Intelligent livestock management needs to detect the behavior of livestock in the pasture and collect information such as the movement path and location of the livestock. The pasture Internet of Things system can be used to obtain the above information. The system architecture based on wireless sensor networks can meet the communication needs between various sampling nodes. However, due to the remote geographical location of ranches and the complex network environment, problems such as connection interruption and packet loss often occur when connecting to the Internet. The use of traditional network connection methods may lead to a large amount of data loss. In order to reduce the loss of sampled data when the connection to the Internet is interrupted, a ranch IoT data transmission scheme based on opportunistic networks is proposed, and the communication between sensors in the ranch environment is analyzed. Communication status, three ways of communication between sensors are summarized, and modeled and analyzed respectively. Under the premise of limited storage capacity of sensors, using the principle of opportunistic network, a calculation method of access point density is proposed, and the intrinsic relationship between livestock movement speed, sensor node storage capacity and data transmission packet loss rate is summarized, thus ensuring System data loss is within the allowable range of the design. The correctness of the method is demonstrated through experiments that evaluate and verify the theoretical results of the proposed method.
With the widespread application of Internet of Things technology in agriculture [1], information methods are gradually used in animal husbandry production [2]. Livestock behavior monitoring in pastures is a basic condition for intensive and intelligent management of animal husbandry.
The monitoring of livestock behavior can be effectively realized using Internet of Things technology [3-6]. Generally, ranches are in relatively remote areas, and ranch IoT is often at the edge of the network service area or outside the service area. The communication capabilities of sensor nodes are limited. In the ranch environment, it can only be connected to the nearest base station through a wireless public network [7-11] Due to the transmission of information, continuous IoT monitoring data often cannot be transmitted through the network, resulting in network anomalies such as signal interruptions.
Opportunistic networks (ON) [12-14] are a kind of self-organizing network, which are specially designed for the situation where wireless network links are often disconnected and there may not be a complete link between the source node and the target node, using the movement of nodes. , data exchange occurs when nodes enter the mutual communication coverage area to complete data communication. Therefore, opportunistic networks are often used to solve problems such as wildlife monitoring [15-16] and network access in remote areas [17-18].
This article analyzes the data transmission environment of the Internet of Things under the special conditions of the ranch, and uses the characteristics of opportunistic network technology to design network data transmission for the livestock industry Internet of Things, so that the data transmission of the pasture Internet of Things can adapt to the frequent interruptions of the sensor network in the pasture environment. Provide an effective solution to solve the problem of data transmission interruption of the Internet of Things in livestock farms.
1 Analysis of farm Internet of Things system
1.1 Structure of the pasture Internet of Things system The pasture Internet of Things system can realize the collection and processing of livestock behavior information. The system collects physical system information (such as livestock location information, body temperature information, movement status information, etc.) through sensors, and uses the network (2G/3G/4G and other wireless public networks) to transmit the information to the application server for processing. Figure 1 is a structural diagram of a livestock information monitoring Internet of Things system in a pasture. The smart sensors in the perception layer have certain storage, computing and communication capabilities. The data collected by sensors use sensor networks for short-distance data transmission, such as networks composed of protocols such as ZigBee. At the same time, nodes use wide area network access technology, generally using wireless methods, such as GPRS networks, to transmit sensor data collection to the Internet. This part of the network organization forms the network layer of the ranch Internet of Things system.
In the animal husbandry Internet of Things system, the sensors are in a mobile state during the information collection process, and the collection nodes continuously change their spatial positions. This change in the spatial topology makes the animal husbandry Internet of Things different from other fields in facility agriculture that are currently widely discussed, such as intelligent Internet of Things systems such as greenhouses; at the network layer, during the data communication process between sensor nodes and the Internet, due to the mobility of the nodes themselves, and when the nodes use wireless communication methods, the distance between the nodes and the base station often exceeds the coverage of the base station, making The network is often down.
1.2 Analysis of Communication Status of Ranch Internet of Things System In the ranch environment, network transmission mainly uses wireless networks to exchange data between base stations and communication nodes. There are three situations: sensor network nodes are all distributed within the communication coverage of the base station; sensor network The node distribution part is within the communication coverage of the base station; the sensor network nodes are all distributed outside the communication coverage of the base station.
Use sets to represent the above three situations: A = {a1, a2, a3}. There are also three situations in how sensor network communication is organized: all sensor network nodes can communicate directly with the base station, and sensor network nodes cannot communicate with each other; some of the sensor network nodes can communicate with the base station, and sensor network nodes can communicate with each other; all sensor network nodes can communicate with each other. Sensor network nodes can communicate with the base station and the nodes can also be connected to each other.
