Predictive Energy Harvesting Wireless Body Area Network Emergency Data Resource Allocation Method

Abstract

This invention discloses a prediction-based method for emergency data resource allocation in wireless body area networks (WBANs), belonging to the field of WBAN technology. Key technical points include: sensor nodes reporting status information to the coordinator; a mobile edge computing platform predicting channel quality based on a temporal convolutional network and calculating the minimum transmission power and maximum transmittable data volume for each time slot; allocating time slots according to node data requirements, energy status, and transmission capacity; employing a minimum cost flow algorithm combined with segmented minimum service constraints for time slot scheduling to limit the longest waiting interval for nodes; the coordinator broadcasting the allocation results; and nodes adaptively selecting and executing energy harvesting, transmission, or switching modes based on real-time data and energy status. This invention can reduce emergency data transmission latency, improve reliability, and enhance system throughput and long-term operational stability under uncertain energy conditions.

Description

Technical Field

[0001] This invention belongs to the field of wireless body area network technology, and particularly relates to a method for emergency data resource allocation in a predictive energy harvesting wireless body area network. Background Technology

[0002] Wireless body area networks (BPANs) are wireless sensor networks that monitor the human body. They typically consist of a coordinator and multiple wearable, implantable, or portable sensor nodes. These nodes collect vital signs data such as electrocardiogram (ECG), blood pressure, blood glucose, and body temperature, and then transmit the data to the coordinator for forwarding to a remote medical center, enabling continuous monitoring and analysis of human health status. Due to the long-term, continuous, and real-time requirements of health monitoring data, BPANs need to minimize system power consumption while ensuring transmission reliability. Currently, most BPAN sensor nodes are battery-powered. While this provides stable power and can support some bursts of data uploads, battery replacement or charging causes monitoring interruptions and degrades user experience. Furthermore, battery configurations increase device size and weight and pose an environmental burden. Given the limitations of node size and energy, simply optimizing media access control protocols, sleep scheduling, or power control is insufficient to fundamentally achieve long-term stable operation. Therefore, energy harvesting technology has been introduced into BPANs, converting ambient energy into electrical energy to improve system endurance. Radio frequency (RF) energy harvesting, with its strong controllability and adaptability to human wearing scenarios, has attracted widespread attention in BPANs.

[0003] However, the available energy of energy harvesting body area network (BNB) nodes is dynamic and uncertain, making resource allocation and data scheduling more complex, especially in sudden scenarios such as falls or cardiac arrhythmias, where emergency data needs to meet the requirements of low latency and high reliability transmission. Existing dynamic resource allocation strategies for BNBs often focus on maximizing average throughput or improving fairness under energy harvesting constraints, and usually do not specifically optimize for the low latency requirements of sudden emergency data. For example, during routine monitoring, when a node suddenly generates emergency data such as a sudden drop in heart rate, the lack of a priority guarantee mechanism for emergency data in existing scheduling schemes often forces this data to queue up with other routine data, resulting in excessive transmission latency and potentially missing the best opportunity for medical intervention.

[0004] Meanwhile, existing time-slot scheduling schemes lack a clear mechanism to further reduce the latency of emergency data transmission while maintaining system throughput. Current resource allocation strategies generally adopt fixed transmission modes, meaning that nodes often execute only a single working mode within a scheduling cycle, such as dedicating all their time to energy harvesting or all their time to data transmission, or using a fixed energy harvesting-then-data transmission process. They lack the ability to adaptively switch based on changes in node instantaneous energy status and data load. When a node has low energy but urgently needs to send data, the fixed energy harvesting-then-transmission mode leads to increased waiting time for emergency data; conversely, if a node has sufficient energy but only needs to transmit a small amount of regular data, the fixed pure transmission mode wastes energy and affects long-term operational stability. Therefore, a resource allocation method suitable for energy harvesting body area networks (ENAs) for radio frequency energy harvesting is needed to achieve low-latency transmission of emergency data under dynamic energy changes and improve the system's adaptability and overall performance to sudden data bursts. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes an emergency data resource allocation method for energy harvesting wireless body area networks based on prediction, thereby resolving the issues present in the prior art.

[0006] In a first aspect, to achieve the above objectives, the present invention provides a method for emergency data resource allocation in a predictive energy harvesting wireless body area network, comprising the following steps:

[0007] Each sensor node reports node status information to the coordinator, which includes at least energy status, data status, and link channel information.

