WO2017166539A1 - 一种基于多代表节点与多层融合的异构传感数据收集方法 - Google Patents

一种基于多代表节点与多层融合的异构传感数据收集方法 Download PDF

Info

Publication number
WO2017166539A1
WO2017166539A1 PCT/CN2016/090088 CN2016090088W WO2017166539A1 WO 2017166539 A1 WO2017166539 A1 WO 2017166539A1 CN 2016090088 W CN2016090088 W CN 2016090088W WO 2017166539 A1 WO2017166539 A1 WO 2017166539A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
data
base station
value
hop count
Prior art date
Application number
PCT/CN2016/090088
Other languages
English (en)
French (fr)
Inventor
刘语欣
刘安丰
淡州阳
廖志军
Original Assignee
中南大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中南大学 filed Critical 中南大学
Publication of WO2017166539A1 publication Critical patent/WO2017166539A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30—Services specially adapted for particular environments, situations or purposes
    • H04W4/38—Services specially adapted for particular environments, situations or purposes for collecting sensor information
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
    • H04W52/02—Power saving arrangements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
    • H04W52/02—Power saving arrangements
    • H04W52/0209—Power saving arrangements in terminal devices
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W84/00—Network topologies
    • H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Definitions

  • the invention belongs to the field of wireless network heterogeneous data collection, and particularly relates to a sensor data collection method capable of improving network life.
  • the wireless sensor network is a wireless network composed of a large number of self-organizing and multi-hopping nodes through multi-hop wireless links and communication, which can be widely used in industrial monitoring, agriculture, civil, environmental monitoring, battlefields. , marine, fire and other special environments and applications, is considered to be one of the key infrastructure technologies of the future important IoT network.
  • a crucial issue in wireless sensor network research is how to transmit network-aware data to the base and increase the lifetime of the network. The lifetime of the network is the primary issue in wireless sensor network research. Therefore, the proposed research is also extremely numerous, and only relevant research related to the present invention will be described below. Its background is discussed as follows:
  • An important way to save energy is to reduce the number of nodes participating in the perception. Only a subset of the nodes are selected from the nodes to participate in the monitoring of events or targets. This will achieve the following two objectives: (1) Select only a part of the nodes from the event or the target area for sensing, thereby reducing the number of nodes participating in the perception, so that the nodes that do not participate in the sensing can enter the sleep state, thereby saving energy. (2) Corresponding to reduce the perception of nodes with small event or target perception information, and selecting nodes with high quality of event or target monitoring information can reduce the data sent to the base station without reducing the quality of event and target monitoring. Quantity, thereby increasing network life.
  • each node sends the sensed data independently to the base station, so even if fewer nodes are selected to monitor the event and the environment, there is still a large energy consumption.
  • a method for representing nodes is proposed by substituting a value called a representative node in the data coverage set to approximate the perceptual value of all nodes in the entire set. Therefore, it is not necessary to send each node's perceived value to the base station, so the representative node method can further reduce the amount of data transmitted to the base station, thereby saving energy, but this requires the application system to tolerate a certain data error range.
  • the nodes deployed in the wireless sensor network have correlations between the perceived information. For example, when an event occurs, multiple nodes sense the event at the same time, so it is not necessary to send the node data of each sensory event information to the base station. Instead, the number of data perceived by these nodes is counted before being sent to the base station. According to the fusion, only the data that can accurately express the event information is extracted, that is, the data is fused and then sent to the base station. In this way, the amount of data that the node needs to send can be reduced without reducing the information acquired by the system.
  • the existing research problems are: (1) In the representative node method, the only research can only deal with a single perceptual value network, but cannot handle multiple perceptual value networks. (2) The proposed node selection methods are very complicated and require a lot of energy to be consumed. (3) Each representative node is sent to the base station independently, so that the energy consumption is relatively large. Data fusion technology was not used. There are still major improvements in the current technology. Therefore, it is necessary to design a heterogeneous sensor data collection method to improve network lifetime.
  • the technical problem to be solved by the present invention is to provide a heterogeneous sensing data collection method based on multi-representative node and multi-layer fusion, and the heterogeneous sensing data collection method based on multi-representative node and multi-layer fusion can reduce transmitted data. Quantity, improve network life.
  • a heterogeneous sensing data collection method based on multi-representative nodes and multi-layer fusion.
  • each node can sense multiple types of sensing values, such as different types of sensing values such as temperature, humidity, and air pressure.
  • Class-aware data is called heterogeneous data; the following two methods are used to improve the network life, that is, the information of the perceived data meets the application requirements, and the network lifetime is maximized as much as possible:
  • Method 1 For each type of perceptual data, multiple data coverage sets are formed in the entire wireless sensor network, and the data coverage set is recorded as DCS, and the difference of such perceptual data values of each data coverage set node is smaller than the application setting. Threshold, each data coverage set has a head node as a representative node;
  • Method 2 For each type of sensing data, the representative node initiates a route to the base station to transmit the type of sensing data; the routing of the same type of sensing data is collected in a routing path as much as possible in the process of routing to the base station. The data is transmitted and transmitted to the base station.
  • Each node randomly selects a number between 0 and 1. For node A, if the selected number is less than the threshold P(A), the node acts as a set head node; where P(A) is calculated as follows:
  • mod is the remainder function
  • p is the initial probability that the node is elected as the set head node, which is set in advance according to the application requirements
  • ⁇ is the number of rounds currently cycled
  • G is the most recent The round is not elected as the set of nodes of the set head node
  • E avg and E A represent the average energy of node A and its neighbor nodes and the current energy of node A, respectively.
  • each node sets its own hop count to ⁇ , that is, infinity
  • the broadcast message is CM A , indicating that the message of the set member node A is created; the broadcast message CM A of the node A contains 6 items:
  • Type of perceived value ⁇ is the type of perceived value, The total number of types of table-aware values
  • hop count ⁇ A is less than
  • d(.) is the difference function
  • ⁇ k represents the difference threshold of the k-th data perceptual value, which is a preset known value
  • the node A' that fails to be the set head node and is not able to become the data set set member automatically becomes the set head node.
  • the data fusion routing method is adopted, and the same type of data is merged, and then the representative node sends the data to the aggregation node;
  • the data fusion routing method includes the following steps:
  • Step S1 The hop count of the base station reaching the base station is 0, and each node sets the hop count of each type of observation value to reach the base station as ⁇ , thus sharing Then, the base station broadcasts out, and the node that receives the broadcast message compares the hop count of its own k-th sensed value to the base station and the hop count in the broadcast packet, if the hop count in the broadcast packet is less than 1 saved by itself. If the k-th sense value reaches the hop count of the base station, the number of hops of the broadcast packet plus 1 is used instead of the saved hop count. Then, the node broadcasts the hop count of the base station itself to the same broadcast, so that each node gets the arrival base station. The minimum hop count of the kth class of perceived values;
  • Step S2 When the data of the kth set head node needs to be transmitted to the base station, the shortest path algorithm is used, and the node that is nearly one hop away from the base station is sequentially selected as a relay node for routing, and the data is transmitted to the base station;
  • Step S3 The node performing the kth type data routing in step S2 sets the hop count of the kth class to the base station to 0; then, the node whose hop count becomes 0 outspreads the hop count of the base station to reach the base station, and the method thereof Similar to step S1, the nodes in the affected area change their own hop counts to the base station, so that other similar data routes are attracted to the first created route, so that multiple data routes are collected on one route for data fusion. And then sent to the base station to reduce the energy consumption by reducing the amount of data that needs to be sent, thereby improving the network life.
