Interference mitigation method for inter-body area network and intra-network joint optimization based on channel prediction

Abstract

The invention discloses an inter-body area network and intra-network joint optimization interference mitigation method based on channel prediction, which comprises the following steps of: firstly, predicting channel quality between each node and a coordinator in a body area network under different frequencies by using an autoregression model, classifying the body area network according to neighbor information among the body area networks, and dividing the channel quality of each node in the body area network according to neighbor information; secondly, distributing channels with different frequencies to various body area networks according to the principle of maximizing the average signal receiving strength of the nodes so as to avoid inter-network interference; in each body area network, the number of required time slots is calculated according to the service volume of each node, and when the number of time slots is not enough, the number of time slots of the nodes is reduced according to the principle that the total utility of the nodes is maximized; and then, according to the predicted channel quality condition of the nodes in the network under the allocated frequency, allocating time slots to the nodes by taking the minimum average packet loss probability as a principle. The method provided by the invention can be used for the scene of coexistence interference of the semi-dynamic body area network, and inter-network and intra-network interference can be avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of body area networks, and particularly relates to an interference mitigation method for joint optimization between and within body area networks based on channel prediction. Background Art

[0002] Body area network is a human body monitoring technology in the field of intelligent healthcare, and its core architecture consists of a coordinator and several nodes. The interference problem generated by multiple coexisting body area networks has become a key challenge restricting its reliability and large-scale application. In static or semi-dynamic coexisting body area networks, the stationary time of the body area network is much longer than the moving time.

[0003] Due to the motion characteristics of semi-dynamic body area networks, in recent years, many scholars have been committed to studying methods for allocating different frequency channels to different body area networks through communication negotiation to mitigate inter-network interference problems, such as graph coloring methods. That is, the problem of allocating body area network transmission resources is modeled as a graph coloring problem, and then by allocating different colors (different frequency channels) to adjacent vertices (adjacent body area networks), the effect of mitigating inter-network interference is achieved. However, in the strategies proposed by current scholars, the allocation of different frequency channels to different body area networks is randomly assigned, and no consideration has been given to the link state between the nodes within the body area network and the coordinator under which frequency channel is the best. If the link state of the nodes within the network is poor under a certain frequency channel, even if the inter-network interference problem is avoided by allocating different frequency channels, the communication between the nodes within the network and the coordinator cannot be successfully carried out, which will lead to an increase in the energy consumption of node data transmission and packet loss. Therefore, the design of the interference mitigation strategy should not only solve the inter-network interference but also improve the communication performance of the nodes within the network.

[0004] The current intra-network node time slot allocation scheme in the interference mitigation strategy cannot well solve the intra-network interference problem, that is, the interference problem caused by different nodes within the same body area network transmitting data in the same time slot. On the one hand, in a body area network where the number of time slots is not enough, the existing TDMA protocol will cause transmission conflicts between nodes or a few nodes have no transmission opportunity, thus increasing the delay. On the other hand, although the existing interference mitigation method based on Latin squares allocates time slots to intra-network nodes, this time slot allocation scheme has a large amount of resource waste. Secondly, the existing node time slot allocation scheme allocates a continuous time slot interval to a single node, and this does not take into account that the node needs some time for channel recovery when the channel quality is poor. If a continuous time slot interval with relatively poor channel quality is allocated to a certain node for data transmission, it will cause the node to be unable to send data to the coordinator in time. Therefore, the design of the interference mitigation strategy should fully consider the allocation problem of intra-network node time slot resources to avoid intra-network interference problems. Summary of the Invention

[0005] The main objective of the present invention is to overcome the drawbacks and deficiencies of the prior art, and to provide an interference mitigation method for joint optimization between and within body area networks based on channel prediction. This method can avoid inter-network interference and intra-network interference, and improve the communication performance of nodes within a single body area network.

[0006] To achieve the above objective, the present invention adopts the following technical solutions: An interference mitigation method for joint optimization between and within body area networks based on channel prediction. A body area network consists of a coordinator and several nodes. Between different body area networks, if multiple coexisting body area networks use channels of the same frequency for data transmission, serious inter-network interference will occur; within each body area network, if multiple nodes use the same time slot for data transmission, intra-network interference will also occur; in addition, regardless of whether multiple body area networks use channels of the same frequency, within each body area network, the channel quality between a node and the coordinator is different in each time slot. The interference mitigation method includes the following steps: S1. First, divide all body area networks into regions, and classify the body area networks according to neighbor information in each region. Two body area networks that are neighbors to each other cannot be classified into the same category; S2. When a classification conflict occurs in a body area network within a certain region, adjust the classification in parallel within the local region; S3. Allocate channels of different frequencies to each type of body area network according to the channel prediction results of intra-network nodes at different frequencies and the principle of maximizing the average signal reception strength of intra-network nodes; S4. Within each body area network, calculate the number of time slots required by each node and construct the utility function of the node. When the number of time slots is insufficient, reduce the number of node time slots based on maximizing the total utility of the nodes; S5. Within each body area network, the coordinator allocates time slots to each node based on the channel prediction results of the nodes at the allocated frequencies and the principle of minimizing the average packet loss rate of the nodes.

[0007] Furthermore, the process of classifying the body area networks in step S1 is as follows: S1.1. According to the geographical location, use the K-means algorithm to divide all body area networks into regions. The purpose of regional division is to improve the utilization rate of channels. Since the interference between regions is very small, body area networks in different regions can use channels of the same frequency for data transmission; S1.2. In each area, the body area network shares a neighbor information table in the form of broadcasting. The neighbor information table stores the neighbor information on whether all body area networks in this area are neighbors with each other. All body area networks are classified according to the neighbor information, and two body area networks that are neighbors with each other cannot be classified into the same category. The purpose is to distinguish body area networks that interfere with and coexist with each other through classification. Body area networks classified into the same category are far away from each other and hardly interfere with each other, while body area networks that are close to each other and interfere with each other need to be classified into different categories, which lays the foundation for frequency allocation later; S1.3. When a body area network can be classified into multiple categories, calculate the shortest distances between the body area network in these multiple categories and this body area network respectively. First, classify this body area network into the category with the farthest shortest distance from this body area network to minimize the possibility of classification conflicts occurring in the future due to the movement of the body area network, and put the other categories in the optional category pool of this body area network. The purpose of putting them in the optional category pool is to facilitate quick classification adjustment when classification conflicts occur in the body area network later, so as to enhance the stability of classification.

