High accuracy data fusion algorithm for privacy serving in wireless sensor networks

Abstract

As the main tool to monitor remote networking environment, sensor network integrates various functions of data collection, data collection, data processing and information transmission information data encryption in one, which has been gradually applied in all areas of society. However, sensor networks may suffer attacks during the process of data acquisition and transmission. Therefore, data privacy protection becomes the most important problem in wireless sensor networks. In this paper, privacy protection, high precision and data fusion were studied, and a privacy preserving algorithm based on CRAE was proposed on the existing encryption algorithms. The algorithm can not only reduce the energy consumption of sensor networks, but also have good privacy. Simulation experiments were carried out to compare the proposed algorithm with the existing encryption algorithms and achieve the desired design goal.

Keywords

Privacy protection sensor networks data fusion energy consumption

1 Introduction

Sensor network is the main method of collecting information on the Internet. In a certain range of detection, a large number of sensor nodes are laid out, which can improve the accuracy and robustness of the whole sensor network. By using a large number of sensor nodes, the collected data not only has various forms, but also has the characteristics of large amount of data and complex relationship between data. Each sensor node processes, transfers and stores the acquired information data, which greatly increases the overall energy consumption of the sensor network. Therefore, in order to save the overall energy consumption of sensor networks, data fusion is adopted. Taking the sensor nodes as the center of the collected data, the routing method based on clustering is the main way to reduce energy consumption. But it is worth noting that in the traditional data fusion process, the data in the sensor network is faced with huge security risks. In the process of data fusion, sensor nodes are likely to be attacked and even captured, thus resulting in data leakage, loss, or even tampering in the sensor network. Therefore, in the collection, processing, transmission and data fusion process, how to use advanced algorithm to protect the sensor network security, information security and data privacy is the common issue of the sensor network researchers. In this paper, based on the existing privacy protection algorithm for sensor networks, further improvements are made for its shortcomings [1].

2 State of the art

At present, in the field of privacy protection technology, most of the theoretical achievements published by foreign scholars are still in the majority. However, this technology is not yet mature, and new ideas are constantly pouring out [2]. Especially in recent years, it has attracted worldwide attention, and some major research projects have made breakthrough progress. The technology has been applied and operated in some systems, and achieved satisfactory results. Foreign scholars have applied data fusion privacy protection methods in other fields and gradually transplanted them into the sensor network after improvement, and made a comprehensive and in-depth study of secure multi-party computation, therefore, it is necessary to not only ensure the correctness of the data, but also ensure that they do not disclose any information about their privacy. Finally, some data fusion privacy protection schemes for today’s sensor networks are proposed. The mechanism of integrity protection for applications in sensor networks is constantly proposed and integrated with integrity protection schemes, so as to obtain both schemes [3]. At present, the proposed privacy protection research program has its own application scope, and some of its own shortcomings still haven’t been effectively solved [4]. In China, researchers have begun to pay more attention to the research of privacy protection technology. The research of privacy protection in wireless sensor networks attracts more and more colleges and universities. Several research groups in Fudan University, Peking University and Northeastern University have studied privacy related issues [5]. However, China’s research on privacy protection technology lags behind western developed countries in general. The challenges and opportunities that China faces are coexistence [6]. Data distortion and data encryption technology are the main problems of privacy protection technology. Many basic theories have matured, but some deep-seated problems need to be further solved [7]. Generally speaking, the research on privacy protection technology is still in its infancy, and it has a lot of research space.

3 Methodology

3.1 System structure design

Sensor nodes, users, and system monitoring areas are three basic components of the sensor network. In the classical sensor networks, the sensor nodes are separated to a certain distance from the base station (the sink node), and the position is fixed, which can maintain the absolute stability of energy. Therefore, sensor nodes can obtain precise coordinates of their positions accurately. For sensor nodes in a sensor network, their positions are always fixed and have the same initial energy. Direct communication between sensor nodes can be achieved through the aggregation of sensor nodes. Each sensor node has its own unique ID and coordinates. And each sensor node’s structure and communication structure remain the same, evenly distributed in the detection environment. By debugging the power of sensor nodes, the system can restrict and adjust the power of the signal transmitted by sensor nodes. According to the different conditions of the detection environment, the fixed power of the sensor nodes is debugged, so as to reduce the energy consumption and prolong the service life of the sensor communication network. A typical sensor network system diagram is shown in Fig. 1 [8].

Fig.1

Sensor network system structure.

Fig.2

The relationship between sensor networks, detection environment, and data collection among the three.

