CN101471865A - Method for high-efficiency data fusion in wireless sensor network - Google Patents
Method for high-efficiency data fusion in wireless sensor network Download PDFInfo
- Publication number
- CN101471865A CN101471865A CNA2007103045803A CN200710304580A CN101471865A CN 101471865 A CN101471865 A CN 101471865A CN A2007103045803 A CNA2007103045803 A CN A2007103045803A CN 200710304580 A CN200710304580 A CN 200710304580A CN 101471865 A CN101471865 A CN 101471865A
- Authority
- CN
- China
- Prior art keywords
- node
- fusion
- time
- network
- node set
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Images
Landscapes
- Mobile Radio Communication Systems (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses an efficient data fusion method for a wireless sensor network (WSN), which comprises the following steps: establishing an adjustable node set and setting an initial probing step size; calculating the fusion contributions of all elements in the adjustable node set according to the probing step size; selecting a node corresponding to the maximal value of fusion contribution; judging whether the total fusion time-lag of the node meets a time-lag restriction; adjusting node fusion waiting time or deleting nodes which do not meet the time-lag restriction from the adjustable node set; repeating the procedures until the adjustable node set is null; stopping adjusting the fusion waiting time; issuing the fusion waiting time of the node; and conducting data fusion. The method has low complexity and message communication traffic, ensures real-time property of the network, effectively reduces energy consumption of the node, prolongs the life cycle of the network, remarkably improves fusion effect, and is suitable for different topological networks.
Description
Technical field
The present invention relates to a kind of data fusion method, relate in particular to a kind of method for high-efficiency data fusion in wireless sensor network, belong to the wireless sensor network field.
Background technology
Wireless sensor network is made up of a large amount of sensing nodes, and the mode hop-by-hop by radio communication between node transmits information, the task of finishing institute's coverage area information collection of cooperation.Information has certain real-time, promptly requires to be sent in the regular hour user.The node of sensor network uses powered battery usually, must reduce the consumption of energy as far as possible.Data fusion method is the redundancy that transit node utilizes information, many parts of information vias handle obtain a or a spot of information to reduce the method for cost on network communication.The node of carrying out data fusion is not to transmit to next-hop node immediately, but this metadata cache a period of time, is called time of fusion after receiving a piece of data; When the time of fusion end, this node obtains all data of receiving portion or a spot of information and is sent to next-hop node in time of fusion after handling.Usually, time of fusion is long more, and the data that obtain in this time period are just many more, just can more significant reduction data volume, the just data traffic of the more effective reduction network of energy after the data processing; Otherwise time of fusion is short more, and the data of acquisition are few more, and the effect of fusion is just poor more, and the reduction of data traffic is just not obvious more.Yet data need to arrive (center) aggregation node via the transmission of multi-hop, and the time of fusion of transit node increases, and can influence the real-time of data, therefore must consider to realize data fusion efficiently under data are delayed time affined condition as far as possible.
Along information transmission path, it is the principal element that influences syncretizing effect that the time of fusion of transit node is distributed, and is the key problem of integration technology.People such as U Roedig have proposed a kind of centralized violence exhaust algorithm (Bu Lute FOX algorithm in 2004, Brute-Force) (reference: U Roedig, A Barroso, CJ Sreenan.Determination of Aggregation Pointsin Wireless Sensor Networks.In:Proceedings of the 30th Euromicro Conference (EUROMICRO), 2004.503-510.), this method calculates that each node carries out the probability of data fusion and the yield value that is produced on the transmission path of all possible time-delay allocative decision correspondence, and the time-delay allocative decision of yield value maximum is institute and asks.Yet owing to adopt similar exhaustive algorithm, its computation complexity is very high and be difficult to realize.This article has proposed a kind of heuritic approach of simplification immediately, is defined in every route all possible delay time is concentrated on the node, is equivalent to a kind of greedy algorithm of coarseness.With the heuritic approach of calling this simplification in the following text is DAP (Determination of AggregationPoint) data fusion method.This algorithm is the probabilistic information of foundation calculating still, but only time of fusion is assigned on the fixing point for the route of every leaf node to the Sink node, has reduced algorithm complex, but has reduced arithmetic accuracy and syncretizing effect simultaneously.In addition, they have provided two kinds of ordinary data fusion time allocation method used thereins, i.e. average distribution system EDA (EvenDistributed Algorithm) and press level exponential decrease method HDA (Hierarchy Distributed Algorithm).The mean allocation method is that all nodes distribute identical time of fusion, equals total time-delay and retrains divided by total number of plies.The exponential decrease method is to distribute maximum time of fusion near the node at center, and the time of fusion of the node in the outer more level is more little, is 2 negative exponent time power and successively decreases.
