RU2528415C1 - Method for distributed traffic balancing in wireless sensor network - Google Patents

Abstract

FIELD: radio engineering, communication.

SUBSTANCE: invention discloses a method and a system for distributed traffic balancing in a wireless sensor network based on an algorithm for routing from a source node to a destination node, where the wireless sensor network is presented as a graph G (N, M), where N denotes network nodes and M boundaries. There are K routes and information is generated at a rate Q_c and transmitted over a communication channel C at a rate q_c. The i-th node has an energy margin E_i, and each boundary ij has a weight/value e_ij which corresponds to energy for transmitting one data packet from node i to j, and the life T_i of each node is defined as

. Each node determines a routing table and stores a message transmission vector, performs analysis of routing alternatives on the most optimum sum vectors calculated from the routing table. To this end, the life of the entire network $${\text{T}}_{\text{sys}}=\underset{i\in N}{{\displaystyle \min }}{\text{ T}}_{\text{i}}(q_c)$$

is determined. Maximisation of the life is defined as maximise T_sys, and to achieve maximum life of the entire network, routes are distributed, where the choice of a route in a network is based on use of the least costly transmissions at each node and the most costly transmissions are excluded.

EFFECT: providing efficient routing, prolonging the life of a network and high reliability.

11 cl, 4 dwg

Description

FIELD OF THE INVENTION

The invention relates to the field of wireless communications and can be used in automated monitoring systems that operate both independently and as part of multi-level information management systems, in particular in real-time monitoring of environmental or industrial parameters with nodes distributed over large territories and not having wired communication lines and power lines.

State of the art

Currently, sensor networks are increasingly taking their place in monitoring applications for various places and events. In connection with the development of wireless communication technology, it became possible to develop wireless distributed sensor networks (RCC). Distributed sensor networks differ from conventional networks in limited energy resources, low computing power, the need for a denser location and low cost of one node. These features from other networks (for example, cellular) determine new goals and objectives of their application. Wireless sensor networks are widely used in many areas of human activity, and therefore they are now receiving great attention.

The distributed sensor network consists of many cheap, autonomous, multi-functional nodes that are located in the monitoring zone. Each node consists of a set of blocks, such as: a sensor used to receive data from the environment, a data transmit-receive block, a microcontroller for processing and controlling signals, and an energy source. The processor is powered by a self-contained battery with a finite energy resource, which leads to significant restrictions on energy consumption. Maintaining sensor nodes, such as replacing batteries, is expensive, especially when nodes are located in hard-to-reach places, so most sensor networks are maintenance free and work until the battery runs out. This property of sensor networks is very important in the development of routing algorithms in the BSS, allowing to increase the efficiency of energy consumption of the network.

So, there are many ways to save energy nodes in the sensor network, and figure 1 shows their classification. The methods can be divided into three large groups - this is energy conservation using work cycles, based on the amount of information transmitted and mobility.

Work cycles include topology control and energy management. Topology control is aimed at using or reducing redundant network connections in order to save resources. You can manage consumption by using various energy-saving media access control protocols (MAC protocols) and device operating modes. The second class of ways to save energy is based on the amount of information transmitted, as well as on the receipt of this information in economical ways. The energy spent on processing information is incomparably less than the energy required to transmit it, so intranet data processing, data compression or prediction is used. Repeaters are also used to save power from sensor networks.

Routing methods can be divided into the following categories: direct, hierarchical, and routing depending on geographic location.

Direct routing involves the transmission of messages from node to node in the network, where each node performs the same function of transmission and / or relay, in contrast to hierarchical, where one or more nodes of information collection and processing are allocated. The disadvantage of direct routing is that networks that collect information from some area will send a lot of redundant information, especially with a significant density of the sensor network. In order to avoid redundancy of information, they use special algorithms aimed at obtaining information not from nodes, but from a specific area of the network. For example, the Sensor Protocols for Information via Negotiation (SPIN) algorithm is known, where the base station sends a request to a specific region of the sensor network. Upon receiving the request, the nodes of the region fulfill the request requirement, locally exchange data and send back a generalized response.