[0008] Further, when there is a classification conflict among body area networks in the cell in step S2, parallel adjustment of classification is performed in the local area. This is because semi-dynamic body area networks have a certain degree of mobility. As the human body moves, body area networks that were originally classified into the same category and did not interfere with each other may get closer to each other and become new neighbors. Therefore, the original category no longer applies to the current situation, resulting in classification conflicts among body area networks. After the conflict occurs, in order to improve the efficiency of classification adjustment, the method of local adjustment can reduce the traversal of body area networks that do not require classification adjustment. At the same time, parallel processing can also be performed in multiple non-overlapping local areas to improve the efficiency of classification adjustment. The specific process is as follows: The process of locally adjusting the body area network with classification conflicts in step S2 is as follows: S2.1. In the case of a classification conflict among body area networks, limit the classification conflict within three layers of body area networks, and perform classification adjustment in this local area, or perform classification adjustment simultaneously in multiple non-overlapping local areas; Among them, three layers of body area networks refer to taking a certain body area network as the first layer of body area network, all neighbor body area networks of this body area network as the second layer of body area network, and all neighbor body area networks of the neighbor body area networks of this body area network as the third layer of body area network; the local area refers to the spatial area composed of these three layers of body area networks; multiple local areas are non-overlapping spatial areas; The purpose of this step is to reduce the traversal of body area networks that do not require classification adjustment. In addition, parallel processing can also be performed in multiple local areas to perform classification adjustment simultaneously, greatly improving the efficiency of classification; S2.2. In each round of classification adjustment within each local area, the order of classification adjustment for multiple personal area networks is as follows: First, preferentially select the personal area network with the fewest number of neighbors for classification adjustment; second, select the personal area network with the largest number of categories in the optional category pool for classification adjustment; then, based on the neighbor information and the optional category pool of the personal area network, reclassify the personal area network. The basis for no classification conflict after classification is still that two personal area networks that are neighbors to each other cannot be classified into the same category; if there are still personal area networks with classification conflicts in the current local area after the first round of classification adjustment, then these personal area networks enter the second round of classification adjustment; for the personal area networks that still need to undergo the second round of classification adjustment, first expand the optional category pool of these personal area networks to the current maximum number of categories, and then reclassify the personal area network based on the neighbor information and the optional category pool of the personal area network. The basis for no classification conflict after classification is still that two personal area networks that are neighbors to each other cannot be classified into the same category; if there are still personal area networks with classification conflicts in the current local area after the second round of classification adjustment, then these personal area networks enter the third round of classification adjustment; for the personal area networks that still need to undergo the third round of classification adjustment, classify these personal area networks into new categories. Since the new categories will surely not conflict with the currently existing categories, the classification adjustment is completed.

[0009] The purpose of this step is to enable the personal area networks with classification conflicts to quickly readjust their classifications, enhancing the stability of classification to avoid new inter-network interference caused by the movement of personal area networks.

[0010] Furthermore, in step S3, different frequencies of channels are allocated to various types of personal area networks based on the channel prediction results of in-network nodes at different frequencies and the principle of maximizing the average signal reception strength of in-network nodes. The purpose is to maximize the overall channel quality of all nodes within the personal area networks in the entire region over a certain period of time in the future, enabling more nodes to transmit data in time slots with good channel quality. The specific process is as follows: S3.1. Nodes within each type of personal area network detect the channel quality between them and the coordinator at each frequency respectively, and predict the channel quality in a short period of time in the future based on the autoregressive model. Take the average value of the predicted values as the average signal reception strength in a short period of time in the future, and take the average value of the average signal reception strengths of all nodes in this type of personal area network as the average signal reception strength of this type of personal area network at a certain frequency. The autoregressive model is selected for prediction because this model is applicable to static or semi-dynamic personal area networks with stable channel quality changes and has good autocorrelation. Secondly, the autoregressive model not only has high prediction accuracy but also low computational complexity, and is very suitable for personal area networks with limited capabilities. Regarding the average signal reception strength of each type of personal area network at a certain frequency: Let the total number of nodes in the j-th type of personal area network be , then the average signal reception strength predicted by all nodes in the j-th type of body area network at the h-th frequency is expressed as , wherein, is the average signal reception strength predicted by the i-th node in the j-th type of body area network at the h-th frequency according to the detection result within a short period of time in the future; S3.2. With the goal of maximizing the average signal reception strength between all nodes in all types of body area networks and the coordinator within a short period of time in the future, different frequencies of channels are assigned to different types of body area networks, and the same type of body area network is assigned the same frequency. The purpose of this step of operation is to improve the overall communication quality of the nodes within the network within a short period of time while avoiding inter-network interference, so that more nodes can perform data transmission in time slots with good channel quality, maximize the reliability of communication between nodes and the coordinator, and reduce the transmission power during the data transmission process of nodes to save the energy consumption of the device.

[0011] Among them, the assignment formula is as follows:

[0012] wherein, represents taking the maximum value of , is the average signal reception strength predicted by all nodes in the j-th type of body area network at the h-th frequency, means not assigning the h-th frequency to the j-th type of body area network, means assigning the h-th frequency to the j-th type of body area network; S3.3. Convert the solution of the optimization assignment problem into the solution of the minimum cost matrix problem to obtain the frequency assignment scheme between body area networks.

[0013] Since the matrix in this optimization problem is a sparse matrix, the Hungarian algorithm is originally used to solve the assignment problem of minimizing the cost, and the computational complexity is low. The existing inter-network frequency method randomly assigns frequencies, but the frequency assignment scheme solved by the present invention takes into account the channel quality of the nodes within the network.

[0014] Furthermore, in step S4, within each body area network, calculate the number of time slots required by each node and construct the utility function of the node. When the number of time slots is insufficient, reduce the number of node time slots based on the maximization of the total node utility. The purpose is to reduce the possibility of packet loss caused by transmission conflicts due to insufficient available time slots for nodes within the body area network, so as to avoid in-network interference between nodes. In addition, the designed node utility function considers three factors: node priority, traffic volume, and channel stability factor, ensuring the reasonableness and fairness of the allocation of the number of time slots for each node within the body area network, and avoiding the situation where the delay increases due to the lack of transmission opportunities for a small number of nodes when the number of time slots is insufficient. The specific process is as follows: S4.1. In each personal area network, nodes detect the channel quality between them and the coordinator at the assigned frequencies. The coordinator predicts the signal reception strength at each time slot during the data transmission phase using the autoregressive method based on multiple detection results. Here, the communication between nodes and the coordinator is completed through superframes. Each superframe includes a channel detection phase, an algorithm operation phase, an information allocation phase, and a data transmission phase. Nodes send polling signals to the coordinator during the channel detection phase. The coordinator reads the signal reception strength from the received polling signals and collects it as historical data for prediction. The algorithm operation phase is used for the coordinator to calculate the number of time slots for each node and allocate transmission time slots to nodes. The information allocation phase is used for the coordinator to inform each node of the specific allocation determined in the algorithm operation phase. The data transmission phase is used for each node to send data to the coordinator. S4.2. Calculate the number of time slots required for each node and the number of time slots required for each area according to the traffic volume. This process is carried out during the algorithm operation phase of the superframe. Calculating the number of time slots required for each node according to the traffic volume of each node is to avoid resource waste during the time slot allocation process, and use the number of time slots required for each node as the upper limit for allocating time slots to that node. This is also one of the constraints in solving the optimization problem in the following step S4.4. The j-th node in the i-th personal area network The formula for the number of required time slots is