People such as Jae Young Choi proposed a kind of distributed allocation method ATC (time of fusion control algolithm of consulting formula in 2006, Aggregation Time Control) (reference: Jae Young Choi, Sunghyun Choi, Wook Hyun Kwon, and Hong Seong Park.Aggregation Time Control Algorithm for Time constrained Data Deliveryin Wireless Sensor Networks.In:Proceedings of IEEE VTC 2006-Spring, Melboume, Australia, May 7-10,2006.).In the network application process, each sensor node is according to certain rule, the time of autonomous increase transmission delay.Aggregation node then is responsible for checking whether each grouping that receives is overtime.In case find that grouping is overtime, aggregation node starts time-out information broadcasting, makes node corresponding reduce delay time.This process constantly repeats, up to arriving stable state.It is overtime that this strategy can effectively suppress packet, but its syncretizing effect is general, and when can reach the also very difficult control of stable state.The state transition diagram of this algorithm as shown in Figure 1.
The problem that the data fusion method of therefore present wireless sensor network mainly exists is that syncretizing effect is poor, the deficiency that data processing complex is high.
Summary of the invention
The present invention will propose a kind of method for high-efficiency data fusion in wireless sensor network under the condition that guarantees the time-delay constraint, the data processing complex degree of the inventive method is lower, and syncretizing effect has clear improvement.
The distribution method of time of fusion is a kind of centralized time of fusion distribution method based on the fusion contribution degree in the method that the present invention proposes, be called CFDD (Contribution-based Fusion Delay Distribution), by the flow of periodic data collection type network is analyzed, we have quantized the flow distribution of network after the application data convergence strategy, and fully taken into account the alternate position spike opposite sex and internodal influence each other of each node in routing tree, proposed to merge contribution degree and be described as the percentage contribution of specified node distribution time of fusion the whole network syncretizing effect.And then the time of fusion assignment problem is summed up as nonlinear polynary optimization problem, and the times of utilizing the climbing method of fixed step size to carry out many wheels distribute, solved in the periodic data acquisition applications, under time-delay constraints, data fusion problem efficiently.
This programme execution that periodically circulates, each cycle comprises 2 stages, is called initial phase and actual motion stage.Finish the collection of network topological information, the calculating of node time of fusion and the issue of result of calculation at initial phase.In the actual motion stage, node is carried out the data mixing operation according to the result of calculation of time of fusion.
The core of this programme is the calculating of node time of fusion.The climbing method of the fixed step size of the similar function of many variables optimization problem of time of fusion distribution method that the present invention proposes is promptly constantly selected optimum probing direction, up to the maximum that reaches target function.To the last bound constrained of data time-delay, be divided into some parts by exploring step number, every part length is for exploring step-length.Be assumed to be all nodes in the time of fusion distribution method of the present invention at first and distribute zero propagation, progressively choose subsequently and may improve maximum node the network integration, and attempt this node is increased a time of fusion of exploring step-length, if this node still satisfies the time-delay constraint after increasing, then the time of fusion of this exploration step-length is distributed to this node, otherwise no longer increase time of fusion for this node.Repeat said process, finish up to distributing.The present invention has defined the index that merges contribution degree, distributes the influence of time to network integration gain with quantitative sign node.It is short more to explore step-length, and the precision of normally result is high more, and effect is good more, but operand is big more; Otherwise precision reduces deleterious, but operand reduces.Exploring step-length and weigh according to node processing ability and network demand, is the parameter of presetting.
The above analysis, technical scheme of the present invention is:
A kind of method for high-efficiency data fusion in wireless sensor network the steps include:
1) routing iinformation of aggregation node collection network node and set to explore the step-length next-hop node of each node (be number);
2) set up the time of fusion that to adjust node set and initialization node;
3) can adjust element in the node set to all, calculate it and merge contribution;
4) try to achieve maximum and the corresponding node that node merges contribution degree;
5) total fusion of calculating whole paths at this node place is delayed time,
If the time-delay constraint is satisfied in total fusion time-delay, then step-length is explored in one of the time of fusion of this node increase, forwards step 3) to; Otherwise this node can be adjusted the node set rejecting certainly and judge whether can adjust node set is empty;
6) if can adjust node set for empty, then finish to explore the step-length adjustment, the time of fusion of publisher node is carried out data fusion; Otherwise return step 3).
The method of the routing iinformation of described aggregation node collection network node is:
1) aggregation node is at first initiated route foundation by broadcast query message,
2) sensor node is received behind the query messages according to routing rule echo reply message, and by information method incidentally the routing state information of self is communicated to described aggregation node;
3) aggregation node stores the route topological structural information of the whole network of obtaining, uses in order to subsequent step.