  • the information perceived by multiple nodes may be very similar. If the range of information perceived by the node is within a specified threshold, the difference between the perceived information of a node and its perceived information may be used at the threshold. The perceived value of other nodes in the range. This can reduce the amount of data that the network needs to transmit to increase network lifetime.
  • a heterogeneous data collection method of multi-representative nodes and multi-layer fusion is further proposed according to the above principles.
  • the present invention as shown in FIG. 1, not only nodes whose perceptual values are within a certain threshold range constitute a set, but the perceptual value of all nodes of the set is represented by a perceptual value representing the node, and only the value of the node needs to be sent. Go to the base station.
  • the method for collecting heterogeneous sensing data based on multi-representative node and multi-layer fusion is a distributed representative node selection method, which can be completed by randomly electing a representative node from any node, thereby making the algorithm energy consumption small. Simple and effective; able to handle multiple sensing values; the data is merged and sent to the base station, which further reduces the amount of data and improves network life.
  • the improvements of the present invention are as follows: (a) Previous studies have been able to handle only a single perceptual value. In the actual network, the sensor nodes are often equipped with a variety of sensor components, such as a sensor node can often sense the temperature, humidity, pressure and other perceived values. Therefore, the heterogeneous perceptual value cannot be represented by a representative value, but is represented by a plurality of different types of representative values. Therefore, different heterogeneous perceptual values are represented by different representative values in the present invention; (b) Relaxing the need for appropriate selection of representative nodes in previous studies to reduce network transmission data.
  • the complexity of the method is often increased in order to select a suitable representative value.
  • the present invention relaxes this constraint by simply specifying that the value of the node that is not in the set is a representative value. Because the present invention also performs secondary data fusion on the data transmitted to the base station by the representative node, on the one hand, the amount of data transmitted to the base station is greatly reduced, and the complexity of selecting the representative value is reduced.
  • Another important improvement of the present invention is that when the value of the representative node is sent to the base station, the route convergence method is adopted, so that the routes of the same type of perceived value are merged, thereby performing data fusion, which can greatly reduce the transmission to the base station.
  • the amount of data increases the life of the network.
  • routing branch 1, routing branch 2, routing branch 3 data are finally converged on one route, so that these same kind of data is then data fusion to reduce the amount of data transmission.
  • each route is separately routed to the base station, so that the amount of data transmission is large, and thus the method of the present invention can greatly improve the network lifetime.
  • the method of the invention can simplify the previous algorithm, expand the application range of the algorithm, reduce the amount of data transmitted to the base station, and improve the network life.
  • the inventive method is divided into two components. 1: For each type of perceptual data, a plurality of data coverage sets are formed in the entire wireless sensor network, and the difference between the data in the same data coverage set is less than a prescribed threshold. Thus, each data coverage set can use a representative node to represent the perceived value of the entire set. This is the first level to reduce the amount of data transmitted by the network; 2: the data representing the node in the process of routing to the base station, the representative nodes of the same type of data will be merged into one as many as possible according to the routing algorithm proposed by the present invention. On the path, the same type of data is re-data fused, again reducing the amount of data that needs to be transmitted to the base station. Thus the method of the invention can significantly increase network lifetime.
  • FIG. 1 is a schematic diagram of heterogeneous sensing data routing based on multi-representative node and multi-layer fusion;
  • FIG. 2 is a schematic view of a collective formation process
  • FIG. 3 is a schematic diagram of a process of forming a route
  • Figure 4 is a comparison diagram of the amount of data undertaken by nodes under different methods
  • Figure 5 is a comparison of the energy consumption of nodes under different methods.
  • Figure 6 is a comparison diagram of the maximum energy consumption of the network
  • Figure 7 is a comparison diagram of maximum energy consumption under different data correlation coefficients
  • Figure 8 is a comparison of the remaining energy rates of the network under different methods.
  • each node can sense multiple types of sensing values, such as temperature, humidity, and air pressure.
  • the perceptual value of the type, the multi-class perceptual data is called heterogeneous data; the following two methods are used to improve the network life, that is, the information of the perceptual data is maximized as long as the information needs are met:
  • Method 1 For each type of perceptual data, multiple data coverage sets are formed in the entire wireless sensor network, and the data coverage set is recorded as DCS, and the difference of such perceptual data values of each data coverage set node is smaller than the application setting. Threshold, each data coverage set has a head node as a representative node;
  • Method 2 For each type of sensing data, the representative node initiates a route to the base station to transmit the type of sensing data; the routing of the same type of sensing data is collected in a routing path as much as possible in the process of routing to the base station. The data is transmitted and transmitted to the base station.
  • Each node randomly selects a number between 0 and 1. For node A, if the selected number is less than the threshold P(A), the node acts as a set head node; where P(A) is calculated as follows:
  • mod is the remainder function
  • p is the initial probability that the node is elected as the set head node, which is set in advance according to the application requirements
  • ⁇ is the number of rounds currently cycled
  • G is the most recent The round is not elected as the set of nodes of the set head node
  • E avg and E A represent the average energy of node A and its neighbor nodes and the current energy of node A, respectively.
  • each node sets its own hop count to ⁇ , that is, infinity
  • the broadcast message is CM A , indicating that the message of the set member node A is created; the broadcast message CM A of the node A contains 6 items:
  • Type of perceived value ⁇ is the type of perceived value, The total number of types of table-aware values
  • hop count ⁇ A is less than
  • (II) compare the value of the CM A. ⁇ data of the CM A. ⁇ with the value of the CM A. ⁇ of the set head node CM A .h, if Then B sends the join message JM to the set head node A to the set head node A, and continues the operation of (III); the process of joining the message (JM) is relatively simple, and the node receiving the JM message is minimized.
  • the routing method forwards the JM to the set head node, so that the final JM arrives at the set head node, and the set head node records each member node in the data overlay set;
  • d(.) is the difference function
  • ⁇ k represents the difference threshold of the k-th data perceptual value, which is a preset known value
  • the node A' that fails to be the set head node and is not able to become the data set set member automatically becomes the set head node.
  • the data fusion routing method is adopted, and the same type of data is merged, and then the representative node sends the data to the aggregation node;
  • the data fusion routing method includes the following steps:
  • Step S1 The hop count of the base station reaching the base station is 0, and each node sets the hop count of each type of observation value to reach the base station as ⁇ , thus sharing Then, the base station broadcasts out, and the node that receives the broadcast message compares the hop count of its own k-th sensed value to the base station and the hop count in the broadcast packet, if the hop count in the broadcast packet is less than 1 saved by itself. If the k-th sense value reaches the hop count of the base station, the number of hops of the broadcast packet plus 1 is used instead of the saved hop count. Then, the node broadcasts the hop count of the base station itself to the same broadcast, so that each node gets the arrival base station. The minimum hop count of the kth class of perceived values;
  • Step S2 When the data of the kth set head node needs to be transmitted to the base station, the shortest path algorithm is used, and the node that is nearly one hop away from the base station is sequentially selected as a relay node for routing, and the data is transmitted to the base station;
  • Step S3 The node performing the kth type data routing in step S2 sets the hop count of the kth class to the base station to 0; then, the node whose hop count becomes 0 outspreads the hop count of the base station to reach the base station, and the method thereof Similar to step S1, the nodes in the affected area change their own hop counts to the base station, so that other similar data routes are attracted to the first created route, so that multiple data routes are collected on one route for data fusion. And then sent to the base station to reduce the energy consumption by reducing the amount of data that needs to be sent, thereby improving the network life.