In order to ensure the practicability of sensor network communication, the number of sensor nodes in sensor networks is very large, and it reaches a very high coverage density in the detection environment. But because the resource of a sensor node is limited, the computing power and energy storage of sensor nodes are limited, which restricts the communication ability of sensor networks. Because of the limited resources, unattended nodes and unreliable information transmission methods, sensor networks may be attacked by the outside world at any time. The security of typical sensor network data can’t be effectively guaranteed. Attackers can even attack network nodes at random, destroy and steal transmission information [9]. Wireless sensor networks communicate with each other and form a self-organizing type network. The data in the detection environment is transmitted wirelessly to the outside world. The transmission process will also be subject to magnetic field, noise and clutter interference, and its data security can’t be guaranteed. Therefore, it is necessary to design a data fusion algorithm for privacy preserving data fusion with high accuracy [10].

3.2 Privacy preserving algorithm based on CRAE

In the existing sensor networks, almost every step of the communication process needs to take into account the security issues in research and development. Privacy protection has become one of the most important parts of the design of sensor networks. Based on the existing research results, the privacy protection technology in the current sensor networks can be divided into three kinds in general. The first is data perturbation technique. The core of this technology is to fill in the random data of the original data transmitted in the sensor network, exchange encrypted perturbation data with network neighbors, and then transmit it. Finally, the data is re processed with the relevant formula, and the original attributes of the data remain unchanged. The second is data encryption technology. The core of this technique is to convert the data source to function using cryptographic keys. The data receiver uses the same encryption key to process the data function, so that it can be transformed into raw data, thus ensuring the confidentiality of data. The third is the restricted publishing technology. The core of this technology is to selectively release part of the data according to the specific circumstances of the data. But to some extent, it may cause data loss [11].

In this paper, the CRAE based privacy protection algorithm, CPDAA, is used. The algorithm is to further cluster according to the CRAE algorithm. Sensor nodes in a sensor network diverge and cluster information into many sensor nodes. After the sensor node receives the relevant information, a value is generated randomly inside the sensor node. When the value generated by the sensor node is less than the threshold T (formula (1)), the sensor node will transmit the information to other sensor nodes according to the information scattered by the node, which is regarded as the cluster head. Other sensor nodes which are not cluster heads should judge that they should join a cluster and send a signal to respond to the nearest cluster head node according to the information scattered by the cluster node and the strength of the signal. The algorithm proposed in this paper aims at the shortcomings of the CPDA algorithm and makes corresponding improvements. It can reduce the total energy consumption of the sensor network, reduce the cost of the sensor network, prolong the overall lifetime of the sensor network, and guarantee the security of the sensor nodes to transmit data [12]. $T = {\begin{matrix} 0, \\ \frac{p}{1 - p * (h \mod \frac{1}{p})} \end{matrix}$ (1)

In the formula, P represents the percentage of the number of cluster heads in the total number of sensor nodes, h represents the current number of election rounds, and mod is modulo operation. In the process of computing in a cluster, the content of the random information released by the cluster node is simplified to that each cluster model has a cluster head V0 and 11 common nodes. In the single cluster model, Y groups are set up to exchange data and transmit data. For ease of understanding, set Y = 3. Sensor nodes, as cluster heads, send information to other sensor nodes as members. Other sensor nodes, which are members of the cluster, generate random values to the cluster hair, indicating the identity of the nodes within the cluster. Among the 3 switching groups, one is randomly chosen as a tool for data exchange and data transmission. The sensor nodes within the cluster will send information from the sensor nodes to the cluster head. The detailed diagram is shown in Fig. 3. After the cluster head has obtained the related data information, the data in the three groups will be collected according to the complete CPDA algorithm, and the information data will be exchanged and further encrypted and fused. The data transmission in the sensor network is encrypted so as to ensure the privacy protection of the sensor network [13].

Fig.3

In group, nodes transmit data to cluster head.

After the collection node sends information to a large number of sensor nodes, the sensor nodes form a cluster between each other to exchange and encrypt the information data in the cluster. Then, the encrypted information is transmitted to the sub cluster header nodes through cluster heads. The sub cluster header then transfers the encrypted information data to the collection node. The above process is recycled. After the entire sensor node repeats the operation for a period of time, the sensor network performs an overall reboot according to the design. After the restart, the next round of nodes sends information, the random cluster heads and the cluster heads are selected, and new clusters are reestablished. In the form of intra cluster computing, the algorithm can ensure the privacy and security of the information data to a maximum extent, and also greatly reduce the amount of data traffic. The overall energy consumption of the sensor network is low, and the lifetime of the sensor network can be extended to the maximum extent [14].