In hierarchical routing, it is required to use nodes with a large supply of energy for collection and processing, which, although it allows saving on the transfer of already processed data of a much smaller amount, is often unacceptable due to the homogeneity of the devices used or other difficulties. In order not to use specialized nodes, there are several technologies. So, the Low-Energy Adaptive Clustering Hierarchy (LEACH) technology is known, when the collection function is taken in turn by several sensor network nodes selected according to a certain algorithm, thereby distributing the load of the collection node.

Routing depending on geographic location is also called geometric routing, because the geometrical direction to the base station is used to find the route. There is also routing by virtual coordinates, which are built not only depending on the actual position of the node, but also take into account natural surface irregularities, obstacles, the level of the transmission channel, etc.

Multithread routing is also known, where message delivery from one node is possible in several ways. Recently, much attention has been paid to on-demand routing at the base station, for example, based on finding the shortest path and maintaining it taking into account the bad channel and the failure of nodes. However, nodes located at the shortest distance are quickly depleted, which leads to disconnection and a decrease in the lifetime of the network, which is often understood as the lifetime of the first failed node. Therefore, there is a need to create a technology for maximizing the lifetime of the sensor network, which is solved by one or another linear programming method.

So, patent RU 2439812 C1 is known as an essentially technical solution, published 2012-01-10, IPC H04W 36/00, where a self-configuring sensor network of a plurality of sensors and actuators based on routing depending on geographical location is disclosed. The sensor network consists of a central data processing device (DLC) and N base stations (BS) located uniformly or randomly along the boundaries of the network coverage area, where the BSs are spatially referenced to global positioning coordinates and contain memory for storing the value of the confidence coefficient, which is a number in the range from the specified minimum and maximum values. The confidence coefficient for the BS is set approximately equal to the maximum value. Inside the coverage area of the sensor network, M nodes are evenly or randomly located, with M >> N. The nodes are equipped with a memory designed to store the coordinates of the spatial reference, which is initialized with random values during the production process, and to store the confidence coefficient, which is initialized with a value approximately equal to the minimum value of the confidence coefficient. Each node and BS establish a connection with no more than K neighboring nodes and BS, and the value of K depends on the characteristics of the bandwidth of the communication channel, the performance and power consumption of microprocessors that are part of them. After the connection is established, the nodes and the BS perform the operation of mutual determination of the values of spatial coordinates. For this, each node or BS cyclically transmits its own memory values to store the values of the coordinates of the spatial reference and memory to store the value of the confidence coefficient. In each processing cycle, the node receives the values of coordinates and confidence coefficients from all neighboring devices with which it is connected, and determines the calculated values of its own coordinates and its own confidence coefficient using the method of weighted averaging of the values of its own coordinates and coordinates of neighboring devices, using confidence coefficients the device itself and neighboring devices. Thus, the nodes of the sensor network receive spatial reference. To route a message from the central control center to a node with coordinates (x, y, z), it transmits a message to one or more BSs closest to the required coordinates. The indicated BS transmit the message to the nearest nodes, and the nodes sequentially - to their nearest nodes in the direction of the vector directed to the desired point (x, y, z). The nodes spatially attached to points located at a distance not exceeding the radius of sensitivity of the sensor network r perceive the message as addressed to them. Further arbitration of nodes for selecting the final addressee of the message, as well as sending confirmation of the receipt of the message, is carried out as necessary, based on the technical requirements for the functioning of the network. To route a message from a node to a central control center, nodes are additionally equipped with memory to store a list of coordinates of the nearest BSs. To transmit the DLCM message, the node transmits the message to one or several neighboring nodes in the direction of the vector directed to the point with the BS coordinates, when the message reaches the BS, it sends the message directly to the DLC and, if necessary, sends a transmission confirmation message to the sending node.

The disadvantage of such a self-configuring sensor network and the method of its operation is the complexity of the equipment used, associated with the need to set and use the coordinates of the spatial reference of nodes and base stations, and also this solution does not provide a long life time of the entire network as a whole.

As the closest analogue to the prototype, we can propose the method of routing with the maximum lifetime in the wireless Ad-hoc network, disclosed in the publication Arvind Sankar and Zhen Liu, Maximum Lifetime Routing in Wireless Ad-hoc Networks, INFOCOM 2004, Twenty-third Annual Joint Conference of the IEEE, Computer and Communications Societies, vol.2, pp1089-1097, where the problem of maximizing the lifetime of a sensor network, which is solved by linear programming, is formulated, namely, an algorithm is proposed to minimize the sum of potential functions of all queues.