  • the data coverage set algorithm proposed in accordance with the present invention forms a set of data coverages, each set representing the value of the data coverage set by a perceptual value called an overlay set head node.
  • each type of data has multiple sets of data coverage. Any node that has such a perceptual value must belong to one of such data coverage sets.
  • Run the same algorithm for each type of perceptual value so that each type of perceptual value in the network is covered by multiple sets of data coverage.
  • each data coverage set is initiated by the set head node to the base station.
  • the initiated route has the characteristics that the same type of routes are brought together, and thus, the multiple set head nodes are collected in the routing process. On one path (routing path), data fusion can be performed, thereby further reducing the amount of data that needs to be transmitted to the base station, thereby improving network lifetime.
  • the following process of forming a set header node refers to a process of forming a set head node of a class k data overlay set (DCS). Since the formation process of other class data cover sets is similar to the process of the kth class data cover set process, Just need to perform the following process together Once, it is possible to form a data coverage set for the perceptual values of all classes in the network.
  • DCS class k data overlay set
  • each node randomly selects a number between 0 and 1, and if the number selected by the node A is smaller than the threshold P(A), the node acts as a set head node.
  • P(A) is calculated as follows:
  • Equation (8) Equation (8) that the node that has selected the cluster head will not become the cluster head in the next 1/p round cycle.
  • E avg and E c represent the average energy of node A and its neighbor nodes and the current energy of node A, respectively.
  • each node sets its own hop count to ⁇ for each type of DCS set head node. After the previous operation set header node is generated, each node that becomes the collection head node performs subsequent operations:
  • the broadcast message CMA of the node A mainly includes six main contents: (a) the ID of the set head node, which is recorded as CM A .h. (b) Node A's own ID, set to CM A .ID. (c) The number of hops that the node reaches the set head node A (d) Limit the value ⁇ of the broadcast within a certain number of hops, denoted as CM A . ⁇ .
  • CM A. ⁇ The perceived value of the aggregated head node A is denoted as CM A. ⁇ .
  • (II) the value of the perceived value of the CM A . ⁇ CM A . ⁇ their first category of data and CM A .h perceived value of the collective head node compared, if Then B sends a join message (JM) to the set head node A to join the set head node A, and continues the operation of (III).
  • JM join message
  • the process of joining the message (JM) is relatively simple.
  • the node receiving the JM message is in a minimum routing method.
  • the set head node forwards the JM, and thus the final JM arrives at the set head node, and the set head node records the data overlay set for each member node.
  • CM A . ⁇ -1 0, all operations are completed. Otherwise, the content of the broadcast packet is updated and then broadcast out.
  • the formation process of the data set members is shown in Figure 2. If the node A becomes the aggregate head node in the competition, the node A broadcasts to the outside, and the node in the broadcast range receives the broadcast information of the node A, and if it finds that it belongs to the A. If the distance of the perceived value is within the error range, then the join is performed to the set head node A. At this time, the nodes P, B, G, J, and D are added to the set head node A, and become the members of the DCS where the node A is located, as shown in Fig. 2 (1).
  • the nodes P, B, G, J, D, and C are broadcasted again, and the nodes N, E, K, H, S, I, and O are added, as shown in Fig. 2 (2). If set Then, the node within the 2-hop range can join the DCS where the aggregate head node A is located. Similarly, if F also competes as a set head node, node F does the same operation, thereby forming the DCSs shown in Fig. 2 (3).
  • the routing process is shown in Figure 3.
  • the routing protocol starts with a hop-like diffusion protocol, so that the nodes know the number of hops they have reached the base station. But unlike ordinary hop-diffusion protocols, each node has Class-aware values, peer-like perception values can be fused. Therefore, the purpose of heavy data fusion routing is to aggregate similar sensing values into the same route for data fusion. Thus each node is saved The minimum number of hops that the class-aware value reaches the base station. At the beginning, therefore, the base station will itself reach the base station with a hop count of zero. Each node sets the number of hops for each type of observation to reach the base station to be ⁇ , thus sharing Awkward.
  • the node that receives the broadcast message compares the number of hops of the kth class of the sensed value to the base station and the number of hops in the broadcast packet, if the number of hops in the broadcast packet is less than 1 of the k-type perception saved by itself. When the value reaches the hop count of the base station, the number of hops of the broadcast packet plus 1 is substituted for the saved hop count. Then, this node broadcasts the same number of hops that it has reached the base station. Thus, each node gets the minimum hop count of the kth class of perceived values arriving at the base station.
  • Figure 3 contains the routes of two types of data.
  • the solid line color is one type, and the dotted line route is one type.
  • the same type of data routing should be brought together as much as possible for integration. Data that is not homogeneous does not need to be merged.
  • FIG. Fig. 3(1) shows the minimum hop count of each type of observations arriving at the base station recorded by each node after performing the hop count broadcast of step S1, and Fig. 3 is provided with two types of perceptual data.
  • all the different sensing data of the same node arrive at the base station with the same hop value.
  • the node routes to the base station according to the category of the data using the minimum hop routing mechanism of this category.
  • the route trace shown in Fig. 3 (1).
  • the node on the route sets the hop count of its own kth class to the base station to 0.
  • the real line as shown in Fig. 3 (2) is composed of tracks. Then, the node whose hop count becomes 0 spreads out, and the node within the affected range changes its own hop count to the base station, as shown in Fig. 3 (3). Since the number of hops of the node on the route to the base station is 0, other similar data routes are attracted to the first created route, so that multiple data routes are collected on one route and sent to the base station. As shown in Figure 3 (4).
  • the method of the present invention is compared with the following: (1)
  • the first type is that each node generates data and directly transmits it to the base station, which is called a direct data collection (DDC) method in the present invention
  • DDC direct data collection
  • the second type is the literature [ Hung C C, Peng W C, Lee W C. Energy-aware set-covering approaches for approximate data collection in wireless sensor networks. IEEE Transactions on Knowledge and Data Engineering, 2012; 24(11): 1993-2007.
  • Each data coverage set uses a representative node to represent the perceived value of its set and reports directly to the base station, which is referred to as a R-node data collection (RNDC).
  • RNDC R-node data collection
  • the method of the present invention is MRRF.
  • the MRRF method proposed by the present invention can equalize the network energy consumption.
  • the energy consumption is much smaller than the DDC and RNDC methods, while in the far base station region, the energy consumption is high.
  • the DDC and RNDC methods The reason for this phenomenon is that the MRRF method uses the representative node and re-convergence routing method to make the energy consumption of the near-base station area node much lower than the other two methods.
  • the MRRF method uses a larger DCS.
  • the energy consumption of the node to create and maintain on the DCS is much larger than the other two methods, so that the energy consumption of the MRRF method in the far base station area is rather high.
  • this does not reduce the performance of the MRRF method, but makes full use of the network energy consumption and improves the energy utilization of the network.
  • the maximum energy consumption of the network of the method of the present invention is much smaller than the other two methods.
  • the maximum energy consumption of the network determines the network lifetime, that is, the method of the present invention can greatly Improve network life. They are 1.6 times to 3.2 times the network lifetime under the representative node data collection method and the direct data collection method.
  • Figure 7 shows the comparison of the maximum energy consumption under different data correlation coefficients. It can be seen from the figure that the maximum energy consumption of the network under the representative node data collection method and the direct data collection method is 2.12 to 3.73 times that of the method of the present invention, 4.24. ⁇ 7.13 times. It is indicated that the method of the invention can significantly reduce the energy consumption of a node.
  • Figure 8 shows the comparison of the remaining energy rates of the network under different methods. It can be seen from the figure that the residual energy rate of the network under the method of the invention is only about 10%, while the other two methods are as high as 70%, indicating that the method of the present invention can Use energy efficiently.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明公开了一种基于多代表节点与多层融合的异构传感数据收集方法，发明方法分为二个组成部分。1：对于每一类感知数据，在整个无线传感器网络中形成多个数据覆盖集合，同一数据覆盖集合内数据之间的差值小于规定的阀值。这样，每一个数据覆盖集合就可用一个代表节点来代表整个集合的感知值。这是第一个层面的减少网络所传输的数据量；2：代表节点的数据在向基站路由的过程中，同一类数据的代表节点会依据本发明提出的路由算法尽可能多的汇合到一条路径上，从而使同一类数据进行再次数据融合，从而再次减少需要传输到基站的数据量。因而本发明方法能够显著的提高网络寿命。