[0015] In the formula, is the total traffic volume that node needs to send; is the maximum traffic volume that can be sent in one time slot, represents taking the smallest integer greater than or equal to ; Assume there are sensor nodes in a personal area network. Then the number of time slots required for this personal area network is expressed as

[0016] In the formula, represents the number of time slots required for the j-th node in the personal area network, Assume there are personal area networks in a certain area. Then the number of time slots that need to be allocated for this area is expressed as

[0017] In the formula, represents the number of time slots required for the m-th personal area network, represents taking the maximum value of ; S4.3. Construct the utility function of the node. The utility function of the node includes the priority of the node, the traffic volume of the node, and the channel stability factor of the node. The existing time slot reduction scheme does not consider the channel stability of the node in future time slots. Allocating too many time slots to nodes with poor channel stability will lead to waste of transmission resources and cannot improve the utility of the node. Among them, the node 's priority is defined as

[0018] In the formula, represents the data received by node ; represents the normal range of the health parameter; 1, 0.75, 0.5, 0.25 represent four priority levels from high to low; represents the data type sent by node , and the value is defined as follows:

[0019] In the formula, the node with a priority of 1 has the highest priority, followed by the node with a priority of 0.75, then the node with a priority of 0.5, and finally the node with a priority of 0.25; The traffic volume of node is defined as In the formula,

[0020] In the formula, is the total traffic volume that node needs to send; is the maximum traffic volume that can be sent in one time slot, is the number of time slots actually allocated to node , is the number of time slots required by node ; The channel stability factor of node is defined as , in the formula, , where is the total number of time slots, refers to the predicted signal reception strength of the node in the i-th time slot after min-max normalization; Finally, the utility function of node is defined as

[0021] In the formula, the traffic volume , the channel stability factor , the priority , is the weight coefficient of priority, traffic volume, and channel stability factor, adjustable within the range of 0 to 1; S4.4. Considering the entry and exit of body area networks in the area, when the number of available time slots in a body area network is less than the number of time slots required by each node, it is necessary to reduce the number of time slots for each node. Among them, the reduction scheme is expressed as

[0022] In the formula, represents taking the maximum value of ; is the number of sensor nodes in a body area network; is the number of time slots allocated to a node; is the number of time slots required by a node; is the number of available time slots allocated to m body area networks in this area. The two constraints of this optimization problem are: (1) Each node is allocated at least one time slot and at most the number of time slots it needs; (2) The sum of the number of time slots required by all nodes in a body area network cannot exceed the number of available time slots allocated to its area; The strong memory search performance of dynamic programming quickly solves the corresponding optimal time slot allocation scheme when the total number of available time slots is different, improving the solution efficiency. At the same time, the dynamic programming method is applicable to any form of node utility function.

[0023] Next, use the dynamic programming method to solve the time slot allocation for each node for the above reduction scheme. The state transition equation in the dynamic programming method used is expressed as

[0024] In the formula, is the maximum utility generated by allocating time slots to the first nodes; is the utility generated by allocating time slots to the th node.

[0025] Further, in each body area network in step S5, the nodes allocate time slots to each node based on the minimum average packet loss rate of the nodes according to the channel prediction results at the assigned frequency. The purpose is to make full use of the transmission opportunities of the nodes and improve the overall success rate of data transmission among the nodes in the network. In each body area network, the predicted channel quality intensity values between each node and the coordinator are different at different transmission time slots. Nodes can reduce the transmission power and the packet loss rate by transmitting data in time slots with high signal reception intensity, while the probability of data loss will increase when transmitting data in time slots with low signal reception intensity. Therefore, when allocating time slots for the nodes, overall planning is required to minimize the average packet loss rate of the nodes in the network as much as possible, so that as many data packets as possible can be successfully received by the coordinator. The specific process is as follows: S5.1. In each body area network, first convert the predicted signal reception intensity of each node at each time slot during the data transmission phase into a signal-to-noise ratio, then convert the signal-to-noise ratio into an error rate, and finally convert the error rate into a packet loss rate, so as to obtain the predicted packet loss rate of each node at each time slot during the data transmission phase; convert the predicted signal reception intensity of each node at each time slot during the data transmission phase of the superframe into a packet loss rate, aiming to quantify the probability of data packet loss during the data transmission process of each node in the network at each time slot. At the same time, enable the nodes to make full use of the transmission opportunities, that is, each node tries to select a time slot with high signal reception intensity to transmit data, and nodes that need to be allocated multiple time slots can also select intermittent time slots for data transmission; S5.2. Allocate time slots to each node based on the minimum average packet loss rate of the nodes in each body area network. The purpose of this step is to improve the overall success rate of data transmission among the nodes in the network and allocate appropriate transmission time slots to each node as much as possible. Therefore, when allocating time slots for the nodes, global optimization is required to minimize the average packet loss rate of the nodes in the network as much as possible, so that as many data packets as possible can be successfully received by the coordinator. The time slot allocation scheme is expressed as follows:

[0026] In the formula, denotes taking the minimum value of The number of time slots allocated to node is Node either sends data or does not send data in a certain time slot, denotes that node does not send data in the th time slot, denotes that node sends data in the th time slot. Only one node is allowed to send data in each time slot, denotes that node sends data in the The packet loss rate predicted in a time slot; S5.3. Solve the time slot allocation scheme for each node using a greedy branch and bound method; the basic principle of the branch and bound method is to continuously cut the feasible solution space and prune the branches, gradually approaching the optimal solution; S5.4. The coordinator notifies each node of the time slot allocation information, and each node performs data transmission in the allocated time slots. These two processes are completed in the information allocation phase and the data transmission phase of the superframe respectively. Among them, each node may perform data transmission in separated time slots or in a continuous time slot interval, which depends on the channel quality of each node in each time slot. Most of the existing time slot allocation methods allocate continuous time slot intervals to nodes and cannot make good use of the transmission opportunities effectively. Based on the channel prediction results, this method not only minimizes the average packet loss rate of the nodes in the network to the greatest extent, but also enables each node to make full use of the transmission opportunities effectively, that is, each node can perform data transmission in separated time slots or in a continuous time slot interval according to the channel quality conditions. Secondly, in the present invention, if the channel quality of a certain node in the allocated time slots is less than the minimum threshold that can be received by the coordinator, then the node may not perform data transmission in this frame. The node only needs to record the starting time of the current frame and send the data in other frames with good channel quality in the future based on the maximum delay allowed to it.