The disadvantage of this method is the low efficiency, since nodes located at the shortest distance are often quickly depleted, which leads to disconnection and a decrease in the network lifetime.

Thus, there is a need to solve the above problems of the prior art.

SUMMARY OF THE INVENTION

The technical result to which the proposed invention is directed is, in particular: providing efficient routing and extending the life of the wireless sensor network for monitoring various objects and parameters in real time, where the information of each node is important, increasing functionality, reliability and reducing the cost of use systems for monitoring. Using the proposed solution will improve the efficiency of operation of the controlled object due to the longer life of the autonomous power supply battery, which will allow to register and transmit data on the parameters of the object and / or environment for a longer time.

The essence of the proposed method of distributed traffic balancing in a wireless sensor network is to apply a new routing algorithm from the source node to the destination node. Communication between the nodes in the sensor network is, for example, via the Zigbee protocol, or in the unlicensed radio frequency range, or via a mobile digital radio network, or any other suitable wireless communication protocol. A distributed sensor network can be represented as a graph G (N, M), which defines a set of the mentioned nodes and the connections between them, where N are network nodes, and M faces, there are also K routes. Information is generated at a speed of Q _c and transmitted over communication channel C at a speed of q _c , with the i-th node having an energy reserve E _i , and each face ij has a weight / price e _ij that corresponds to the energy for transmitting one data packet from node i to j, while the lifetime T _{i of} each node is defined as

Next, a routing table is determined at each node and a message transmission vector is aligned, an analysis of possible route options is carried out according to the most optimal total vectors, which are calculated from the routing table, for this the lifetime of the entire network T _{sys is} determined

Thus, the maximization of the lifetime is defined as maximize T _sys , and to achieve the maximum lifetime of the entire network, routes are allocated for the transmitted information, while the choice of the traffic route in the network is based on the use of the least expensive transmissions at each node, and the most costly nodes based on its calculated T _i .

At least one source node contains a sensor for measuring and monitoring physical parameters (quantities) with autonomous power, which monitors in a given area of the network and transmits messages (data packets) with measured parameters to at least one destination node.

Alternatively, in each node, in order to bring the monitoring data to a uniform form, primary processing of physical parameters obtained from sensors can be performed, for example, by storing them in memory, averaging, and analog-to-digital conversion to the corresponding code. As the measured parameters for monitoring, for example, the environment, various parameters are used, such as temperature, pressure, humidity, light, smoke, vibration level, etc.

Alternatively, the choice of the route when creating and / or updating the routing table is made in accordance with combinations of criteria such as the length of the route, measured by the number of routers through which you must go to the destination node; communication channel bandwidth; predicted total transmission time; communication channel cost; the amount of residual energy on the site.

Alternatively, the method further updates the values of the lifetime T _{i of} each node or the lifetime of the entire system T _sys in accordance with the mentioned combination of criteria, carried out when sending a message from the source node to the destination node or when a connection break between the nodes is detected.

Alternatively, after constructing a routing table, the function of transmitting packets along the optimal paths (route) is implemented when a packet is sent, each network node places the address of the next node in the packet header at the media access control level (MAC level).

Also proposed is a distributed traffic balancing system in a distributed sensor network based on a routing algorithm from a source node to a destination node in a distributed sensor network according to the proposed method, comprising: a destination node connected by a wireless communication channel to a source node, which is a sensor module that houses a transceiver , a sensor of physical parameters, a microcontroller for processing and control and an autonomous source of their power, and the destination node contains a transceiver a sensor, means for accumulating the received information, and means for processing and displaying the received information from sensor modules for constructing a model of the object or space under study.

Alternatively, the sensor modules can be divided into groups, and each group is connected to the destination node by wireless communication through its transceiver. Real-time monitoring of environmental or industrial parameters is carried out pointwise in a given area, where the first subset of the mentioned set of source nodes performs monitoring functions, and the second subset of source nodes performs only the transceiver functions of data packets with measured physical parameters obtained from the first subset of source nodes.