Described routing iinformation comprises route level, father and son's nodal information.
Setting the method for exploring step-length in the described method is: five equilibrium is carried out in the constraint of will delaying time, and its quantity is called the exploration step number, and every part time span is called the exploration step-length.
The route number of plies of described exploration step number and network design is at the same order of magnitude.
Described exploration step number equals the network route number of plies.
The computing formula of described fusion contribution is:
In the formula, i, j, k ..., be the path of node i to aggregation node; Ui is the input flow rate of node i, and Ti is that the fusion of node i distributes the time.
The described time of fusion of node of adjusting in the node set is initialized as zero.
Distribute and described routing iinformation according to described time of fusion in the described method, calculate the described fusion contribution of adjusting node in the node set.
The invention has the beneficial effects as follows:
The method that the present invention proposes is considered different nodes residing alternate position spike opposite sex in routing tree emphatically, and adopt complexity hill-climbing algorithm lower, progressively refinement that time of fusion is distributed, promptly guaranteed the network real-time, greatly improve syncretizing effect again, and be applicable to the network of different topology.Compare with technology in the past, the fusion gain of notebook data fusion method is significantly improved, and specifically: to identical network configuration, under different time-delay constraints, this method has optimum fusion gain, as shown in Figure 2; Under identical time-delay constraint, for different network sizes, this method has optimum fusion gain, as shown in Figure 3; Under time-delay constraint and all identical condition of network size, when heterogeneous networks was disposed, this method had good network adaptability (wherein, A, B, C, D, E distinguish six kinds of network designs at random), as shown in Figure 4.The raising of merging gain shows that the information communication amount of network reduces, and can effectively reduce the energy consumption of node, prolongs the life cycle of network.
Following table has provided the time of fusion distribution method of the present invention's proposition and the pluses and minuses of typical algorithm in the past compare:
| The DAP algorithm | The ATC algorithm | The CFDD algorithm | |
| The time method of salary distribution | Centralized | Distributed | Centralized |
| Precision is adjusted mode | Adjust the time discrete degree | Do not have | Adjust step-length |
| Computation complexity | High | Low | Low |
| Implementation complexity | Low | High | Low |
| Merge the gain index | Better | Generally | Good |
| Network real-time index | Better | Good | Good |
| Network topology adaptability | Better | Good | Good |
Description of drawings
Fig. 1 is the state transition diagram of ATC algorithm;
Fig. 2 compares for the fusion gain between the different fusion methods under the different delayed time constraint;
Fig. 3 is that the fusion gain of different fusion methods under the heterogeneous networks scale compared;
Fig. 4 is the adaptation situation that different fusion methods are disposed heterogeneous networks;
Fig. 5 is the execution phase of data fusion method of the present invention.
Embodiment
Below in conjunction with accompanying drawing 5 and embodiment the present invention is described in further detail: data fusion method of the present invention is divided into a plurality of fixed length cycle with the network operation phase, and each cycle is divided into initial phase and actual motion stage.Finish the collection of network routing iinformation, the distribution of node time of fusion and the distribution of result of calculation at initial phase.In the actual motion stage, node only needs to merge wait with reference to allocation result and gets final product.In the netinit stage, aggregation node (SINK) is at first initiated route by broadcast query message (REQUEST message) and is set up, sensor node is received after the REQUEST message according to certain routing rule echo reply message (RESPONSE message), and by information method incidentally the routing state information of self (as the route level, father and son's nodal information etc., wherein the father node information of node is the most basic) be communicated to the SINK node, and make up the route topological structure of the whole network successively, realize the collection of routing iinformation, and set and explore step-length.This method is periodically carried out the dynamic change that can be more suitable in network topology.
The calculating of time of fusion is the core content of the inventive method.Time of fusion is progressively calculated, and the time-delay constraint is divided into some equal parts, and its quantity is called the exploration step number, and every part time span is called explores step-length (promptly merging step-length).This method is to be that unit progressively appends distribution to the time of fusion of node to explore step-length.Concrete allocation step is stated as follows now:
● STEP-1: initialization: set up and can adjust node set, when initial all nodes in the network are all joined and can adjust in the node set, and the time of fusion of all nodes of initialization is 0
● STEP-2: can adjust element in the node set to all,, calculate it and merge contribution according to the routing iinformation that current time of fusion is distributed and obtained
● STEP-3: try to achieve maximum and corresponding node that node merges contribution degree
● STEP-4: total fusion time-delay of calculating whole paths at this node place, after if this node is adjusted, the time-delay constraint is still satisfied in total fusion time-delay of any paths at this place, then accept this adjustment (time of fusion that promptly increases this node is former distribution time and exploration step-length sum), repeat STEP-2 to STEP-5; Otherwise the adjustment to this node can not be satisfied delay requirement, this node can be adjusted the node set rejecting certainly and judge whether can adjust node set is empty.