[0027] Compared with the prior art, the present invention has the following advantages and beneficial effects: (1) The present invention designs a classification scheme for body area networks based on neighbor information and adjusts the classification, which improves the frequency utilization rate and the stability of classification. The existing classification methods for semi-dynamic body area networks do not consider the distance between current body area networks, which is likely to cause classification conflicts in the future, and are inefficient when adjusting classification conflicts, increasing the traversal of body area networks that do not require classification adjustment. The present invention preferentially classifies a body area network into the category that is currently the farthest from this body area network when the body area network can be divided into multiple categories, reducing the possibility of future classification conflicts for the body area network; and when adjusting classification conflicts, by introducing a conflict limit mechanism, the classification conflicts are limited within three-layer body area networks, and the classification adjustment is performed in a local area, reducing the traversal of body area networks that do not require classification adjustment. In addition, in multiple local areas, parallel processing can be performed, and classification adjustments can be carried out simultaneously, greatly improving the classification efficiency.

[0028] (2)The present invention also designs a scheme for inter-network frequency allocation based on the channel prediction results of nodes in the body area network at different frequencies and the highest average signal reception strength of nodes in the body area network, which maximally improves the overall channel quality of all nodes in the body area network in a future period of time in the entire area, enabling more nodes to perform data transmission in time slots with good channel quality. The existing inter-network frequency method randomly allocates frequencies, only avoiding inter-network interference by allocating channels with different frequencies to the body area network, without considering the overall channel quality of in-network nodes under this inter-network frequency allocation scheme, which may cause interference to the link quality between nodes and the coordinator during in-network communication, resulting in packet loss. The present invention selects the best inter-network frequency allocation scheme based on maximizing the average signal reception strength of in-network nodes through the channel prediction results of nodes at different frequencies, avoiding inter-network interference and maximizing the overall communication performance of in-network nodes to the greatest extent.

[0029] (3)The present invention also designs a time slot reduction scheme based on maximizing the total utility of in-network nodes. The existing time slot reduction scheme does not consider the channel stability of nodes in future time slots. Allocating too many time slots to nodes with poor channel stability will result in waste of transmission resources. This method avoids the possibility of transmission conflicts between nodes in the body area network with insufficient time slots and ensures that all nodes can be allocated at least one time slot. The utility function designed in this scheme considers the priority, traffic volume, and channel stability factor of nodes. Allocating time slots to nodes with higher priority, larger traffic volume, and better channel stability will generate greater utility, and the number of node time slots is reduced based on maximizing the total utility of in-network nodes until the time slot requirements are met. In addition, this scheme uses dynamic programming to solve the node time slot allocation scheme. The strong memory search performance of dynamic programming can quickly solve the optimal time slot allocation scheme when the total number of available time slots is different, improving the solution efficiency.

[0030] (4)The present invention also designs a scheme for allocating time slots to each node based on the channel prediction results of nodes at the allocated frequencies and minimizing the average packet loss rate of nodes, which maximally reduces the possibility of packet loss of in-network nodes and makes full use of the transmission opportunity. The existing time slot allocation scheme does not consider the channel quality of nodes in each time slot. Nodes performing data transmission in time slots with poor channel quality will not only increase the transmission power but also increase the possibility of packet loss. Secondly, the channel quality of each node fluctuates with time slots, and it takes time for the channel quality to recover. Most of the existing time slot allocation methods allocate continuous time slot intervals to nodes, and cannot make good use of the transmission opportunity effectively. This method not only maximally reduces the average packet loss rate of in-network nodes based on the channel prediction results, but also enables each node to make full use of the transmission opportunity effectively, that is, each node can perform data transmission in separated time slots or continuous time slot intervals according to the channel quality situation. Brief Description of the Drawings

[0031] To more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0032] Figure 1 is the topology diagram of the body area network in the embodiment of the present invention; Figure 2 is the interference type diagram of the body area network in the embodiment of the present invention; Figure 3 is the superframe structure diagram of node and coordinator communication in the embodiment of the present invention; Figure 4 is the distance schematic diagram between two moving body area networks; Figure 5 is the body area network distribution diagram in a certain area; Figure 6 is the body area network channel allocation diagram in a certain area; Figure 7 is the node time slot allocation diagram; Figure 8 is the flow chart of the interference mitigation method for joint optimization between body area networks and within the network based on channel prediction in the embodiment of the present invention; Figure 9 is the schematic diagram of dividing the network into multiple regions; Figure 10 is the classification result diagram of the body area network in a certain area; Figure 11 is the comparison diagram of the average packet loss rate of nodes under different numbers of coexisting body area networks for three body area network interference mitigation methods; Figure 12 is the comparison diagram of the average packet loss rate of nodes under different numbers of idle channels for three body area network interference mitigation methods.

[0033] Figure 13 is the comparison diagram of the average power consumption of nodes under different numbers of coexisting body area networks for three body area network interference mitigation methods.

[0034] Figure 14 is the comparison diagram of the channel utilization rate under different numbers of coexisting body area networks for three body area network interference mitigation methods. Detailed Implementation Manner

[0035] To enable those skilled in the art to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of this application.

[0036] In this application, the mention of "embodiment" means that the specific features, structures or characteristics described in combination with the embodiment can be included in at least one embodiment of this application. The appearance of this phrase in various positions in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described in this application can be combined with other embodiments.

[0037] Embodiment 1 This embodiment discloses an interference mitigation method for joint optimization between and within body area networks based on channel prediction. Among them, a body area network consists of a coordinator and several nodes, which are usually attached to the human body. For reference, see Figure 1 ; The interference types in the communication process of the body area network are generally divided into interference within the body area network, interference between body area networks, and interference with other devices. For reference, see Figure 2 ; The relationship between the interference between two mutually moving body area networks and the distance between the two body area networks. For reference, see Figure 4 ; The communication between the coordinator and the nodes uses a superframe as the time unit. The superframe consists of a channel detection phase (CDP), an algorithm operation phase (ARP), a sending allocation information phase (SDP), and a data transmission phase (MAP). The structure of the superframe can be seen in Figure 3 ; In the CDP phase of the superframe, the coordinator can read the corresponding signal reception strength from the polling signals received from the nodes and collect it as historical data for predicting the channel quality of its nodes in future time slots. Then, an autoregressive model is used to predict the signal reception strength for the future time slot. The main process of the interference mitigation method for joint optimization between and within body area networks based on channel prediction in the embodiments of the present invention is as shown in Figure 8 and the method includes the following steps: S1. First, divide all body area networks into regions, and classify the body area networks according to neighbor information in each region. Two body area networks that are neighbors to each other cannot be classified into the same category. Specifically: S1.1. According to the geographical location, use the K-means algorithm to divide all body area networks into regions. The specific implementation can be seen in Figure 9 ; S1.2. In each region, the body area network distribution map in a certain region can be seen in Figure 5 , and the body area network shares a neighbor information table in the form of broadcast. The neighbor information table stores the neighbor information on whether all body area networks in this region are neighbors to each other. The neighbor information table is shown in Table 1. All body area networks are classified according to the neighbor information. Two body area networks that are neighbors to each other cannot be classified into the same category. The specific classification results can be seen in Figure 10 ; Table 1. Neighbor Information Table