4 Result analysis and discussion

4.1 Establishment of experimental environment

In the third chapter, the basic principle of CRAE based privacy preserving high-precision algorithm is introduced. In order to further illustrate the advantages of the proposed algorithm compared to the traditional privacy preserving algorithm, the experimental simulation environment is established in this chapter for further simulation experiments on the algorithm. Test work needs to be built on the following principles. Firstly, the functionality and content of the test covers all of the user’s actions and performance requirements. The specific tests include model function test, user rights test, database fault tolerance test and model crash test. Secondly, in the process of testing, the test staffs are not allowed to judge by subjective judgment whether the software model’s function has reached the design standard. Therefore, setting strict testing acceptance standards is very important. In the test, the staff should be strictly in accordance with test standards to determine the standard. Thirdly, software testing is an important part of software development process. In the model coding phase, developers can perform individual use case tests of modular functions. In this way, problems arising from testing in the development phase can be quickly located and resolved, while problems occurring during the testing phase are relatively difficult to find. How to log data collection under the condition of network monitoring point deployed is a problem. A data acquisition algorithm is proposed to solve the problem that whether the captured log data in the transmission bandwidth and the data acquisition is effective and timely. The algorithm reduces the number of transmissions and the amount of data transmitted by the network. When the data collected by the network management is fitted, the data fitting results can well reflect the intensity and the degree of the change of the network state of the collected points. In addition, the expected results are obtained.

In order to guarantee the validity of the data fusion algorithm based on CRAE privacy protection, the MATLAB platform is adopted to simulate the experiment. 200 sensor nodes are set up in the simulation environment. The 200 nodes are distributed randomly in the range of 100m*100m. The main parameters involved in this simulation experiment are shown in Table 1.

Table 1
simulation environment parameter table

Parameter name parameter values

E 40nJ/bit

P 3%

eamp 80nJ/bit/m2

Initial energy of node 4J

Data packet 2000bits

Parameter name	parameter values
E	40nJ/bit
P	3%
eamp	80nJ/bit/m2
Initial energy of node	4J
Data packet	2000bits

4.2 Result analysis of simulation experiments

In the analysis of energy consumption of cluster head nodes, cluster head is the main part of energy consumption in sensor networks. The energy consumption of the cluster head determines the overall energy consumption of the sensor network, and indirectly affects the lifetime of the sensor network. Therefore, in order to complete the lateral contrast and show the advantages of the algorithm, the improved CPDA algorithm is introduced into the test section. The energy dissipation of the cluster head is shown in Fig. 4. It can be seen from the diagram that the energy consumption of cluster head is lower than that of traditional CPDA algorithm in the simulation of each round of CPDAA algorithm.

Fig.4

Comprehensive comparison of energy consumption per cluster head.

In the whole sensor network, if someone attacks the sensor network, tries to obtain the data detected by the sensor nodes in the sensor network, and accesses the cluster head through technical means or eavesdropping methods, he can acquire the encrypted information in the cluster head in the case that there is no authorized sensor node to hold the communication key. In another way, the cluster head is tapped by attacking a plurality of nodes adjacent to the cluster head, and the information in the cluster head is collected. In the simulation experiments, it is assumed that the probability of obtaining a shared secret by an attacker is P overhear. The probability that attackers eavesdrop on cluster heads and obtain data is P colludes. The probability that the overall information of the sensor network is obtained by attackers is Pq. The probability of leakage of the information data in the sensor nodes is: $\underset{1}{p} (q) = \sum_{k = \underset{min}{d}}^{\underset{max}{d}} p (Z = k) \overset{k}{(1 - (1 - \overset{k - 1}{q})})$ (2)

It is assumed that the members in the cluster are Z; d is the number of sensor nodes that are assigned to members in the cluster; P (Z = k) is the probability that the member in the cluster is k. In this simulation experiment, each sensor node has the same probability of becoming cluster head, and other sensor nodes which are members of the cluster send data to the cluster head. At the same time, assuming that the probability of exposing the identity of the sensor nodes within the cluster is the same, the probability of attack of the two algorithms can be compared horizontally. In addition, the privacy protection of the two algorithms can be compared. After the simulation experiment and the calculation, the structure is shown in Fig. 5.

Fig.5

Comparison of data privacy protection.