These and other structural and functional features and advantages of the proposed invention will become apparent from a detailed description of its options, which should be read in conjunction with the drawing.

Brief Description of the Drawings

Figure 1 shows a well-known classification of methods for conserving the energy of nodes in a sensor network.

Figure 2 shows the algorithm for constructing a sensor network based on a survey.

Figure 3 shows the sensor network in the form of a graph G (N, M).

Figure 4 shows the options for determining routes.

DETAILED DESCRIPTION OF THE INVENTION

An algorithm is proposed on which the technology of automated data collection and transmission is based on the proposed RCC (network of autonomous wireless self-organizing mobile devices) to a single point for constructing a model of the object or space under study. This model can mainly be used to build networks for monitoring environmental or industrial parameters in real time, to monitor the condition in the life cycle of buildings and structures, during the design and construction of recreation areas and facilities for sanatorium-resort construction, as well as in various other areas of the automotive industry, railway transport, in road construction, in medicine.

The proposed invention can significantly increase the functionality, reliability and reduce the cost of using such systems for monitoring. The cost reduction is inextricably linked with the constructive, functional and software unification of the parts that make up the system, which requires a thorough analysis of the requirements and research on how to build a universal software and hardware platform for creating environmental monitoring systems based on wireless sensor technology. For this, various parameters are studied: temperature, pressure, humidity, light, smoke, vibration, which are collected through self-organizing sensor networks. RCC consists of end devices, intermediate routers, a network coordinator and a dedicated data collection point, sometimes called a network gateway, it serves to convert data from a radio channel to a network organized on optical or copper wires - Ethernet. Sensors for collecting physical parameters are attached to network nodes - end devices, which are built into a single structure through the network coordinator, for example, using the ZigBee protocol. This allows you to deploy a network for monitoring in a short period of time with minimal cost and sufficiently high reliability.

Each RCC node is equipped with an autonomous power source, which allows them to be installed in hard-to-reach places to take the required readings with minimal labor. A feature of the proposed invention is the creation of a unique scalable software and hardware, consisting of a set of modules necessary for implementation, which allows you to control devices for the maximum possible operating time, and at the same time create an authentic model of a spatial heterogeneous environment in automatic mode. Communication between devices takes place over the air in various communication standards, including the Zigbee protocol, in the unlicensed frequency range or over the mobile digital radio network. The data collected for processing allows you to use such a system to build an ecological 3D model of the environment / space or object under study, significantly reducing the amount of time required for processing and obtaining information and financial resources. The essence of the proposed algorithm, called two ladder-logic, is to control the RCC elements, which allows balancing the load on the network nodes so that the transmitted data is sent to the nearest network node not randomly, but to the one that has the largest energy reserve at the current time . The RCC functioning algorithm used allows you to change the load on the network nodes in such a way that the entire network remains operational for as long as possible.

The use of RCC can provide significant advantages in both technological and economic aspects, over traditional data collection and processing systems. A fundamental increase in the performance of the collection and processing of digital telemetry, achieved through the use of PCC, allows you to aggressively enter the market and switch to technological solutions of a new generation, thereby making possible the emergence of new automated systems operating in real time based on cloud technologies. With the development of technology, a transition from connected local monitoring networks to large-scale monitoring, surveillance and prediction systems based on wireless RCC should occur.

Figure 2 shows an example of routing and building a sensor network based on a survey. RCC consists of many cheap, autonomous, multifunctional nodes that are located in the monitoring zone. Each node consists of a set of blocks, such as a sensor used to receive data from the environment, a data reception and transmission unit, a microcontroller for processing and controlling signals, and a small-sized energy source. The processor is powered by a self-contained battery with a finite energy resource, which leads to significant restrictions on energy consumption. Maintenance of sensor nodes, such as replacing a stand-alone battery, is expensive, especially when nodes are located in hard-to-reach places, so most sensor networks are maintenance-free and operate until the battery is depleted.

The routing algorithm allows you to build a route based on requests and responses. Network Coordinator 1 sends a HELLO broadcast request and receives responses from the router (router) 2. Each router also sends a broadcast request and receives responses from neighboring devices, these can be other routers or end devices 3. Based on the received responses (signal strength, response time and other parameters) the coordinator builds a routing table on each router. Further, the route selection is carried out in the standard algorithm by determining the weight graph with the minimum total value.