Description

一种基于多代表节点与多层融合的异构传感数据收集方法 技术领域
本发明属于无线网络异构数据收集领域，特别涉及一种能提高网络寿命的传感数据收集方法。
背景技术
无线传感器网络是由大量的彼此之间通过多跳无线链路和通信的传感器节点以自组织和多跳的方式构成的无线网络，可以广泛的运用到工业监测，农业，民用，环境监测，战场，海洋，火灾等各种特殊环境与应用中，被认为是未来的重要物联网络的关键基础技术之一。无线传感器网络研究中存在的一个至关重要的问题是如何即能够将网络感知的数据传送到基又能够提高网络的寿命。网络的寿命是无线传感器网络研究的首要问题。因而，提出的研究也格外多，下面仅介绍与本发明相关的有关研究。其背景技术论述如下：
减少节点的能量消耗，从而就能够提高网络寿命。因而节省节点能量的研究成为无线传感器网络中的一个重要的研究内容。节省能量的一个重要方法是减少参与感知的节点数量,从节点中只选择一部分节点来参与事件或者目标的监测。这样会达到如下2个目标:(1)从事件或者目标区域只选择一部分节点进行感知,从而减少了参与感知的节点数量,这样这些不参与感知的节点就可以进入睡眠状态,从而节省能量。(2)相应的减少对事件或者目标感知信息量小的节点的感知,而选择对事件或目标监测信息质量高的节点能够在不减少对事件和目标监测质量的前提下减少发往基站的数据量,从而提高网络寿命。
在已经提出的方法中,每个节点将感知的数据都独立的发往基站,因而即使选择较少的节点来监测事件与环境的情况下仍然会有较大的能量消耗。因而文献[Hung C C,Peng W C,Lee W C.Energy-aware set-covering approaches for approximate data collection in wireless sensor networks.IEEE Transactions on Knowledge and Data Engineering,2012；24(11):1993-2007.]提出了一种代表节点的方法,这种方法是通过在数据覆盖集合中选取一个称为代表节的值来代近似替整个集合所有节点的感知值的方法。因而,并不需将每个节点感知值都发往基站,因而代表节点方法可以更进一步减少发送到基站的数据量，从而可节省能量,但这要求应用系统能够容忍一定数据误差范围。
另一类重要的节省能量的方法是数据融合。无线传感器网络部署的节点由于感知的信息之间存在相关性,例如:当某事件发生后,同时有多个节点感知到事件,因而完全没有必要将每个感知事件信息的节点数据发送给基站,而是在发往基站前将这些节点感知的数据进行数 据融合,只将能够准确表达事件信息的数据提取出来,也就是数据融合后再发往基站。这样,即可以减少节点需要发送的数据量,又不减少系统所获取的信息。
但是，现有的研究存在的问题是：(1)代表节点方法中，仅有的研究只能处理单一感知值的网络，而不能处理多个感知值的网络。(2)现在提出的代表节点选择方法都非常复杂，需要消耗的能量非常多；(3)每一个代表节点都是独立的发往基站，从而能量消耗比较大。未能用到数据融合技术。导致目前的技术还存在较大的改进之处。因此，有必要设计一种异构传感数据收集方法以提高网络寿命。
发明内容
本发明所要解决的技术问题是提供一种基于多代表节点与多层融合的异构传感数据收集方法，该基于多代表节点与多层融合的异构传感数据收集方法能减少传输的数据量，提高网络寿命。
发明的技术解决方案如下：
一种基于多代表节点与多层融合的异构传感数据收集方法，在无线传感器网络中，每一个节点能传感多类感知值，比如温度、湿度、气压等不同类型的感知值，多类感知数据称为异构数据；通过如下二个方法以提高网络寿命，即将感知数据的信息在满足应用需求的前提下，尽可能使得网络寿命最大化：
方法1：针对每一类感知数据，在整个无线传感器网络中形成多个数据覆盖集合，数据覆盖集合记为DCS，每一个数据覆盖集合内节点的此类感知数据值的差值小于应用设定的阀值，每一个数据覆盖集合中设有一个头节点作为代表节点；
方法2：针对每一类感知数据，由代表节点发起向基站的路由以传输该类感知数据；同一类感知数据的路由在向基站路由的过程中，尽可能地汇集在一条路由路径上，经过数据融合后传输到基站。
方法1中，集合头节点的形成过程如下：
每个节点在0～1之间随机选择一个数，对于节点A如果选择的数小于阈值P(A)，则该节点就充当集合头节点；其中P(A)的计算如下：
Figure PCTCN2016090088-appb-000001
式中，mod为取余函数；p是节点当选为集合头节点的初始概率，依据应用的需求而事先设定；γ是目前循环进行的轮数；G是最近
Figure PCTCN2016090088-appb-000002
轮未当选为集合头节点的节点集合；Eavg和EA分别表示节点A以及其邻居节点的平均能量和节点A的当前能量。
方法1中，数据覆盖集合中的集合成员节点的形成过程如下：
(1)初始化时，每个节点设置自己到达每一类DCS集合头节点的跳数为∞，即无穷大；
(2)对应任一个集合头节点A，首先进行广播，广播消息为CMA，表示创建集合成员节点A的消息；节点A的广播消息CMA包含6项内容:
(a)集合头节点的ID,记为CMA.h；
(b)节点A自己的ID,设为CMA.ID；
(c)节点到达集合头节点A的跳数
Figure PCTCN2016090088-appb-000003
(d)限制广播在一定跳数内的值Δ,记为CMA.Δ；
(e)感知值的类型
Figure PCTCN2016090088-appb-000004
ψ是感知值的种类，
Figure PCTCN2016090088-appb-000005
表感知值的种类总数；
(f)集合头节点A的感知值,记为CMA.υ；
集合头节点A广播自己的第k类感知值的广播消息为:CMA.h＝A,CMA.ID＝A,
Figure PCTCN2016090088-appb-000006
表示广播的距离广播发起节点的最大跳数(其取值与应用相关，一般设为8),CMA.ψ＝k,CMA.υ＝xA；xA是感知值的具体数值；任何接收到广播消息的节点,如节点B,设收到的广播包为CMA，依次进行如下操作:
(I)首先读取CMA.ψ,比较自己保存的第CMA.ψ类数据到达CMA.h【集合头节点的
ID】的跳数ΗA是否小于
Figure PCTCN2016090088-appb-000007
如果是，丢弃此广播消息,不做任何操作，继续等待新的消息；
如果不是，则更新自己到达集合头节点CMA.h的跳数
Figure PCTCN2016090088-appb-000008
并设置到达第k类数据的集合头节点是CMA.h,到达集合头节点路由的下一跳是CMA.ID,继续下一步操作；
(II)将自己的第CMA.ψ类数据感知值与集合头节点CMA.h的感知值CMA.υ的值进行比较,如果
Figure PCTCN2016090088-appb-000009
则B向集合头节点A发出参加集合头节点A的加入消息JM,并继续第(III)的操作；加入消息(JM)的过程比较简单.为现有技术；收到JM消息的节点以最小路由方法向集合头节点转发JM,因而最终JM到达集合头节点,集合头节点记录 数据覆盖集合中每一个成员节点；
d(.)为差值函数，
Figure PCTCN2016090088-appb-000010
表示
Figure PCTCN2016090088-appb-000011
的差值，εk表示第k类数据感知值的差值阈值，为预先设定的已知值；
(III)如果CMA.Δ-1＝0,则所有操作结束；否则接收到消息的节点B更新广播包的内容后再向外广播进行下一个循环；更新后的广播包内容为:CMA.h＝A,CMA.ID＝B,CMA.Δ＝CMA.Δ-1,CMA.ψ＝k,CMA.υ＝xA。
在第一轮未能成为集合头节点,也未能够成为数据覆盖集合(set member)的节点A'自动成为集合头节点。
在方法2中，采用数据融合路由方法，将同类数据进行融合后，再由代表节点向汇聚节点发送数据；
所述的数据融合路由方法，包括以下步骤：
步骤S1：基站将自己到达基站的跳数为0，每个节点设置每类观察值到达基站的跳数为∞,因而共有
Figure PCTCN2016090088-appb-000012
个∞，然后,基站向外广播,收到广播消息的节点将自己第k类感知值到达基站的跳数与广播包中的跳数比较,如果广播包中的跳数加1小于自己保存的第k类感知值到达基站的跳数,则用广播包的跳数加1代替保存的跳数，然后,此节点将自己到达基站的跳数进行同样的广播，使得每个节点都得到到达基站的第k类感知值的最小跳数；
步骤S2：当第k类集合头节点有数据需要传输到基站时，采用最短路由算法，依次选取距离基站近一跳的节点作为中继节点进行路由，将数据传输到基站；
步骤S3：在步骤S2中进行第k类数据路由的节点将自己第k类到达基站的跳数设置为0；然后,跳数变为0的节点向外扩散自己到达基站的跳数，其方法与步骤S1类似，受影响范围内的节点变更自己的到达基站的跳数，使得其它同类数据路由会被吸引到先创建的路由上,从而多条数据路由会汇集在一条路由上进行数据融合后，再发往基站，通过减少所需要发送的数据量以减少能量消耗，提高了网络寿命。
在无线传感器网络中，多个节点感知的信息可能非常相近，如果节点感知的信息相差的范围在指定的阀值内，就可以用一个节点的感知信息代表与与其感知信息的差值在阀值范围内其它节点的感知值。这样就可以减少网络所需要传输的数据量提高网络寿命。
本发明中依据以上的原理更进一步提出了多代表节点与多层融合的异构数据收集方法。在本发明中，如图1所示，不仅感知值在一定阀值范围内的节点组成一个集合，该集合所有节点的感知值用一个代表节点的感知值代表，只需要将该节点的值发往基站就可以了。
与现有技术相比，本发明的优点在于：