● STEP-5:, be then algorithm end of sky if can adjust node set, aggregation node is broadcasted result of calculation (being the time of fusion of each node correspondence) to the whole network; Otherwise return STEP-2.
Wherein, the computing formula of the fusion contribution degree of node is:
In the formula, node i is i to the path of aggregation node, j, and k ..., SINK; Ui is the input flow rate of node i, and Ti is that the fusion of node i distributes the time.In the formula, the input flow rate U of node
iComprise two parts, promptly the flow of self image data and transmit the flow of other node datas promptly has
To the network that periodic data is gathered, the flow S of node self image data
iBe fixing and identical, depend on sampling period and phase data volume weekly.
It is the output flow sum of all child nodes of node i.To arbitrary node i, its output flow V
i=U
i/ (1+T
iU
i).Because network topology is known, so the I/O traffic of all nodes can be according to the time of fusion decision of current distribution, and then by merging the fusion contribution degree that the computing formula of contributing obtains node.
In this programme, if it is too small to explore step-length, then the time of implementation of algorithm can increase, but effect can be better; Otherwise then algorithm is carried out soon, but the effect possible deviation.In the application of reality, explore the step number maximum route number of plies of making peace greatly and quite can have effect preferably.
Claims (9)
1. a method for high-efficiency data fusion in wireless sensor network the steps include:
1) step-length is explored in the routing iinformation of aggregation node collection network node and setting;
2) set up the time of fusion that to adjust node set and initialization node;
3) can adjust element in the node set to all, calculate it and merge contribution;
4) try to achieve maximum and the corresponding node that node merges contribution degree;
5) total fusion of calculating whole paths at this node place is delayed time,
If the time-delay constraint is satisfied in total fusion time-delay, the time of fusion of then adjusting this node is former time of fusion and explores the step-length sum, forwards step 3) to; Otherwise this node can be adjusted the node set rejecting certainly and judge whether can adjust node set is empty;
6) if can adjust node set for empty, then finish to explore the step-length adjustment, the time of fusion of publisher node is carried out data fusion; Otherwise return step 3).
2. the method for claim 1 is characterized in that the method for described aggregation node collection network node routing iinformation is:
1) aggregation node is at first initiated route foundation by broadcast query message,
2) sensor node is received behind the query messages according to routing rule echo reply message, and by information method incidentally the routing state information of self is communicated to described aggregation node;
3) aggregation node is stored the routing iinformation that obtains, and finishes the collection of network routing iinformation.
3. method as claimed in claim 2 is characterized in that described routing iinformation comprises route level, father and son's nodal information.
4. method as claimed in claim 3, it is characterized in that the method for setting the exploration step-length is: five equilibrium is carried out in the constraint of will delaying time, and its quantity is called the exploration step number, and every part time span is called the exploration step-length.
5. method as claimed in claim 4 is characterized in that the described exploration step number and the network route number of plies are at the same order of magnitude.
6. method as claimed in claim 5 is characterized in that described search step number equals the network route number of plies.
7. the method for claim 1 is characterized in that the computing formula of described fusion contribution is:
In the formula, i, j, k ..., be the path of node i to aggregation node; Ui is the input flow rate of node i, and Ti is that the fusion of node i distributes the time.
8. the method for claim 1 is characterized in that adjusting with described that the time of fusion of node is initialized as zero in the node set.