[0038] S1.3. When a body area network can be classified into multiple categories, calculate the shortest distances between the body area networks in these multiple categories and this body area network respectively. Prioritize classifying this body area network into the category with the farthest shortest distance from this body area network, and put other categories in the optional category pool of this body area network. For example, in Figure 6 the body area network can be classified into the first category or the third category. However, the body area networks in the third category are currently farther from the body area network . Therefore, prioritize classifying the body area network into the third category; S2. When there is a classification conflict among body area networks in the region, perform parallel classification adjustment in the local area. Specifically: S2.1. In the case of a classification conflict among body area networks, limit the classification conflict within three layers of body area networks, and perform classification adjustment in this local area, or perform classification adjustment simultaneously in multiple non-overlapping local areas. For example, in Figure 6 , if there is a classification conflict for the body area network , this local area consists of the body area networks , , , ; S2.2. In each round of classification adjustment in each local area, the order of classifying multiple body area networks is as follows: Prioritize selecting the body area network with the fewest neighbors for classification adjustment; secondly, select the body area network with the largest number of categories in the optional category pool for classification adjustment. For example, in Figure 6 , if both the body area networks and need to be classified and adjusted, prioritize classifying and adjusting the body area network Perform classification adjustment; then, based on the neighbor information and the optional category pool of the body area network, reclassify the body area network. The basis for no classification conflict after classification is still that two body area networks that are neighbors to each other cannot be classified into the same category. If there are still body area networks with classification conflicts in the current local area after the first round of classification adjustment, then these body area networks enter the second round of classification adjustment. For the body area networks that still need to undergo the second round of classification adjustment, expand the optional category pool of these body area networks to the current maximum number of categories, and then reclassify the body area network based on the neighbor information and the optional category pool of the body area network. The basis for no classification conflict after classification is still that two body area networks that are neighbors to each other cannot be classified into the same category. If there are still body area networks with classification conflicts in the current local area after the second round of classification adjustment, then these body area networks enter the third round of classification adjustment. For the body area networks that still need to undergo the third round of classification adjustment, classify these body area networks into new categories. In this example, the new category is the fourth category, and the classification adjustment is completed.

[0039] S3. Allocate channels with different frequencies to various types of body area networks according to the channel prediction results of the nodes in the network at different frequencies and the principle of maximizing the average signal reception strength of the nodes in the network. Specifically: S3.1. The nodes within each type of body area network detect the channel quality between them and the coordinator at each frequency, and predict the channel quality in the next short period of time according to the autoregressive model. Take the average value of the predicted values as the average signal reception strength in the next short period of time, and take the average value of the average signal reception strengths of all nodes in this type of body area network as the average signal reception strength of this type of body area network at a certain frequency. This channel prediction generally needs to re-detect the channel and re-predict every ten-odd superframes in the actual implementation process to ensure timeliness; S3.2. With the goal of maximizing the average signal reception strength between all nodes in all types of body area networks and the coordinator in the next short period of time, allocate channels with different frequencies to various types of body area networks. The body area networks of the same type are allocated the same frequency, which is transformed into an optimal allocation problem. For example, in Figure 5 the area shown, the body area network is divided into three categories and a total of 3 types of frequency channels are required. This optimal allocation problem is expressed as:

[0040] In the formula, represents taking the maximum value of , is the predicted average signal reception strength of all nodes in the j-th type of body area network at the h-th frequency, means not allocating the h-th frequency to the j-th type of body area network, means allocating the h-th frequency to the j-th type of body area network; S3.3. Transform the solution of the optimization allocation problem into the solution of the minimum cost matrix problem to obtain the frequency allocation scheme among body area networks. Among them, the minimum cost matrix can be obtained by subtracting each element from the maximum value of the average signal reception strength matrix. Then, perform row transformation or column transformation on the minimum cost matrix to make zero elements appear in each row and column, and find the minimum number of straight lines covering all zeros. Adjust the matrix until the optimal matching of independent zero elements is found as the final frequency allocation scheme for the body area networks in this area. The schematic diagram of the final frequency allocation of body area networks in a certain area is as shown in Figure 6 the following figure.

[0041] S4. In each body area network, calculate the number of time slots required by each node and construct the utility function of the node. When the number of time slots is insufficient, reduce the number of time slots of the node based on the maximization of the total utility of the nodes. Specifically: S4.1. In each body area network, the node detects the channel quality between it and the coordinator at the allocated frequency. The channel detection phase is in the CDP phase of the superframe structure shown in Figure 3 the following figure. The coordinator predicts the signal reception strength at each time slot in the data transmission phase using the autoregressive method based on multiple detection results. For example, for a certain node, record the predicted value sequence of the signal reception strength at each time slot as , where is the total number of time slots; S4.2. Calculate the number of time slots required by each node and the number of time slots required by each area according to the traffic volume. Among them, the number of time slots required by each body area network is the sum of the number of time slots required by all nodes in the network, and the number of time slots required by each area is the maximum value of the number of time slots required by all body area networks in the area; S4.3. Construct the utility function of the node. The utility function of the node includes three factors: the priority of the node, the traffic volume of the node, and the channel stability factor of the node. Among them, the channel stability factor needs to first perform maximum-minimum normalization processing on each element value in , and then take the mean value of the normalized predicted value sequence to obtain. The larger the value of the channel factor, the better the channel quality of the node; S4.4. Considering the entry and exit of body area networks in the area, when the number of available time slots in a certain body area network is less than the number of time slots required by each node, use the dynamic programming method to reduce the number of time slots of each node based on the maximization of the total utility of the nodes. Assume that the reduction method for nodes allocated time slots is expressed as:

[0042] In the formula, represents taking the maximum value of , is the number of sensor nodes in a personal area network, is the number of time slots allocated to a node, is the number of time slots required by a node, is the number of available time slots allocated to m personal area networks in this area. The two constraints of this optimization problem are: (1) Each node is allocated at least one time slot and at most the number of time slots required by the node; (2) The sum of the number of time slots required by all nodes in a personal area network cannot exceed the number of available time slots allocated to its area; The specific solution needs to be obtained by constructing a two-dimensional dp array and state transition equation, and another two-dimensional path array needs to be constructed to obtain the specific number of time slots allocated to each node through backtracking. Among them, represents the number of time slots allocated to the q-th node when q nodes are allocated s time slots .