本发明的基于多代表节点与多层融合的异构传感数据收集方法，是分布式的代表节点选择方法，只要随机的从任意一节点进行代表节点选举就可以完成，从而使算法能量消耗小，简单有效；能够处理多个感知值的情况；数据再次进行融合后发往基站，从而进一步减少了数据量，提高了网络寿命。
对比以往研究，本发明的改进如下:(a)以往的研究仅能够处理单一感知值的情况。在实际的网络中，传感节点往往搭载了多种传感器件，比如一个传感节点往往能够同时感知温度、湿度、气压等感知值。因而这种异构的感知值不能用一个代表值来代表，而是要用多个不同类型的代表值来代表，因而，本发明中对异构的不同感知值采用不同的代表值来代表；(b)放松了以往研究中需要合适的选择代表节点以减少网络传输数据的不足。在以往的方法中如果选择的代表值不同，则代表节点的个数也不相同，也就是需要传送给基站的数据包个数不同。因而往往为了选择合适的代表值而增大方法的复杂度。本发明放松了这种约束，只需要随机的指定未成为集合中的节点的值为代表值即可。因为，本发明还对代表节点传送给基站的数据进行了二次数据融合，从而一方面大大减少了传送到基站的数据量，而减少了选择代表值的复杂度。
本发明的另一个重要改进是当代表节点的值发往基站时，采用路由汇合的方法，使同一种类型感知值的路由会汇合在一起，从而进行数据融合，这样能够大大减少传送到基站的数据量，提高网络寿命。如图1所示，路由分支1，路由分支2，路由分支3的数据最后都汇合在一条路由上，从而这些同类的数据再进行数据融合以减少数据传输的数据量。而以往的方法是每一条路由都是单独的路由到基站，从而其数据传输量很大，因而本发明的方法能够大大的提高网络寿命。经过如上的二种改进，本发明方法能够简化以往的算法，扩大算法的适用范围，减少传送到基站的数据量，提高网络寿命。
发明方法分为二个组成部分。1：对于每一类感知数据，在整个无线传感器网络中形成多个数据覆盖集合，同一数据覆盖集合内数据之间的差值小于规定的阀值。这样，每一个数据覆盖集合就可用一个代表节点来代表整个集合的感知值。这是第一个层面的减少网络所传输的数据量；2：代表节点的数据在向基站路由的过程中，同一类数据的代表节点会依据本发明提出的路由算法尽可能多的汇合到一条路径上，从而使同一类数据进行再次数据融合，从而再次减少需要传输到基站的数据量。因而本发明方法能够显著的提高网络寿命。
附图说明
图1为基于多代表节点与多层融合的异构传感数据路由示意图；
图2为集合形成过程示意图；
图3为路由的形成过程示意图；
图4为在不同方法下节点承担的数据量的对比图；
图5为在不同方法下节点的能量消耗情况的对比
图6为网络的最大能量消耗对比图；
图7为在不同数据相关系数下的最大能量消耗对比图；
图8为不同方法下网络剩余能量率的对比图。
具体实施方式
为了便于理解本发明，下文将结合说明书附图和较佳的实施例对本文发明做更全面、细致地描述，但本发明的保护范围并不限于一下具体实施例。
除非另有定义，下文中所使用的所有专业术语与本领域技术人员通常理解含义相同。本文中所使用的专业术语只是为了描述具体实施例的目的，并不是旨在限制本发明的保护范围。
实施例：
如图1-3，一种基于多代表节点与多层融合的异构传感数据收集方法，在无线传感器网络中，每一个节点能传感多类感知值，比如温度、湿度、气压等不同类型的感知值，多类感知数据称为异构数据；通过如下二个方法以提高网络寿命，即将感知数据的信息在满足应用需求的前提下，尽可能使得网络寿命最大化：
方法1：针对每一类感知数据，在整个无线传感器网络中形成多个数据覆盖集合，数据覆盖集合记为DCS，每一个数据覆盖集合内节点的此类感知数据值的差值小于应用设定的阀值，每一个数据覆盖集合中设有一个头节点作为代表节点；
方法2：针对每一类感知数据，由代表节点发起向基站的路由以传输该类感知数据；同一类感知数据的路由在向基站路由的过程中，尽可能地汇集在一条路由路径上，经过数据融合后传输到基站。
方法1中，集合头节点的形成过程如下：
每个节点在0～1之间随机选择一个数，对于节点A如果选择的数小于阈值P(A)，则该节点就充当集合头节点；其中P(A)的计算如下：
Figure PCTCN2016090088-appb-000013
式中，mod为取余函数；p是节点当选为集合头节点的初始概率，依据应用的需求而 事先设定；γ是目前循环进行的轮数；G是最近
Figure PCTCN2016090088-appb-000014
轮未当选为集合头节点的节点集合；Eavg和EA分别表示节点A以及其邻居节点的平均能量和节点A的当前能量。
方法1中，数据覆盖集合中的集合成员节点的形成过程如下：
(1)初始化时，每个节点设置自己到达每一类DCS集合头节点的跳数为∞，即无穷大；
(2)对应任一个集合头节点A，首先进行广播,广播消息为CMA，表示创建集合成员节点A的消息；节点A的广播消息CMA包含6项内容:
(a)集合头节点的ID,记为CMA.h；
(b)节点A自己的ID,设为CMA.ID；
(c)节点到达集合头节点A的跳数
Figure PCTCN2016090088-appb-000015
(d)限制广播在一定跳数内的值Δ,记为CMA.Δ；
(e)感知值的类型
Figure PCTCN2016090088-appb-000016
ψ是感知值的种类，
Figure PCTCN2016090088-appb-000017
表感知值的种类总数；
(f)集合头节点A的感知值,记为CMA.υ；
集合头节点A广播自己的第k类感知值的广播消息为:CMA.h＝A,CMA.ID＝A,
Figure PCTCN2016090088-appb-000018
表示广播的距离广播发起节点的最大跳数(其取值与应用相关，一般设为8),CMA.ψ＝k,CMA.υ＝xA；xA是感知值的具体数值；任何接收到广播消息的节点,如节点B,设收到的广播包为CMA，依次进行如下操作:
(I)首先读取CMA.ψ,比较自己保存的第CMA.ψ类数据到达CMA.h【集合头节点的
ID】的跳数ΗA是否小于
Figure PCTCN2016090088-appb-000019
如果是，丢弃此广播消息,不做任何操作，继续等待新的消息；
如果不是，则更新自己到达集合头节点CMA.h的跳数
Figure PCTCN2016090088-appb-000020
并设置到达第k类数据的集合头节点是CMA.h,到达集合头节点路由的下一跳是CMA.ID,继续下一步操作；
(II)将自己的第CMA.ψ类数据感知值与集合头节点CMA.h的感知值CMA.υ的值进行比较,如果
Figure PCTCN2016090088-appb-000021
则B向集合头节点A发出参加集合头节点A的加入消息JM,并继续第(III)的操作；加入消息(JM)的过程比较简单，为现有技术；收到JM消息的节点以最小路由方法向集合头节点转发JM,因而最终JM到达集合头节点,集合头节点记录数据覆盖集合中每一个成员节点；
d(.)为差值函数，
Figure PCTCN2016090088-appb-000022
表示
Figure PCTCN2016090088-appb-000023
的差值，εk表示第k类数据感知值的差值阈值， 为预先设定的已知值；
(III)如果CMA.Δ-1＝0,则所有操作结束；否则接收到消息的节点B更新广播包的内容后再向外广播进行下一个循环；更新后的广播包内容为:CMA.h＝A,CMA.ID＝B,CMA.Δ＝CMA.Δ-1,CMA.ψ＝k,CMA.υ＝xA。
在第一轮未能成为集合头节点,也未能够成为数据覆盖集合(set member)的节点A'自动成为集合头节点。
在方法2中，采用数据融合路由方法，将同类数据进行融合后，再由代表节点向汇聚节点发送数据；
所述的数据融合路由方法，包括以下步骤：
步骤S1：基站将自己到达基站的跳数为0，每个节点设置每类观察值到达基站的跳数为∞,因而共有
Figure PCTCN2016090088-appb-000024
个∞，然后,基站向外广播,收到广播消息的节点将自己第k类感知值到达基站的跳数与广播包中的跳数比较,如果广播包中的跳数加1小于自己保存的第k类感知值到达基站的跳数,则用广播包的跳数加1代替保存的跳数，然后,此节点将自己到达基站的跳数进行同样的广播，使得每个节点都得到到达基站的第k类感知值的最小跳数；
步骤S2：当第k类集合头节点有数据需要传输到基站时，采用最短路由算法，依次选取距离基站近一跳的节点作为中继节点进行路由，将数据传输到基站；
步骤S3：在步骤S2中进行第k类数据路由的节点将自己第k类到达基站的跳数设置为0；然后,跳数变为0的节点向外扩散自己到达基站的跳数，其方法与步骤S1类似，受影响范围内的节点变更自己的到达基站的跳数，使得其它同类数据路由会被吸引到先创建的路由上,从而多条数据路由会汇集在一条路由上进行数据融合后，再发往基站，通过减少所需要发送的数据量以减少能量消耗，提高了网络寿命。
对每一类感知数据，依据本发明提出的数据覆盖集合算法形成一个一个的数据覆盖集合，每个集合由称之为覆盖集合头节点的感知值来代表此数据覆盖集合的值。覆盖集合算法运行后，每一类数据都有多个数据覆盖集合，任意拥有此类感知值的节点必定属于此类数据覆盖集合中的某一个集合。对每一类感知值运行同样的算法，这样网络中每一类感知值都由多个数据覆盖集合所包含。然后，每一个数据覆盖集合由集合的头节点发起向基站的路由，在本发明中，发起的路由具有同类路由会汇集到一起的特性，因而，在多个集合头节点在路由过程中汇集在一条路(路由路径)上，就能够进行数据融合，从而进一步减少所需要传送到基站的数据量，从而提高网络寿命。
以下的集合头节点的形成过程都是指第k类数据覆盖集合(DCS)的集合头节点形成过程. 由于其它类数据覆盖集合的形成过程与第k类数据覆盖集合成过程是类似的,因而只需要将下面的过程执行共