As a rule, sensor nodes are equipped with similar devices with a certain set of functions. After installation, during operation, the sensor nodes must themselves be organized into a communication network, where each node uses only those functions that are necessary to solve the task. Routing also occurs automatically. In addition to the primary routing, regular network rebuilding is also required, because devices can lose the communication channel or fail for reasons related to external or internal factors.

The operation of each sensor unit is aimed at measuring various environmental parameters, such as temperature, pressure, light, humidity, smoke, vibration level, etc. Such a variety of parameters entails various fields of application, for example, data collection and environmental monitoring, monitoring of various production facilities, located both in a separate building and on a large territory, oil and gas industry facilities, transport facilities, military applications, etc. Sensor networks perform various tasks and which can be roughly divided into two categories. The first category of tasks is associated with the detection of events that occur very rarely, but require immediate notification and / or location detection. The second category (monitoring) includes the tasks of continuous measurement of a quantity over a long period of time. Here, the delay time may be equal to the characteristic time of change of the measured parameter. Monitoring can be carried out pointwise over any area, with point measurement, the main part of the nodes plays the role of transmitters, and only a small part of the nodes directly monitors.

A routing algorithm with traffic balancing in a distributed sensor network is proposed. For this, a distributed sensor network can be represented as a graph G (N, M) with N nodes and M faces, which represents a set of existing nodes and possible connections between it, as shown in Fig. 3. Each i-th node initially has an energy reserve E _i . Each face ij has a weight / price e _ij that corresponds to the energy for transmitting one data packet from node i to j. It is believed that there are K routes, and information is generated at a speed of Q _c and transmitted over a communication channel C at a speed of q _c .

The lifetime T _{i of} each node will be equal in such a system

According to the algorithm used, the routing table is determined by the coordinator at each node. Lines a message transfer vector. Next, an analysis of the possible route options is carried out according to the most optimal total vectors, which are calculated according to the routing table. Thus, the goal is to save the total energy consumed in the entire network for the transmission of one packet. This is effective for data networks, when the network lifetime is determined by the time during which the network is able to transmit messages.

In networks where each node simultaneously performs two functions: measuring a quantity and transmitting messages, that is, the sensor network performs the function of monitoring physical quantities in a given area, the importance of each node is important to complete the picture.

Then the lifetime of the entire system T _{sys is} defined as:

The task of maximizing the lifetime will look like: maximize T _sys , and to achieve the maximum lifetime of the entire system, it is necessary to distribute routes for the transmitted information. The essence of the proposed routing method with balancing traffic in the RCC is that the choice of the traffic route in the network is based on the use of the least costly transmissions on each node that can be involved in data transmission. In other words, the most expensive hop-hop jumps (a transit section or a transition in the network between two nodes of the network through which traffic is transmitted) are excluded from possible options for the route of the data packet, thereby saving energy at each node and reducing the probability of failure of the node, which eliminates the collapse of the entire network of measurements due to the fact that one node has already ceased to perform actual measurements.

The choice of route option (shown in Fig. 4) during the formation and updating of the routing table is made in accordance with a combination of criteria such as: the length of the route, measured by the number of routers through which it is necessary to go to the destination; communication channel bandwidth; predicted total forwarding time; communication channel cost; the amount of residual energy on the site.

After constructing a routing table, the algorithm implements the function of transmitting packets along optimal paths by the fact that when sending a packet through a router, each node on the local network places the address of the next recipient in the packet header at the MAC level. Thus, in the example shown in Fig. 3, based on the minimum total cost (weight / price) at the nodes (Fig. 4), route 1 will be selected, with the sum of the cost of weight / price - 9, as the minimum value. Thus, the passage of traffic through the nodes of route 1 will lead to the complete energy depletion of node 4 as soon as possible, which will disable these nodes and exclude the possibility of collecting parameters at the desired points of the study.

However, when using the proposed distributed traffic balancing algorithm based on weighting factors, route 2 will be selected, which will allow the sensor network to exist an order of magnitude longer. This is possible due to the fact that the load on all nodes, in the case of the proposed algorithm, is distributed more systematically across all nodes of the network.