[0043] S5. In each personal area network, the coordinator allocates time slots to each node based on the channel prediction result of the node at the allocated frequency with the principle of minimizing the average packet loss rate of the node. Specifically: S5.1. In each personal area network, first convert the predicted signal reception strength of each node at each time slot during the data transmission phase into the signal-to-noise ratio. For example, if the signal reception strength is -85 dBm and the noise floor is -105 dBm, then the signal-to-noise ratio is 20 dB; then convert the signal-to-noise ratio into the bit error rate BER, which can be achieved through the conversion formula in the present invention; finally, convert the bit error rate BER into the packet loss rate , for example, when , the data packet length bits, , thus obtaining the predicted packet loss rate of each node at each time slot during the data transmission phase; S5.2. Allocate time slots to each node based on minimizing the average packet loss rate of the nodes in each personal area network, which is transformed into an optimization allocation problem. The optimization allocation problem is expressed as:

[0044] In the formula, represents taking the minimum value of , the number of time slots allocated to node is , node either sends data or does not send data in a certain time slot, represents that node does not send data in the -th time slot, Indicates a node sends data in the th time slot, and only one node is allowed to send data in each time slot. Indicates a node The predicted packet loss rate in the th time slot; S5.3. Solve the optimization allocation problem using the greedy branch and bound method to obtain the time slot allocation scheme for each node. Further, the specific implementation steps for solving are as follows: S5.3.1. Generate an initial solution by relaxation substitution. Relax the integer variable to a continuous variable , and then solve the relaxed solution of the corresponding linear programming problem. If the relaxed solution satisfies the integer condition, it is directly used as the optimal solution; otherwise, its objective function value is the initial lower bound (corresponding to the minimization problem), which is the benchmark value for all integer solutions, and the lower bounds of subsequent branches will not be lower than this value. To improve efficiency, then quickly generate a feasible integer solution as the upper bound through a greedy strategy: sort the values in ascending order, and sequentially allocate time slots under the condition of satisfying the time slot number constraint for each node, and calculate its total packet loss rate as the initial upper bound ; S5.3.2. Branching strategy. Select a non-integer solution (the variable closest to 0.5 can be selected) in the relaxed solution to generate two sub-problems: force , that is, the nd time slot must be allocated to the th node, as the left branch; force , that is, the th time slot is not allocated to the th node, as the right branch; S5.3.3. Bounding and pruning. The relaxed solution of each sub-problem provides the lower bound of this branch. If the lower bound is higher than the currently known upper bound , then prune and stop searching this branch; otherwise, retain this branch and quickly generate a new feasible solution upper bound on this branch through a greedy strategy; S5.3.4. Search and iteration. Repeat steps 2 and 3, continue to branch, bound, and prune all sub-problems, but the sub-problem with the smallest lower bound can be preferentially branched or multiple sub-problems can be processed in parallel to accelerate convergence until the integer optimal solution is found.

[0045] S5.4. The coordinator notifies each node of the time slot allocation information, and each node conducts data transmission in the allocated time slots. Among them, each node may conduct data transmission in separated time slots or in continuous time slot intervals. For example, in a specific personal area network, the time slot allocation for each node is as Figure 7as shown

[0046] Embodiment 2 Referring to steps S1 to S5 in the interference mitigation method for joint optimization between and within body area networks based on channel prediction disclosed in Embodiment 1, in this embodiment, matlab is used to simulate and evaluate the performance of the proposed scheme. The specific operation method of network operation is as shown in Embodiment 1. On this basis, in this embodiment, the change of the average packet loss rate of nodes under different numbers of coexisting body area networks and different numbers of idle channels is tested, as well as the comparison with two other schemes. Some important parameters used in the simulation in this embodiment are as follows: the communication radius of the body area network is 1.5 m, the slot length is 10 ms, the data rate is 250 kbps, the packet size is 60 Bytes, and the selected noise is -92.2 dBm. The results obtained in the simulation experiment are the average values of the results obtained by simulating the changes of static body area networks or semi-dynamic body area networks, that is, the positions of the body area networks change multiple times, under the same number of body area networks or the same conditions. Figure 11 and Figure 12 respectively show the comparison of the average packet loss rate of nodes under different numbers of coexisting body area networks and different numbers of idle channels between the proposed scheme (Joint Optimized Scheduling Strategy, JORS) and EFRS proposed by Fan et al. and TSR proposed by Hu et al.

[0047] Figure 11 describes the change of the average packet loss rate of a single node with the increase in the number of coexisting body area networks. It can be easily seen from the figure that when the number of coexisting body area networks in the whole network increases, the congestion degree between wireless body area networks becomes higher and higher, and there are fewer and fewer idle channels for nodes to transmit data, and the average packet loss rate of a single node for data transmission is also larger. In the TSR scheme, a clustering scheme based on distributed coloring simply and efficiently clusters the wireless body area networks in the network, and uses the KM algorithm to allocate time slots for the sensor nodes in each cluster, maximizing the packet acceptance rate of node data transmission in each cluster. In the EFRS scheme, when the number of WBANs is larger, the degree of inter-network interference increases. Since the number of idle channels remains unchanged, the number of orthogonal Latin squares for WBANs to transmit data remains unchanged, and it is difficult to ensure that the WBANs interfering with each other can all select mutually orthogonal Latin squares to allocate transmission resources, and the probability of collision between nodes in different WBANs increases greatly, resulting in an increase in the packet loss rate of nodes. The proposed scheme JORS in this paper adjusts the transmission timing of each node by predicting the channel quality through the in-network time slot allocation scheme, and allocates transmission time slots for nodes based on minimizing the average packet loss rate of in-network nodes, so effectively reducing the average packet loss rate of nodes.

[0048] Figure 12Describes the change in the average packet loss rate of a single node with the increase in idle channels. It can be easily seen from the figure that when the number of WBANs in the network is fixed, the more idle channels there are, the more WBANs can transmit data on channels with different frequencies, reducing the degree of inter-network interference. Therefore, the average packet loss rate when a single node performs data transmission also decreases. In the TSR scheme, by introducing fairness weights, WBANs severely interfered by link quality can have more opportunities to select other time slots with good channel performance for transmission, thus ensuring the reliability of node data transmission to a certain extent. In the EFRS scheme, when the number of body area networks is fixed, as the number of idle channels increases, the order of the Latin square increases, and the number of orthogonal Latin squares is more. This enables different WBANs to select orthogonal Latin squares to allocate channels and time slots, greatly reducing the probability of inter-network interference among nodes. Moreover, as the number of idle channels increases, the transmission resources can sufficiently meet the transmission requirements of nodes at this time, greatly reducing the probability of transmission conflicts among nodes in different WBANs. The proposed scheme JORS in this paper predicts the channel quality between nodes and coordinators at different frequencies when allocating idle channels with different frequencies to different types of WBANs, improving the overall channel quality of all in-network nodes in each small period of time. Therefore, the probability of packet loss during node communication is effectively reduced.

[0049] Embodiment 3 Referring to steps S1 to S5 in the interference mitigation method for joint optimization between and within body area networks based on channel prediction disclosed in Embodiment 1, this embodiment uses matlab to simulate and evaluate the performance of the proposed scheme. The specific operation method of network operation is as shown in Embodiment 1. On this basis, this embodiment tests the change in the average energy consumption of nodes under different numbers of coexisting body area networks and the comparison with the other two schemes. Some important parameters used in the simulation in this embodiment and the calculation method of the results are the same as those in Embodiment 2. Figure 13 and Figure 14 Are respectively the comparison of the average power consumption and channel utilization rate of nodes of the proposed scheme JORS, the EFRS proposed by Fan et al., and the TSR proposed by Hu et al. under different numbers of coexisting body area networks.

