offspring in the existence of a tree network generation parents and child relationship. In the base model, the base station (BS) sends a message to all the network cluster head, encouraging cluster head rebuilding a cluster and export a new SK. BS key to change in order to control the process. Wherein, SK all nodes in the network are common. The operation of the system within a specified period of time execution. Therefore, all nodes must be exported at the same time continuing SK and establish communication with each other. Key change process through the establishment of this centralized model can improve network scalability feasibility, and other applications do not need to prepare a large number of keys. Cluster-head model is similar to the base station model. The difference is that, during the change in the key, each cluster is separated from each other. Each cluster has a session key from other clusters. Clusters in the reconstruction and the end of the session key when executing when the cluster head exported public key is transmitted to the base station, and is used for inter-cluster session. The advantage of this model, each cluster according to the cluster key feature reconstruction. So when the session key is captured, it will not affect other clusters secure communications. Respectively in OmNet platform using LEACH protocol network model and the network model used NRFP simulation. Simulation results show that under the same conditions, using NRFP WSN protocol malicious data received average rate than the use of low-LEACH protocol for WSN. In summary, NRFP is a self-organizing framework, that is, a new key change function protocol. The self-organizing framework covers the key deployment, key establishment and key distribution, and proposes a cluster-based key changes (re-keying) process algorithms. This structure uses only symmetric key cryptography and is based on a simple set of assumptions and guidelines. By placing the system into two collaborative security domains to achieve energy security and balance: the domain in a simple oversight and resource to achieve a compromise between high security, rather than the opposite field supervision. NRFP advantages include distributed communication process, can quickly establish a communication between nodes, and the session key generation algorithm complexity is low, thereby consuming less energy; size of extension of the network; different applications do not require a large number of keys but only to change the generating function of these keys to suit each application. 3 presents a wireless transmission technology energy-saving sensor networks need sufficient time to meet the survival needs of individual applications, and the network's survival needs of network energy support. Energy consumption in the network, the largest proportion of the energy consumption of the communication. In the same nodes within the coverage of duplication of functions and tasks as well as continue to send duplicate messages will result in a tremendous waste of energy networks. When a part of the node death due to energy depletion, the entire network may therefore be suspended, reduced network lifetime. To solve this problem, this paper proposes a role-based task scheduling dormant cells and energy distribution management mechanisms (SRDC). The algorithm is based on the node coverage and work time settings sleeping node, and through the SDRC rotation sleep queue node, nodes in the network in order to optimize the distribution of tasks, saving network energy and prolong network lifetime. Assuming all the nodes in WSN using omni-directional antennas send messages, and the network as an example of a simple cluster. Only by the cluster member that cluster N1, N2 and cluster head NO constitute and define N1 and N2 have the same range in the same sensing domain. When the distance between N1 and NO N1 and N2 is not greater than the distance between, N1 N0 message transmitted to the same will be received by N2. Then, in the N1 send a message to NO after a period of time, the message is received N1 to N2 will also send the message NO cluster head. Therefore, by setting the node N2 N2 dormant avoid sending repetitive messages, effective conservation of energy networks. If you have the N2 in a non-working state, will inevitably lead to excessive energy consumption of N1, while N2 energy surplus. N1 and N2 energy is not balanced equally detrimental to network survivability. Therefore, the work of rotation N1 and N2 order both to avoid repetitive messages sent between the nodes can be balanced energy consumption. Based on this node turns dormant ideas put forward SRDC algorithm, which is divided into two steps: sensing domain identification and sleeping node settings. The neighbor node definition, its characteristics similar to, and have the same frame and cover the same area. When a cluster head node receives a message with the same sensing domain neighbor nodes in the cluster head node to respond after a period of time to send the same message to the cluster head. Therefore, SRDC algorithm based on the node NBR, ID and other information having the same identification sensor node of the domain. SRDC algorithm uses the queue (SRDCQ) method with the same sensing domain node hibernation settings. Each cluster has its own independent SRDCQ, the queue is stored in the sink node. Queue length of the signal sort nodes formula and determined by the order according to their corresponding nodes hibernation. In a cluster, the sensing node of the domain having the same sequence into SRDCQ. The same time, the nodes except the node number of 0 is active, the remaining nodes are in varying degrees of sleep state. SRDC algorithm queue nodes into four activity Status: S0, state S1, S2, and status state S3. Nodes in state SO strongest activity, and its data acquisition, data reception and data transmission functions are active. In this state, a node can complete the task of target detection and information transfer. For the node state S1, the data transmission function is masked, can only receive commands from the sink node. State S2 only sensing element of the node is activated. And it means that the node is in state S3 deep sleep state, which is used remote sensing, data processing, data storage, and communications components are all dormant. In this case, only through the sensor's internal clock wake up the node. These four states under the energy consumed by a node in descending order. Queues nodes using two factors change the order. These two factors were the cycle and the maximum activity. Among them, the maximum activity refers to the activity of the sensor nodes in the state in which the maximum energy consumption. When the energy consumed by the cluster head is equal to or greater than the maximum activity factor, the queue after each cycle in which it will be based on the time (period), and the energy consumption (maximum activity) to change the order of nodes. With a six nodes containing only the cluster as a simple example to describe the working process SDRCQ. Initial state, N0 is elected as the cluster head, N1 to N5 are the cluster member, and N1 and N2, N3 and N4, N4 and N5 have the same sensing domain. N0 to N5 in accordance with the sensing domain in order to enter the queue. SDRC algorithm makes sense in the other nodes in the same domain N2 and N4 dormant. In the round end of the cycle, SDRC algorithm based on each node in the first cycle energy consumption and the use of time of the updated node queue order. As in the S0 state N0 cluster head consumes a lot of energy, so it is discharged out the end, in a dormant state. And Node 3 into the first row, was elected as the cluster head. N4 and N2 continues in a dormant state, N1 and N5 is active, and enter the second round of the cycle. Whether the node in the previous cycle is activated for how long, how much times as long as the node can contain maximum energy priorities ranked in the first column, a cluster head. It can be seen, the clusters are not used on the premise SDRC algorithm may repeatedly transmit the message N2 and N4 resulting waste of energy and shorten its lifetime. And in the context of limited energy supply, N0 since been used as cluster head will consume a lot of energy. The number of cycles increases, the energy consumption of nodes in the cluster surge, and will severely uneven distribution. And after using SDRC algorithm, the algorithm has been reduced as a cluster head N0 energy consumption caused by the pressure, and reasonable to avoid the loss of energy caused by repetitive messages. In a number of cycles, the energy of nodes in the cluster does not produce significant uneven distribution. In OmNet platform LEACH protocol used separately as well as through a network model SDRC algorithm to improve the LEACH protocol network model for simulation. Simulation results show that the SRDC-LEACH protocol network its life time can be maintained up to the first 23 cycles, while the use of LEACH protocol network can only be maintained to 18 cycles. In practice, SRDC algorithm maintains a network of 80-190% of the node energy, and almost all the network nodes to control the die at the same time, ensure that the task of continuing the distribution within the cluster, so that in all network nodes node dies again before work. Overall, SRDC algorithm to maximize the residual energy constrained overall return expectations as the goal of a single application design an optimal transmission strategy, and for the transmission of multiple applications design the optimal control strategy. In the design of optimal transport policy, in accordance with the transfer of programs and the overall energy consumption of the network to establish a convex optimization model, and the use of sophisticated optimization techniques to solve the model to obtain the optimal transmission strategy. In the design of optimal transmission control strategy, the problem is abstracted as a dynamic process. SRDC algorithm to enhance the security of wireless sensor networks. A large number of in-depth simulation results proved that the adoption SRDC mechanisms combined with existing LEACH protocol can achieve low-power communication structure, and thus extend the lifetime of wireless sensor networks. 