Figure PCTCN2016090088-appb-000025
次,就能够将网络中所有类的感知值形成数据覆盖集合.
在集合头节点的选择算法中，每个节点在0～1之间随机选择一个数，对于节点A如果选择的数小于阈值P(A)，则该节点就充当集合头节点。其中P(A)的计算如下：
Figure PCTCN2016090088-appb-000026
式中，mod为取余函数，p是节点当选为集合头节点的初始概率，γ是目前循环进行的轮数；G是最近
Figure PCTCN2016090088-appb-000027
轮未当选为集合头节点的节点集合。由式(8)可知，当选过簇头的节点在接下来的1/p轮循环中将不能成为簇头。Eavg和Ec分别表示节点A以及其邻居节点的平均能量和节点A的当前能量。
(2)集合成员节点的形成过程
在开始时,每个节点设置自己到达每一类DCS集合头节点的跳数为∞。在经过前面的操作集合头节点产生后,每一个成为集合头节点的节点进行后续操作：
下面以集合头节点A为例论述集合成员节点的形成过程。集合头节点A首先进行广播,设广播消息为CMA(表示创建集合成员节点A的消息).节点A的广播消息CMA主要包含6个主要内容:(a)集合头节点的ID,记为CMA.h.(b)节点A自己的ID,设为CMA.ID.(c)节点到达集合头节点A的跳数
Figure PCTCN2016090088-appb-000028
(d)限制广播在一定跳数内的值Δ,记为CMA.Δ.(e)数据的类型
Figure PCTCN2016090088-appb-000029
集合头节点A的感知值,记为CMA.υ.这样集合头节点A广播自己的第k类感知值的广播消息为:CMA.h＝A,CMA.ID＝A,
Figure PCTCN2016090088-appb-000030
CMA.ψ＝k,CMA.υ＝xA.任何接收到广播消息的节点,比如节点B,设收到的广播包为CMA.依次进行如下操作:
(I)首先读取CMA.ψ,比较自己保存的第CMA.ψ类数据到达CMA.h的跳数ΗA是否小于
Figure PCTCN2016090088-appb-000031
如果不是,则更新自己到达集合头节点CMA.h的跳数
Figure PCTCN2016090088-appb-000032
并设置到达第k类数据的集合头节点是CMA.h,到达集合头节点路由的下一跳是CMA.ID,继续第(II)操作.否则,丢弃此广播消息,不做任何操作。
(II)将自己的第CMA.ψ类数据感知值与集合头节点CMA.h的感知值CMA.υ的值进行比 较,如果
Figure PCTCN2016090088-appb-000033
则B向集合头节点A发出参加集合头节点A的加入消息(JM),并继续第(III)的操作.加入消息(JM)的过程比较简单.收到JM消息的节点以最小路由方法向集合头节点转发JM,因而最终JM到达集合头节点,集合头节点记录数据覆盖集合每一个成员节点。
(III)如果CMA.Δ-1＝0,则所有操作结束。否则更新广播包的内容后再向外广播。新的广播包内容为:CMA.h＝A,CMA.ID＝B,CMA.Δ＝CMA.Δ-1,CMA.ψ＝k,CMA.υ＝xA。
数据集合成员的形成过程如图2所示.设节点A竞争中成为集合头节点,则节点A向外广播,广播范围内的节点接收到节点A的广播信息后,如果发现自己是属于与A的感知值的距离在误差范围内,则向集合头节点A申请加入。这时节点P,B,G,J,D加入集合头节点A,成为节点A所在DCS的成员(member),如图2(1)所示。同样,节点P,B,G,J,D,C再向外广播,使节点N,E,K,H,S,I,O加入,如图2(2)所示。如果设置
Figure PCTCN2016090088-appb-000034
则2跳范围内节点都可以加入集合头节点A所在的DCS。同样,如果F也竞争为集合头节点的话,则节点F也做同样的操作,从而形成图2(3)所示的DCSs。
路由过程如图3所示,路由协议开始时采用类似于跳数扩散协议,目的是使节点都知道自己到达基站的跳数。但与普通跳数扩散协议不相同的是,每个节点有
Figure PCTCN2016090088-appb-000035
类感知值,同类感知值才能融合,因而,重数据融合路由的目的是想使同类的感知值汇聚到同一条路由以便于数据融合。因而每个节点保存有
Figure PCTCN2016090088-appb-000036
类感知值到达基站的最小跳数。刚开始时,因而,基站将自己到达基站的跳数为0。每个节点设置每类观察值到达基站的跳数为∞,因而共有
Figure PCTCN2016090088-appb-000037
个∞。然后,向外广播,收到广播消息的节点将自己第k类感知值到达基站的跳数与广播包中的跳数比较,如果广播包中的跳数加1小于自己保存的第k类感知值到达基站的跳数,则用广播包的跳数加1代替保存的跳数。然后,此节点将自己到达基站的跳数进行同样的广播。这样,每个节点都得到到达基站的第k类感知值的最小跳数。我们通过图3给出了重数据融合路由的建立过程。设网络中数据的种类
Figure PCTCN2016090088-appb-000038
图3(1)给出了在进行跳数广播后,每个节点记录的到达基站的每类观察值的最小跳数,这时同一个节点所有类感知值到达基站的值相同。然后,如果集合头节点有数据发送到基站,则节点依据数据的类别采用此类别的最小跳数路由机制路由到基站。这时如图3(1)所示的路由轨迹。与以往路由机制不同的是,在本发明的重数据融合路由方法中,此路由上的节点将自己第k类到达基站的跳数设置为0。如图3(2)所示的黑色路由轨迹。然后,跳数变为0的节点向外扩散,受影响范围内的节点变更自己的到达基站的跳数,如图3(3)所示。由于此条路由上节点到达基站的跳数为0,因而其它同类数据路由就会被吸引到最先创建的路由上,从而多条数据路由会汇集在一条路由上发往基站。如图3(4)所示，圆圈 中的数值如4/4代表到达基站的跳数，每个圈有二个数字，分别表示二类数据到到达基站的跳数。
图3中包含有两类数据的路由，实线色的是一类，点线的路由是一类。同一类数据路由要尽量汇集到一起，以便融合。而不同类的数据则不需要汇合。
通过图3来说明路由过程。图3(1)给出了在进行步骤S1的跳数广播后,每个节点记录的到达基站的每类观察值的最小跳数,图3设有两类感知数据。这时同一个节点所有不同感知数据到达基站的跳数值相同。然后,如果有集合头节点有数据发送到基站,则节点依据数据的类别采用此类别的最小跳数路由机制路由到基站。这时如图3(1)所示的路由轨迹。与以往路由机制不同的是,在本发明的方法2中,此路由上的节点将自己第k类到达基站的跳数设置为0。如图3(2)所示的实线路由轨迹。然后,跳数变为0的节点向外扩散,受影响范围内的节点变更自己的到达基站的跳数,如图3(3)所示。由于此条路由上节点到达基站的跳数为0,因而其它同类数据路由就会被吸引到先创建的路由上,从而多条数据路由会汇集在一条路由上发往基站。如图3(4)所示。
本发明的方法与之比较的方法是:(1)第1种是每个节点产生数据后直接向基站发送,在本发明中称为直接数据收集方法(direction data collection,DDC)[参见文献：Liu A,Ren J,Li X,et al.Design principles and improvement of cost function based energy aware routing algorithms for wireless sensor networks.Computer Networks,2012,56(7):1951-1967].第2种是文献[Hung C C,Peng W C,Lee W C.Energy-aware set-covering approaches for approximate data collection in wireless sensor networks.IEEE Transactions on Knowledge and Data Engineering,2012；24(11):1993-2007.]的方法,每个数据覆盖集合用一个代表节点代表其集合的感知值,并直接向基站报告,称这种方法为代表节点数据收集方法(R-node data collection,RNDC).本发明方法为MRRF。
从图4与5的实验结果可以看出本发明提出的MRRF方法能够均衡网络能量消耗,在近基站区域,其能量消耗远小于DDC与RNDC方法,而在远基站区域时,其能量消耗却高于DDC与RNDC方法。造成这种现象的原因是:MRRF方法由于采用代表节点与重融合路由方法从而使得近基站区域节点的能量消耗大大低于其它2个方法.而在远基站区域,MRRF方法采用较大的DCS,从而节点花费在DCS上创建与维护的能量消耗远大于其它二种方法,从而在远基站区域MRRF方法的能量消耗反而较高。但这并没有降低MRRF方法的性能，反而充分利用了网络能量消耗,提高了网络的能量利用率。
从图6可以看出，本发明方法相对于其它二种方法，网络的最大能量消耗远小于其它二种方法。而网络的最大能量消耗就决定了网络寿命，也就是说，本发明的方法能够大幅度 的提高网络寿命。分别是代表节点数据收集方法、直接数据收集方法下网络寿命的1.6倍到3.2倍。
图7给出了在不同数据相关系数下的最大能量消耗对比，从图中可以看出：代表节点数据收集方法和直接数据收集方法下网络最大能量消耗是本发明方法的2.12～3.73倍，4.24～7.13倍。说明本发明方法能够显著的减少节点的能量消耗。
图8给出了不同方法下网络剩余能量率的对比，从图中可以看出，本发明方法下网络剩余能量率仅为10％左右，而其它二种方法高达70％，说明本发明方法能够有效利用能量。