[0050] Figure 13Describes the variation of the average power consumption of nodes with the increase in the number of coexisting body area networks. It can be easily seen from the figure that when the number of coexisting body area networks is larger, the number of idle channels in the network is not sufficient to enable the mutually interfering WBANs to mitigate inter-network interference by using channels with different frequencies, resulting in an increased probability of transmission conflicts among nodes. Nodes need more energy for data retransmission, etc., and the total energy consumption in the network increases. Therefore, the average power consumption of a single node also increases accordingly. In the JORS scheme, the average power consumption of nodes is the lowest. This is because the JORS scheme has the lowest average packet loss rate, reducing the energy consumption of data retransmission. In addition, the JORS scheme also predicts the link status between nodes and the coordinator in each transmission time slot, enabling in-network nodes to transmit data as much as possible in time slots with good channel quality, thereby reducing the transmission power. In the TSR scheme, the average packet loss rate of nodes is relatively low, reducing the energy consumption of data retransmission. However, when using the KM algorithm to allocate time slots for in-cluster nodes, the energy loss may increase due to the high computational complexity. With the increase in the number of coexisting body area networks, EFRS has the highest average energy consumption. This is because the average packet loss rate of EFRS nodes is the largest, resulting in more and more nodes needing to perform data retransmission, increasing the total energy consumption.

[0051] Figure 14It describes the variation of channel utilization in the entire network with the increase in the number of coexisting body area networks. In this solution, the channel utilization is defined as the ratio of the time for successful data transmission to the total channel time that meets the data transmission threshold. When the number of coexisting body area networks does not increase to the maximum number of coexisting body area networks that the current idle channel can accommodate, more and more nodes need to transmit data, the throughput increases, and the channel utilization is higher; when the number of coexisting body area networks exceeds the maximum number of coexisting body area networks that the current idle channel can accommodate, the degree of inter-network interference increases, the average packet loss rate of nodes increases, and the ratio of the time for successful data transmission to the total channel time that meets the data transmission threshold decreases, that is, the channel utilization decreases. The channel utilization of the JORS scheme is the highest. On the one hand, the JORS scheme assigns frequencies to different body area networks based on the channel prediction results of nodes at different frequencies and then maximizes the average signal reception strength of in-network nodes, which fully improves the signal reception strength of nodes in the body area network in each cell; on the other hand, within each network, the coordinator assigns transmission time slots to nodes based on the channel prediction results of each node at the assigned frequency and then minimizes the average packet loss rate of in-network nodes. Therefore, in the JORS scheme, the probability of successful transmission of nodes in the time slots with higher signal reception strength is the largest, and the channel utilization is the highest. In the TSR scheme, when the number of coexisting body area networks is less than 35, the channel utilization increases with the increase in throughput; when the number of coexisting body area networks exceeds 35, the average packet loss rate of nodes increases, resulting in a decrease in channel utilization. In the EFRS scheme, when the number of coexisting body area networks is less than 40, the channel utilization increases; when the number of coexisting body area networks exceeds 40, the channel utilization decreases. However, when the number of coexisting body area networks reaches about 50, the channel utilization of EFRS is higher than that of the TSR scheme, because when the number of coexisting body area networks is large, the transmission resources in the EFRS scheme are fully utilized, and the orthogonal Latin square can ensure a low probability of inter-network interference. Even when the number of coexisting body area networks is large, there is at most only one transmission conflict between nodes.

[0052] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0053] The above embodiments are the preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications made without departing from the spirit and principle of the present invention shall be equivalent replacement methods and are all included in the protection scope of the present invention.

Claims

1. An interference mitigation method for joint optimization between and within body area networks based on channel prediction, wherein, The body area network consists of a coordinator and several nodes. In each body area network, the channel characteristics between the nodes and the coordinator are time-varying. The interference mitigation method is characterized in that the method comprises the following steps: S1. First, divide all body area networks into regions, and classify the body area networks according to neighbor information in each region. Two body area networks that are neighbors of each other cannot be classified into the same category; S2. When a classification conflict occurs in a body area network in a certain region, adjust the classification in parallel in the local area; S3. According to the channel prediction results of the nodes in the network at different frequencies and based on the principle of maximizing the average signal reception strength of the nodes in the network, allocate channels with different frequencies to each type of body area network; S4. In each body area network, calculate the number of time slots required by each node and construct the utility function of the node. When the number of time slots is insufficient, reduce the number of node time slots based on the maximization of the total node utility; S5. In each body area network, the coordinator allocates time slots to each node based on the channel prediction results of the node at the allocated frequency and with the principle of minimizing the average packet loss rate of the node.

2. The interference mitigation method for inter - body - area - network and in - network joint optimization based on channel prediction according to claim 1, wherein The process of classifying the body area networks in step S1 is as follows: S1.

1. Divide all body area networks into regions by using the K-means algorithm according to the geographical location; S1.

2. In each region, the body area networks share a neighbor information table in a broadcast form. The neighbor information table stores the neighbor information on whether all body area networks in the region are neighbors of each other. All body area networks are classified according to the neighbor information. Two body area networks that are neighbors of each other cannot be classified into the same category; S1.

3. When a body area network can be classified into multiple categories, calculate the shortest distances between the body area networks in these multiple categories and this body area network respectively. First, classify this body area network into the category with the farthest shortest distance from this body area network, and put the other categories in the optional category pool of this body area network.

3. The interference mitigation method for joint optimization between and within body area networks based on channel prediction according to claim 1, characterized in that, The process of locally adjusting the body area network with a classification conflict in step S2 is as follows: S2.

1. In the case of a classification conflict in the body area network, limit the classification conflict within three layers of body area networks, and perform classification adjustment in this local area, or perform classification adjustment simultaneously in multiple non-overlapping local areas; Among them, the three-layer body area network means that taking a certain body area network as the first-layer body area network, all neighbor body area networks of this body area network are the second-layer body area network, and all neighbor body area networks of the neighbor body area networks of this body area network are the third-layer body area network; the local area refers to the spatial area composed of these three layers of body area networks; multiple local areas are multiple local areas where the spatial areas do not overlap with each other; S2.