4 is proposed based on fuzzy logic particle swarm algorithm (PSO) and genetic algorithm (GA) switching evolutionary algorithm (Revolving evolutionary, RE) optimal distribution of cluster head can increase the energy consumption of wireless sensor network communication efficiency, while using the cluster head node for data fusion method can extend the network lifetime. Therefore, the proposed fuzzy logic-based particle swarm optimization (PSO) and genetic algorithm (GA) switching evolutionary algorithm (Revolving evolutionary, RE) to optimize clustering to extend the network lifetime. RE The main idea is clustering nodes and select the optimal cluster head, so as to satisfy three objectives: ① minimize energy consumption; intra-cluster distance is minimized, ie, cluster data to minimize the distance between the vectors . inter-cluster distance is minimized, that is, different cluster centroid minimize the distance between. RE full advantage of the genetic algorithm and particle swarm advantages, beginning from GA optimization algorithm, based on the status of the optimization algorithm using fuzzy logic (FLC) in the switch between GA and PSO. Define the rate of change of the fitness function (RCh) and optimize the schedule (OpP, defined as the end of the actual optimization evolution algebra and algebra biggest evolution ratio) two variables as fuzzy logic system input, and output variables for the next step will be Use of optimization algorithms (COp) and the duration of the next stage (NCPo). Fitness function of the rate of change (RCh) Is the RE algorithms currently used algorithms measure the RCh be normalized, thus ensuring the input variables of fuzzy logic RCh able to adapt to a large range of variation. Two fuzzy logic input variables and output variables NCPo divided into three, and three were used to represent the fuzzy sets and fuzzy output variable COp uses two sets to represent. GA or PSO algorithm for all nodes in the input position and its energy, the output for clustering results and the cluster head. For each node, calculate the distance between nodes and other cluster heads and all the distance between them assigned to their nearest cluster head where the cluster, and then calculating the distance and energy fitness function, and then start FLC to judge whether the algorithm of the switch, and so forth until it meets the ultimate goal. Algorithm uses real coding, applied to wireless sensor networks for different population size (50,100,200) test run 100 times for each set of parameters, compared with LEACH protocol in wireless networks, LEACH PSO, LEACH GA and LEACH RE when the algorithm performance include: ① the convergence speed: RE convergence speed faster than the other algorithms, and RE compared to PSO and GA can get a better global optimum, but also relatively simple to use. ② network lifetime (defined as the time for the simulation with the living number of nodes): simulation results that the use of the network after optimization algorithm will extend the life, and the use of LEACH RE algorithm, the longest network lifetime; This is because the using the optimization algorithm, the network of intra-cluster distance is minimized and the optimal distribution of the cluster head in the network, so the communication process, the members of the cluster nodes, and the communication between the head a short distance, then the energy consumption small. In contrast, in LEACH when used alone, due to network clustering effect is poor, some nodes must pass a long distance to find its cluster head, therefore, large energy consumption. Meanwhile RE GA than PS0 and superior performance is due to the RE fitness function value is smaller. ③ different algorithms, the amount of data sent to the base: The base station receives the data statistics show that more successful RE algorithms can send more data to the base station. Because the cluster head needs to send data to the fusion of a base station located outside the network area, so choose a higher energy cluster head sends a very important message. After using the PSO algorithm, in the steady state before the end of the cluster head is unlikely to run out of energy, which can send more data to the base station. Instead, LEACH not guarantee selection of cluster head having a high energy transmission data to the base station, therefore, some cluster head may be in the process of sending data deaths resulting data to the base station is reduced. This article does not compare these algorithms and K-means algorithm performance, because the K-means can not make network energy consumption is minimized. In this thesis, the results obtained for future research work has laid a solid foundation. As a continuation of this study, future research could focus on LEACH and his security version. WSN to attacks and the resulting performance degradation of the simulation study is meaningful research fields. Furthermore, the key management issues also need more research. On a deep understanding of security authentication mechanisms, such as certificate-based methods and related properties of WSN security applications designed for very valuable.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Automation technology and equipment > Automation components,parts > Transmitter ( converter),the sensor > Sensor applications
© 2012 www.DissertationTopic.Net Mobile