The simulation results show that compared with the energy balance before the improved clustering routing algorithm in wireless sensor networks (CRAE), high accuracy fusion algorithms for privacy preserving in sensor networks, privacy protection algorithm based on CRAE, have advantages in energy consumption and transmission probability of sensor networks. In the sensor network, the energy consumption and computing power of each sensor node are restricted to a certain extent by the restriction of hardware facilities. It is impossible to perform complex calculation and storage of the data. The algorithm uses multiple sensors to cluster and encrypt the data according to the instruction of the cluster nodes. It can reduce the energy consumption of the whole sensor network to a maximum extent. It has strong privacy protection, real-time and reliable information and data transmission. In addition, it prolongs the service life of sensor network with lower energy consumption, and has more effective data fusion effect.

5 Conclusion

Based on the privacy protection algorithm of CRAE, a high accuracy fusion algorithm for privacy preserving was proposed in this paper, so as to effectively reduce the overall energy consumption of the sensor network. The algorithm was further improved on the basis of CPDA algorithm. In this algorithm, the nodes are sent information to the sensor nodes, and the clusters are randomly formed. It can randomly select the identity of sensor nodes in the cluster for further processing, encrypting and transmitting the collected data. Then, the data is transferred the cluster head to the collection node to realize the privacy of the data. In the event of external interference or targeted attacks, the algorithm can be addressed to some extent, and can reduce the amount of data traffic. At the same time, the privacy protection algorithm is fused, and the simulation experiment is carried out. From the simulation results, the algorithm is lower in energy consumption than the traditional privacy encryption algorithm. It has better privacy and security, and basically achieves the desired goal. However, the algorithm proposed in this paper is still insufficient. It can’t detect the attacker’s malicious deletion of data, adding, and other attacks, which need further improvement.

References

Zhang

X.Y.

, Peng

and Chen

, State-of-the-art survey of privacy-preserving data aggregation in wireless sensor networks, Journal on Communications 10(39) (2018)130–142.

Chen

and Shen

E.T.

, A Source-Location Privacy Preservation Protocol in Wireless Sensor Networks Using Source-Based Restricted Flooding, Chinese Journal of Computers 09 (2010)78–84.

Wang

T.C.

and Ding

E.T.

, Privacy-Preserving and Itinerary-Based Data Aggregation Algorithm in Wireless Sensor Networks, Acta Electronica Sonica 6(45) (2017)1334–1341.

Chen

, An Improved Routing Algorithm for Source Location Privacy Protection in Wireless Sensor Network, Chinese Journal of Sensors and Actuators 3(30) (2017)438–449.

Wang

W.X.

and Li

P.Z.

, A privacy protection method of source location in wireless sensor networks, Journal of Chongqing University 03(41) (2018)100–108.

Jiang

W.X.

and Zhang

Z.X.

, Support vector machine description of data fusion protocol in wireless sensor networks, Huazhong Univ. of Sci. & Tech.(Natural Science Edition) 08(46) (2018)99–103.

Ren

X.L.

and Ji

P.S.

, Fuzzy-weighted Algorithm for Data Fusion Based on Clustering in WSN, Computer Engineering 03(44) (2018)109–113+118.

Ozdemir

and Yang

, Integrity protecting hierarchical concealed data aggregation for wireless sensor networks, Computer Networks 8 (2011)147–154.

Sicari

, Grieco

L.A.

, Boggia

and Coen-Porisini

, A dynamic data aggregation scheme for privacy aware wireless sensor networks, The Journal of Systems & Software 1 (2011)78–83.

10.

H.J.

, Lin

and Li

K.Q.

, Energy-efficient and high-accuracy secure data aggregation in wireless sensor networks,, Computer Communications 4 (2011)54–59+63.

11.

Lin

X.D.

, Lu

R.X.

and Shen

X.M.

, Multidimensional privacy-preserving aggregation scheme for wireless sensor networks, Wirel. Commun. Mob, Comput 6 (2010)98–103.

12.

Szewczyk

, Tygar

J.D.

, Wen

and Culler

D.E.

, Security Protocols for Sensor Networks, Adrian Perrig, Wireless Networks 5 (2002)56–62.

13.

Cao

D.L.

et al., A Fault-Tolerant Algorithm for Event Region Detection in Wireless Sensor Networks, Chinese Journal of Computers 10(30) (2007)1770–1776.

14.

Zhao

Z.G.

and Qu

J.F.

, Temperature Control System of Relay Room Based on WSN and Confidence Interval Computing, Computer Engineering and Applications 30(47) (2011)219–223.