2. In each round of classification adjustment within each local area, the order of classification adjustment for multiple personal area networks is as follows: First, preferentially select the personal area network with the fewest neighbors for classification adjustment; second, select the personal area network with the largest number of categories in the optional category pool for classification adjustment; then, based on the neighbor information and the optional category pool of the personal area network, reclassify the personal area network. The basis for no classification conflict after classification is still that two personal area networks that are neighbors to each other cannot be classified into the same category; if there are still personal area networks with classification conflicts in the current local area after the first round of classification adjustment, then these personal area networks enter the second round of classification adjustment; for the personal area networks that still need to undergo the second round of classification adjustment, first expand the optional category pool of these personal area networks to the current maximum number of categories, and then reclassify the personal area networks based on the neighbor information and the optional category pool of the personal area network. The basis for no classification conflict after classification is still that two personal area networks that are neighbors to each other cannot be classified into the same category; if there are still personal area networks with classification conflicts in the current local area after the second round of classification adjustment, then these personal area networks enter the third round of classification adjustment; for the personal area networks that still need to undergo the third round of classification adjustment, classify these personal area networks into new categories to complete the classification adjustment.

4. The interference mitigation method for inter-body area network and in-network joint optimization based on channel prediction according to claim 1, characterized in that The process of step S3 is as follows: S3.

1. Nodes within each type of personal area network detect the channel quality between them and the coordinator at each frequency, and predict the channel quality in a short period of time in the future according to the autoregressive model. Take the average value of the predicted values as the average signal reception strength in a short period of time in the future, and take the average value of the average signal reception strengths of all nodes in this type of personal area network as the average signal reception strength of this type of personal area network at a certain frequency. For the average signal reception strength of each type of body area network at a certain frequency: Let the total number of nodes in the j-th type of body area network be , then the average signal reception strength predicted by all nodes in the j-th type of body area network at the h-th frequency is expressed as Wherein, is the average signal reception strength predicted by the i-th node in the j-th body area network at the h-th frequency according to the detection result within a short period of time in the future; S3.

2. With the goal of maximizing the average signal reception strength between all nodes within all types of personal area networks and the coordinator in a short period of time in the future, allocate channels of different frequencies to different types of personal area networks, and allocate the same frequency to personal area networks of the same type. Among them, the allocation formula is as follows: In the formula, represents taking the maximum value of ; is the predicted average signal reception strength of all nodes in the j-th type of body area network at the h-th frequency, means not allocating the h-th frequency to the j-th type of body area network, means allocating the h-th frequency to the j-th type of body area network; S3.

3. Transform the solution of the optimization allocation problem into the solution of the minimum cost matrix problem to obtain the frequency allocation scheme between personal area networks.

5. The interference mitigation method for joint optimization between and within body area networks based on channel prediction according to claim 1, characterized in that The process of step S4 is as follows: S4.

1. Within each personal area network, nodes detect the channel quality between them and the coordinator at the allocated frequency. The coordinator predicts the signal reception strength at each time slot during the data transmission stage using the autoregressive method based on multiple detection results. Among them, the communication between the node and the coordinator is completed through a superframe. Each superframe includes a channel detection stage, an algorithm operation stage, an information allocation stage, and a data transmission stage; nodes send probe signals to the coordinator during the channel detection stage, and the coordinator reads the signal reception strength from the received probe signals and collects it as historical data for prediction; the algorithm operation stage is used for the coordinator to calculate the number of time slots of each node and allocate transmission time slots for the nodes; the information allocation stage is used for the coordinator to inform each node of the specific allocation determined during the algorithm operation stage; the data transmission stage is used for each node to send data to the coordinator. S4.

2. Calculate the number of time slots required for each node and the number of time slots required for each area according to the traffic volume. The j-th node in the i-th personal area network The calculation formula for the number of required time slots is , where is the total traffic volume that the node needs to send; is the maximum traffic volume that can be sent in one time slot, represents taking the smallest integer greater than or equal to upward. Suppose there are sensor nodes in a personal area network, then the number of time slots required for the personal area network is expressed as , where represents the number of time slots required for the j-th node in the personal area network. Suppose there is a personal area network in a certain area, then the number of time slots to be allocated in this area is expressed as In the formula, represents the number of time slots required for the m-th personal area network, represents taking the maximum value of ; S4.

3. Construct the utility function of the node. The utility function of the node includes the priority of the node, the traffic volume of the node, and the channel stability factor of the node. Among them, the priority of the node is defined as In the formula, represents the data received by the node ; represents the normal range of health parameters; 1, 0.75, 0.5, 0.25 represent four priority levels from high to low; represents the data type sent by the node and its value definitions are as follows: Among them, the node with a priority of 1 has the highest priority, followed by the node with a priority of 0.75, then the node with a priority of 0.5, and finally the node with a priority of 0.25; Node business volume is defined as Wherein, is the total traffic volume that the node needs to send; is the maximum traffic volume that can be sent in one time slot, is the number of time slots actually allocated to the node ; is the number of time slots required by the node . Node channel stability factor of is defined as , where is the total number of time slots refers to the predicted signal reception strength of the node at the i-th time slot after min-max normalization; Finally, the node has its utility function defined as Wherein, the traffic volume , the channel stability factor , the priority , are the weight coefficients of the priority, the traffic volume, and the channel stability factor, which are adjustable within 0 to 1; S4.

4. Considering the entry and exit of the body area network in the area, when the number of available time slots in a body area network is less than the number of time slots required by each node, it is necessary to reduce the number of time slots of each node. Among them, the reduction scheme is expressed as In the formula, represents taking the maximum value of ; is the number of sensor nodes in a certain personal area network, is the number of time slots allocated to a node, is the number of time slots required by a node, is the number of available time slots allocated to m personal area networks in this area. The two constraints of this optimization problem are: (1) Each node is allocated at least one time slot and at most the number of time slots required by the node; (2) The sum of the number of time slots required by all nodes in a personal area network cannot exceed the available time slots allocated to its area; Next, the dynamic programming method is used to solve the time slot allocation of each node for the above reduction scheme. The state transition equation in the used dynamic programming method is expressed as In the formula, is the maximum utility generated by allocating time slots to the first nodes, is the utility generated by allocating time slots to the th node.

6. The interference mitigation method for inter-body area network and in-network joint optimization based on channel prediction according to claim 5, wherein The process of allocating time slots to each node based on the minimum average packet loss rate of the nodes in step S5 is as follows: S5.

1. In each personal area network, the signal reception strength predicted by each node in each time slot during the data transmission stage is first converted into a signal-to-noise ratio, then the signal-to-noise ratio is converted into a bit error rate, and finally the bit error rate is converted into a packet loss rate, so as to obtain the packet loss rate predicted by each node in each time slot during the data transmission stage; S5.

2. Assign time slots to each node based on minimizing the average packet loss rate of nodes within each personal area network. The time slot allocation scheme is described as follows: Wherein, represents taking the minimum value of , the number of time slots allocated to node is . Node either sends data or does not send data in a certain time slot, represents that node does not send data in the -th time slot, represents that node sends data in the -th time slot. Only one node is allowed to send data in each time slot, represents the predicted packet loss rate of node in the -th time slot; S5.

3. Use the greedy branch and bound method to solve the time slot allocation scheme of each node; S5.

4. The coordinator notifies each node of the time slot allocation information, and each node transmits data in the allocated time slots. Among them, each node may transmit data in separated time slots or in continuous time slot intervals.

Images