Claims (5)

  1. 一种基于多代表节点与多层融合的异构传感数据收集方法，在无线传感器网络中，每一个节点能传感多类感知值，其特征在于，通过如下二个方法以提高网络寿命：
    方法1：针对每一类感知数据，在整个无线传感器网络中形成多个数据覆盖集合，数据覆盖集合记为DCS，每一个数据覆盖集合内节点的此类感知数据值的差值小于应用设定的阀值，每一个数据覆盖集合中设有一个头节点作为代表节点；
    方法2：针对每一类感知数据，由代表节点发起向基站的路由以传输该类感知数据；同一类感知数据的路由在向基站路由的过程中，尽可能地汇集在一条路由路径上，经过数据融合后传输到基站。
  2. 根据权利要求1所述的基于多代表节点与多层融合的异构传感数据收集方法，其特征在于，方法1中，集合头节点的形成过程如下：
    每个节点在0～1之间随机选择一个数，对于节点A如果选择的数小于阈值P(A)，则该节点就充当集合头节点；其中P(A)的计算如下：
    Figure PCTCN2016090088-appb-100001
    式中，mod为取余函数；p是节点当选为集合头节点的初始概率，依据应用的需求而事先设定；γ是目前循环进行的轮数；G是最近
    Figure PCTCN2016090088-appb-100002
    轮未当选为集合头节点的节点集合；Eavg和EA分别表示节点A以及其邻居节点的平均能量和节点A的当前能量。
  3. 根据权利要求1所述的基于多代表节点与多层融合的异构传感数据收集方法，其特征在于，方法1中，数据覆盖集合中的集合成员节点的形成过程如下：
    (1)初始化时，每个节点设置自己到达每一类DCS集合头节点的跳数为∞，即无穷大；
    (2)对应任一个集合头节点A，首先进行广播,广播消息为CMA，表示创建集合成员节点A的消息；节点A的广播消息CMA包含6项内容:
    (a)集合头节点的ID,记为CMA.h；
    (b)节点A自己的ID,设为CMA.ID；
    (c)节点到达集合头节点A的跳数
    Figure PCTCN2016090088-appb-100003
    (d)限制广播在一定跳数内的值Δ,记为CMA.Δ；
    (e)感知值的类型
    Figure PCTCN2016090088-appb-100004
    ψ是感知值的种类，
    Figure PCTCN2016090088-appb-100005
    表感知值的种类总 数；
    (f)集合头节点A的感知值,记为CMA.υ；
    集合头节点A广播自己的第k类感知值的广播消息为:CMA.h＝A,CMA.ID＝A,
    Figure PCTCN2016090088-appb-100006
    Figure PCTCN2016090088-appb-100007
    表示广播的距离广播发起节点的最大跳数，CMA.ψ＝k,CMA.υ＝xA；xA是感知值的具体数值；任何接收到广播消息的节点,如节点B,设收到的广播包为CMA，依次进行如下操作:
    (Ⅰ)首先读取CMA.ψ,比较自己保存的第CMA.ψ类数据到达CMA.h的跳数ΗA是否小于
    Figure PCTCN2016090088-appb-100008
    如果是，丢弃此广播消息,不做任何操作，继续等待新的消息；
    如果不是，则更新自己到达集合头节点CMA.h的跳数
    Figure PCTCN2016090088-appb-100009
    并设置到达第k类数据的集合头节点是CMA.h,到达集合头节点路由的下一跳是CMA.ID,继续下一步操作；
    (Ⅱ)将自己的第CMA.ψ类数据感知值与集合头节点CMA.h的感知值CMA.υ的值进行比较，如果
    Figure PCTCN2016090088-appb-100010
    则B向集合头节点A发出参加集合头节点A的加入消息JM，并继续第(III)的操作；收到JM消息的节点以最小路由方法向集合头节点转发JM，因而最终JM到达集合头节点，集合头节点记录数据覆盖集合中每一个成员节点；
    d(.)为差值函数，
    Figure PCTCN2016090088-appb-100011
    表示
    Figure PCTCN2016090088-appb-100012
    的差值，εk表示第k类数据感知值的差值阈值，为预先设定的已知值；
    (III)如果CMA.Δ-1＝0,则所有操作结束；否则接收到消息的节点B更新广播包的内容后再向外广播进行下一个循环；更新后的广播包内容为:CMA.h＝A,CMA.ID＝B,CMA.Δ＝CMA.Δ-1,CMA.ψ＝k,CMA.υ＝xA。
  4. 根据权利要求3所述的基于多代表节点与多层融合的异构传感数据收集方法，其特征在于，在第一轮未能成为集合头节点,也未能够成为数据覆盖集合(set member)的节点A'自动成为集合头节点。
  5. 根据权利要求1-4任一项所述的基于多代表节点与多层融合的异构传感数据收集方法，其特征在于，在方法2中，采用数据融合路由方法，将同类数据进行融合后，再由代表节点向汇聚节点发送数据；
    所述的数据融合路由方法，包括以下步骤：
    步骤S1：基站将自己到达基站的跳数为0，每个节点设置每类观察值到达基站的跳数为 ∞，因而共有
    Figure PCTCN2016090088-appb-100013
    个∞，然后,基站向外广播，收到广播消息的节点将自己第k类感知值到达基站的跳数与广播包中的跳数比较,如果广播包中的跳数加1小于自己保存的第k类感知值到达基站的跳数，则用广播包的跳数加1代替保存的跳数，然后，此节点将自己到达基站的跳数进行同样的广播，使得每个节点都得到到达基站的第k类感知值的最小跳数；
    步骤S2：当第k类集合头节点有数据需要传输到基站时，采用最短路由算法，依次选取距离基站近一跳的节点作为中继节点进行路由，将数据传输到基站；
    步骤S3：在步骤S2中进行第k类数据路由的节点将自己第k类到达基站的跳数设置为0；然后，跳数变为0的节点向外扩散自己到达基站的跳数，受影响范围内的节点变更自己的到达基站的跳数，使得其它同类数据路由会被吸引到先创建的路由上，从而多条数据路由会汇集在一条路由上进行数据融合后，再发往基站。
PCT/CN2016/090088 2016-03-31 2016-07-15 一种基于多代表节点与多层融合的异构传感数据收集方法 WO2017166539A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610196665.3A CN105898679B (zh) 2016-03-31 2016-03-31 一种基于多代表节点与多层融合的异构传感数据收集方法
CN201610196665.3 2016-03-31