The proposed invention can be implemented using various functional and / or hardware, software, special-purpose processors and / or combinations thereof. Preferably, the invention is implemented as a combination of hardware and software. The software is preferably implemented as an application program materially implemented on a program storage / reader. An application program can be downloaded or executed by means of a computer containing any architecture, and is implemented on a computing platform that has hardware: one or more central processors, random access memory and input-output interfaces. The various embodiments of the invention described above are provided for understanding only and as an example, and should not be limited to these examples.

Claims (11)

1. A method of distributed traffic balancing based on a routing algorithm from a source node to a destination node in a distributed sensor network,
in this case, the distributed sensor network is represented as the graph G (N, M), which characterizes the set of the mentioned nodes and the connections between it, where N network nodes, and M faces, there are K routes, and information is generated at a speed of Q _c and transmitted over the communication channel C with a speed q _c , and the i-th node has an energy reserve E _i , and each face ij has a weight / price e _ij , which corresponds to the energy for transferring one data packet from node i to j,
wherein the lifetime T _{i of} each node is defined as

a routing table is determined on each node and a message transmission vector is aligned,
the analysis of possible route options is carried out according to the most optimal total vectors, which are calculated according to the routing table, for this the lifetime of the entire network T _{sys is} determined

in this case, the maximization of the lifetime is defined as maximize T _sys , and to achieve the maximum lifetime of the entire network, routes are allocated for the transmitted information, while the choice of the traffic route in the network is based on the use of the least expensive transmissions at each node, and the most costly ones are excluded when constructing the route. 2. The method according to claim 1, characterized in that in at least one node of the source there is a self-powered sensor that measures and monitors physical parameters in a given area and transfers data packets with measured physical parameters to at least , to one destination node. 3. The method according to claim 2, characterized in that the sensors are used to measure physical parameters to monitor the environment based on the control of the following parameters: temperature, pressure, humidity, light, smoke, vibration level. 4. The method according to claim 3, characterized in that in at least one source node, primary processing of physical parameters obtained from said sensors is performed, for example, accumulation, averaging, analog-to-digital conversion. 5. The method according to claim 1, characterized in that the communication between nodes in the sensor network is carried out according to the Zigbee protocol, or in the unlicensed radio frequency range, or via a mobile digital radio network, or by any other wireless communication protocol. 6. The method according to claim 1, characterized in that in the communication channel between the source node and the destination node contains a router that interacts with these nodes. 7. The method according to claim 1, characterized in that the choice of the route when generating and / or updating the routing table is made in accordance with combinations of criteria such as the length of the route, measured by the number of routers through which it is necessary to go to the destination node, the bandwidth of the communication channel , the predicted total transmission time, the amount of residual energy at the node, the cost of the communication channel. 8. The method according to claim 1, characterized in that after constructing the routing table, the function of transmitting packets along the optimal routes is realized when sending a packet, where each network node puts the address of the next node in the packet header at the level of medium access control (MAC level) . 9. The method according to any one of claims 1, 6, 7, characterized in that the method further includes the step of updating the lifetime values T _{i of} each node or the lifetime of the entire system T _sys in accordance with the mentioned combination of criteria, carried out when sending a message from a source node to a destination node or when a disconnection is detected between nodes. 10. A system of distributed traffic balancing in a wireless sensor network for monitoring physical parameters according to the method according to any one of claims 1 to 9, comprising a plurality of source nodes interconnected and a destination node connected to at least one source node, which It is a sensor module, which houses a transceiver, a sensor of physical parameters, a microcontroller for processing and control and an autonomous power source, the sensor modules are divided into groups and each group is connected to a node values through its transceiver, while the destination node contains a transceiver, means for accumulating the received information and means for processing and displaying the received information from the sensor modules to build a model of the object or space under study. 11. The system of claim 10, characterized in that the monitoring is carried out pointwise in a predetermined area, where at least one subset of the aforementioned set of source nodes performs monitoring functions by means of its sensors of physical parameters, and another subset of the source nodes performs by means of its transceivers only the functions of the reception and transmission of data packets with measured physical parameters obtained from the mentioned subset of source nodes.

Images