Expressed as a set: B = {b1, b2, b3}. Therefore, in the pasture scene, the possible way of network transmission is C=A×B={a1, a2, a3}×{b1, b2, b3}={(a1, b1), (a1, b2), ( a1, b3), (a2, b1), (a2, b2), (a2, b3), (a3, b1), (a3, b2), (a3, b3)} among which (a1, b1), (a1 , b2), (a1, b3) In the three cases, all nodes of the sensor can directly transmit data to the Internet. In this case, there is no network interruption, etc., and the connection processing model is consistent with the Internet processing method; (a3, b1) , (a3, b2), (a3, b3), the network is completely interrupted; therefore, the issues discussed in this article focus on the three situations: (a2, b1), (a2, b2), (a2, b3).
2 Ranch Internet of Things modeling based on opportunistic networks
In view of the particularity of data communication of the Internet of Things in the animal husbandry environment and pastures, based on the analysis of the system architecture of the Internet of Things in animal husbandry, and using the principles of opportunistic networks, a reference for designing the system using the principles of opportunistic networks in system implementation is proposed. model, thereby solving problems such as unstable connections and frequent changes in the spatial location of sensor nodes during Internet of Things data transmission in animal husbandry.
2.1 Overview of Opportunistic Networks Opportunistic networks are a network organization scheme for which there may not be a complete path for communication between network nodes. The physical architecture solution of this network organization form comes from DTN (Disruption Tolerant Network) ks) network [19-20]. Opportunistic networks can be regarded as DTN networks under a wireless self-organizing network. The technical route of its architecture is to complete communication tasks in accordance with the storage-carry-forward routing mechanism. The function of “carrying” packet information in opportunistic networks is achieved by adding a bundle layer [14] to the network architecture. Figure 2 shows a comparison between an opportunistic network architecture and an ordinary TCP/IP network architecture. Through the bundle layer, opportunistic networks can meet the requirements of reliable transmission of data in special environments such as delays and frequent network interruptions.
This paper proposes an engineering solution for data transmission in pastures based on opportunistic networks. In order to meet the needs of wide applicability and system stability in actual work, a data sending strategy based on a single copy of the flooding algorithm is adopted in the data forwarding mechanism. The mobility aspect is based on the movement of the feeding process in the pasture. Under this condition, the relationship between equipment memory, data sampling speed, feeding radius and animal movement speed required for the project is calculated, providing theoretical basis and methods for calculating the memory, data sampling period, etc. required by the system design in project implementation.
2.2 Analysis and modeling of farm Internet of Things scenarios. Section 1.2 discusses the data transmission situations in three network environments: (a2, b1), (a2, b2), (a2, b3) respectively. In Figure 3, respectively Scene 1, scene 2 and scene 3 are represented. Here we discuss the connection conditions of the opportunistic network required for the transmission of information on the Internet of Things in the pasture under these three situations. 2.2.1 Scenario 1 The communication relationship between the sensor network node and the base station when the connection is (a2, b1). Some of the sensor nodes are connected to the base station and some are not connected. The sensor nodes can move, that is, the sensor nodes Part of the movement range is within the base station signal coverage. At this time, there is no communication association between sensors. In this case, it is equivalent to forming a network between each sensor network node and the base station. In the network configuration shown in Scenario 1, each sensor node has the opportunity to connect and communicate with the base station only when it reaches the base station signal coverage. The base station signal coverage and livestock activity radius are shown in Figure 4. Let the livestock activity range be In a circular area (for convenience of explanation, it is related to the shape of the pasture in actual grazing), the radius of activity of livestock is r (m), which is the length of line segment A1O2 in the figure, and the distance between the base station and the center of livestock activity is D (m) , that is, the length of O1O2. The base station signal coverage range is R (m), which is the length of A1O1. The gray diagonal area in Figure 4 of the base station is the signal instability area. When the sensor node transmits signals within this range, there will be packet loss. The packet loss law obeys the function φ. The function φ is a function related to the distance between the node and the base station and the maximum delay limit of the network. , this function can be set accordingly according to different system design requirements and corresponding transmission methods. The width of the gray area is d (m). According to the geometric relationship in Figure 4, the angle formed by the livestock activity center and the intersection point of the two circles can be calculated as ∠A1O2A2, which is expressed as α1 = 2arccos ((r2 + D2 – R2) / (2rD)). The angle between the intersection points of the circles is ∠A1O1A2, expressed as α2 = 2arccos ((R2 + D2 – r2) / (2RD)). The area where the base station signal coverage intersects with the livestock activity range is S1 = r2α1/2 + R2α 2/2 -sin(α1/2)rD
Palabras clave: Pasarela de Internet de las Cosas