Publications (1)

Publication Number Publication Date
WO2017166539A1 true WO2017166539A1 (zh) 2017-10-05

Family

ID=57011727

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/090088 WO2017166539A1 (zh) 2016-03-31 2016-07-15 一种基于多代表节点与多层融合的异构传感数据收集方法

Country Status (2)

Country Link
CN (1) CN105898679B (zh)
WO (1) WO2017166539A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111417084A (zh) * 2020-03-26 2020-07-14 仲恺农业工程学院 农情获取任务的分配方法、系统、介质和设备
CN111800284A (zh) * 2019-04-08 2020-10-20 阿里巴巴集团控股有限公司 边缘云节点集合的选择方法和装置以及电子设备
CN114554566A (zh) * 2022-01-27 2022-05-27 哈尔滨海能达科技有限公司 无线链路的路由方法、装置、电子设备及存储介质
CN115695289A (zh) * 2023-01-03 2023-02-03 顺霆科技(无锡)有限公司 一种异构多跳拓展网络
CN115714770A (zh) * 2022-09-23 2023-02-24 中南大学 一种云－边协同的数据服务化网络均衡计算方法

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107733875A (zh) * 2017-09-21 2018-02-23 深圳市盛路物联通讯技术有限公司 一种数据传输方法及系统
CN110557185B (zh) * 2018-05-31 2022-08-19 成都鼎桥通信技术有限公司 一种中继设备的运行模式确定方法和装置

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101415011A (zh) * 2008-10-31 2009-04-22 北京工业大学 无线传感器网络中安全有效的数据聚合方法
CN201426183Y (zh) * 2009-05-22 2010-03-17 广东工业大学 一种基于簇的无线传感器网络路由系统

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100943178B1 (ko) * 2007-12-17 2010-02-19 한국전자통신연구원 무선 센서 네트워크에서의 라우팅 및 자원 할당 방법
CN101877827B (zh) * 2010-05-18 2012-12-12 浙江工业大学 面向楼宇监测的无线传感器网络数据收集方法
CN101917750B (zh) * 2010-07-30 2012-12-05 南京工业大学 最小努力交付的跨区虚拟簇多跳路由方法
CN102802175A (zh) * 2012-02-24 2012-11-28 东北电力大学 一种基于能量消耗的无线传感器网络路由优化算法
CN104411000A (zh) * 2014-12-15 2015-03-11 南昌航空大学 一种无线传感器网络中分层路由协议簇头选择方法

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101415011A (zh) * 2008-10-31 2009-04-22 北京工业大学 无线传感器网络中安全有效的数据聚合方法
CN201426183Y (zh) * 2009-05-22 2010-03-17 广东工业大学 一种基于簇的无线传感器网络路由系统

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
FENG, CHENWEI: "Data Transmission Strategy for Wireless Sensor Network Based on Virtual MIMO", JOURNAL OF LUJIANG UNIVERSITY, vol. 23, no. 3, 30 June 2015 (2015-06-30), pages 52 - 53, ISSN: 1673-4432 *
HUNG, CHIHCHIEH ET AL.: "Energy-Aware Set-Covering Approaches for Approximate Data Collection in Wireless Sensor Networks", IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, vol. 24, no. 11, 30 November 2012 (2012-11-30), pages 1993 - 2007, XP011463875, ISSN: 1041-4347 *
WANG, KANG. ET AL.: "The Research on Data Fusion Algorithms Based on Clustering Routing Protocol in WSN", CHINA MASTER'S THESES FULL-TEXT DATABASE (ELECTRONIC PERIODICAL), vol. 1140, no. 02, 15 February 2015 (2015-02-15), ISSN: 1674-0246 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111800284A (zh) * 2019-04-08 2020-10-20 阿里巴巴集团控股有限公司 边缘云节点集合的选择方法和装置以及电子设备
CN111417084A (zh) * 2020-03-26 2020-07-14 仲恺农业工程学院 农情获取任务的分配方法、系统、介质和设备
CN111417084B (zh) * 2020-03-26 2022-07-12 仲恺农业工程学院 农情获取任务的分配方法、系统、介质和设备
CN114554566A (zh) * 2022-01-27 2022-05-27 哈尔滨海能达科技有限公司 无线链路的路由方法、装置、电子设备及存储介质
CN115714770A (zh) * 2022-09-23 2023-02-24 中南大学 一种云－边协同的数据服务化网络均衡计算方法
CN115714770B (zh) * 2022-09-23 2024-04-12 中南大学 一种云－边协同的数据服务化网络均衡计算方法
CN115695289A (zh) * 2023-01-03 2023-02-03 顺霆科技(无锡)有限公司 一种异构多跳拓展网络

Also Published As

Publication number Publication date
CN105898679B (zh) 2019-09-20
CN105898679A (zh) 2016-08-24

Similar Documents

Publication Publication Date Title
WO2017166539A1 (zh) 一种基于多代表节点与多层融合的异构传感数据收集方法
WO2017101575A1 (zh) 一种无线自组网的路由方法及装置
US9788257B2 (en) Method and system for dynamically forming service aware bluetooth low energy (BLE) mesh network
US9450830B2 (en) Node apparatus and communication method
KR102038343B1 (ko) 무선 메시 네트워크에서의 액세스 포인트의 집중 및 자동 디스에이블을 위한 시스템 및 방법
US10827410B2 (en) System, apparatus and method for robust transmission of visual data through visual sensor network
CN105794284B (zh) 使节点同步至数据宿的方法以及用于形成无线网络的节点
KR101055416B1 (ko) 무선 센서 네트워크에서의 라우팅 경로 설정 방법 및 이를 수행하기 위한 장치
CN108337166A (zh) 一种面向航空集群网络的低时延高可靠路由协议
KR101079137B1 (ko) 무선 센서 네트워크에서의 패밀리-기반 주소 부여 방법과, 이를 이용한 계층적 라우팅 경로설정 방법
CN107612914B (zh) 一种基于参考节点策略的Ad Hoc网络安全信任方法
CN106304229B (zh) 一种基于内容存储的软件定义无线Mesh网络的方法
CN106713177B (zh) 一种多控制器的wmSDN组网方法
CN109639588B (zh) 一种面向航空集群的网络拥塞控制路由方法
CN110996370A (zh) 一种无人机的网络通讯路由器协议方法
CN106656795B (zh) 一种无线传感器执行器网络分簇路由选择方法
US20120320812A1 (en) Method for Associating Time Slots with Links Between Network Nodes of a Wireless Interconnected Network
CN102083164A (zh) 基于能量感知的无线传感器网络机会路由方法
CN110691380A (zh) 一种降低软件定义无人机集群网络通信开销与时延的方法
Del-Valle-Soto et al. An efficient multi-parent hierarchical routing protocol for WSNs
CN107222901B (zh) 基于信道分配的认知无线网络路由协议的实现方法
Li et al. Reliable ant colony routing algorithm for dual‐channel mobile ad hoc networks
CN109510656A (zh) 一种适用于无人机群体的自适应数据分发方法
WO2018098749A1 (zh) 分布式网络的消息广播方法及节点
Tan et al. A distributed and dynamic data gathering protocol for sensor networks

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16896311

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205 DATED 04/02/2019)

122 Ep: pct application non-entry in european phase

Ref document number: 16896311

Country of ref document: EP

Kind code of ref document: A1