9. method as claimed in claim 8 is characterized in that according to described time of fusion and described routing iinformation, calculates the described fusion contribution of adjusting node in the node set.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2007103045803A CN101471865B (en) | 2007-12-28 | 2007-12-28 | Method for high-efficiency data fusion in wireless sensor network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2007103045803A CN101471865B (en) | 2007-12-28 | 2007-12-28 | Method for high-efficiency data fusion in wireless sensor network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101471865A true CN101471865A (en) | 2009-07-01 |
| CN101471865B CN101471865B (en) | 2011-05-11 |
Family
ID=40829002
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2007103045803A Expired - Fee Related CN101471865B (en) | 2007-12-28 | 2007-12-28 | Method for high-efficiency data fusion in wireless sensor network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101471865B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101835237A (en) * | 2010-05-14 | 2010-09-15 | 南京邮电大学 | Data aggregation method in wireless sensor network |
| CN102123349A (en) * | 2011-01-10 | 2011-07-13 | 张俊虎 | Method for accumulatively searching data set of wireless sensor network |
| CN105245370A (en) * | 2015-10-13 | 2016-01-13 | 湘潭大学 | Self-adaptive layering cross-media data fusion method in Internet of Things |
| CN110278596A (en) * | 2018-03-13 | 2019-09-24 | 国基电子（上海）有限公司 | Wireless sensor network combination system and method |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100542138C (en) * | 2006-10-26 | 2009-09-16 | 中国科学院上海微系统与信息技术研究所 | L 3 architecture for radio sensor network |
| CN100583900C (en) * | 2006-11-16 | 2010-01-20 | 南京邮电大学 | Radio sensor network data convergence path planning method based on the intelligent agent |
| CN101094138B (en) * | 2007-06-15 | 2010-10-06 | 武汉大学 | Method for prolonging lifecycle of wireless sensor network based on D5 algorithm |
-
2007
- 2007-12-28 CN CN2007103045803A patent/CN101471865B/en not_active Expired - Fee Related
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101835237A (en) * | 2010-05-14 | 2010-09-15 | 南京邮电大学 | Data aggregation method in wireless sensor network |
| CN102123349A (en) * | 2011-01-10 | 2011-07-13 | 张俊虎 | Method for accumulatively searching data set of wireless sensor network |
| CN102123349B (en) * | 2011-01-10 | 2014-06-18 | 张俊虎 | Method for accumulatively searching data set of wireless sensor network |
| CN105245370A (en) * | 2015-10-13 | 2016-01-13 | 湘潭大学 | Self-adaptive layering cross-media data fusion method in Internet of Things |
| CN105245370B (en) * | 2015-10-13 | 2019-03-19 | 湘潭大学 | Across the media data fusion method of adaptive layered in a kind of Internet of Things |
| CN110278596A (en) * | 2018-03-13 | 2019-09-24 | 国基电子（上海）有限公司 | Wireless sensor network combination system and method |
| CN110278596B (en) * | 2018-03-13 | 2023-02-14 | 富联国基(上海)电子有限公司 | Wireless sensing network merging system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101471865B (en) | 2011-05-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102256267B (en) | The node clustering method of energy priority in radio sensing network | |
| Zhan et al. | Incentive mechanism design in mobile opportunistic data collection with time sensitivity | |
| CN101959244B (en) | Method for controlling hierarchical type route suitable for wireless sensor network | |
| CN101471865B (en) | Method for high-efficiency data fusion in wireless sensor network | |
| CN106507423A (en) | A kind of wireless sensor network routing method based on the reachable domain of dynamic | |
| CN102299854B (en) | Opportunistic network environment-oriented multi-object routing decision making system | |
| CN106452513A (en) | Mobile energy supplementation method in delay-constrained wireless sensor network | |
| CN104301965A (en) | Wireless sensor network inhomogeneous cluster node scheduling method | |
| CN102752395B (en) | A kind of on-line scheduling method of distributing for distributive data center real time business | |
| Wan et al. | AHP based relay selection strategy for energy harvesting wireless sensor networks | |
| Allam et al. | Enhanced zone-based energy aware data collection protocol for WSNs (E-ZEAL) | |
| CN109862612A (en) | Data collection and wireless charging method based on the planning of difunctional trolley movement routine | |
| CN109451556A (en) | The method to be charged based on UAV to wireless sense network | |
| CN109982452A (en) | A kind of wireless chargeable sensor network charging schedule method based on matroid | |
| Mishra et al. | A delay-bound efficient path design algorithm for mobile sink in wireless sensor networks | |
| CN105302858A (en) | Distributed database system node-spanning check optimization method and system | |
| CN101282291B (en) | Communication method for shunting information of wireless cluster sensing network | |
| Zhang et al. | An improved routing protocol for raw data collection in multihop wireless sensor networks | |
| CN102238561B (en) | Based on the node deployment method of the layering cooperation overlay model of Energy Efficient | |
| Boukerche et al. | A novel two-mode QoS-aware mobile charger scheduling method for achieving sustainable wireless sensor networks | |
| CN101471952B (en) | Fusion method for layered structure data of wireless sensor network | |
| Ma et al. | New AODV routing method for mobile wireless mesh network (MWMN) | |
| CN110072206A (en) | A kind of unmanned plane based on optimum capacity efficiency-internet of things data acquisition method and system | |
| Puviarasu et al. | Dynamic uneven clustering protocol for efficient energy management in EH-WSNs | |
| Adumbabu et al. | An Improved Lifetime and Energy Consumption with Enhanced Clustering in WSNs. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110511 Termination date: 20171228 |