[0008] The coordinator sends the node status information to the mobile edge computing platform;

[0009] The mobile edge computing platform predicts the channel quality of each time slot during the data transmission phase based on a prediction model.

[0010] Based on the prediction results, the mobile edge computing platform calculates the minimum transmission power and maximum transmittable data volume of each node in each time slot to meet the packet loss rate threshold constraint.

[0011] The mobile edge computing platform allocates the number of data transmission time slots to each node based on the node's data requirements, energy status, and transmission capabilities.

[0012] The mobile edge computing platform uses an optimization algorithm to allocate specific time slot sets and scheduling order to each node under the condition of meeting the preset minimum service constraints.

[0013] The coordinator broadcasts the resource allocation results to each node, and the resource allocation results include at least the set of time slots and the corresponding transmission parameters.

[0014] Each node adaptively selects the energy harvesting mode, transmission mode, or energy harvesting-transmission switching mode before the arrival of the allocated time slot based on its own real-time data status and energy status, and performs the corresponding energy harvesting and data transmission operations within the time slot.

[0015] Optionally, the process by which each sensor node reports its node status information to the coordinator includes:

[0016] During the polling phase of the superframe, polling time slots are allocated to each sensor node;

[0017] Each sensor node adopts an energy harvesting-transmission switching mode within its polling time slot, dividing the polling time slot into an energy harvesting phase and a data transmission phase.

[0018] The node collects energy during the energy harvesting phase and sends link channel quality, node data status, and remaining energy to the coordinator during the data transmission phase.

[0019] Optionally, the process by which the mobile edge computing platform predicts the channel quality for each time slot during the data transmission phase based on a predictive model includes:

[0020] Based on the historical link channel information reported by each node, an input sequence for channel prediction is constructed.

[0021] The input sequence is fed into a temporal convolutional network prediction model, which outputs a prediction sequence for the channel quality of each time slot in the next superframe.

[0022] Optionally, the process by which the mobile edge computing platform allocates data transmission time slots to each node based on the node's data requirements, energy status, and transmission capabilities includes:

[0023] Calculate the normalized data demand factor based on the node's data cache size, regular data arrival rate, and urgent data volume;

[0024] Calculate the normalized energy factor based on the node's remaining energy and predicted collectable energy;

[0025] Calculate the normalized transmission capacity factor based on the maximum amount of data that a node can transmit in each time slot.

[0026] The number of time slots is allocated to each node based on the node score factor composed of the data demand factor, energy factor, and transmission capacity factor.

[0027] Optionally, the mobile edge computing platform employs an optimization algorithm to allocate specific time slot sets and scheduling orders to each node under preset minimum service constraints. This process includes:

[0028] The total time slots during the data transmission phase are divided into multiple consecutive segments;

[0029] Set a minimum service constraint for each segment, requiring each node to be allocated at least a preset number of time slots in each segment;

[0030] A time-slot scheduling problem is constructed with the goal of maximizing the total amount of data that can be transmitted. This problem is then transformed into a minimum-cost flow problem for solution, resulting in a specific set of time slots for each node.

[0031] Optionally, the process by which the coordinator broadcasts the resource allocation results to each node includes:

[0032] During the algorithm execution and energy broadcasting phase of the superframe, the coordinator sends radio frequency energy signals to each sensor node;

[0033] During the beacon phase of the superframe, the coordinator broadcasts to each node the resource allocation results, which include the time slot set, minimum transmission power, time switching factor, average maximum transmittable data volume, and emergency data reserve energy.

[0034] Optionally, the process by which each node adaptively selects its working mode based on its own real-time data and energy status includes:

[0035] The node calculates the transmission pressure factor based on the remaining data volume, the remaining number of scheduling time slots, and the average maximum transmittable data volume.

[0036] Based on the transmission pressure factor, the presence of emergency data, and the comparison between remaining energy and emergency data reserve energy, the node selects one of the following: energy harvesting mode, transmission mode, or energy harvesting-transmission switching mode.

[0037] Optionally, before each node performs the corresponding energy harvesting and data transmission operations within the time slot, a mode indication process is also included, comprising:

[0038] At the beginning of the allocated time slot, the node indicates the selected operating mode to the coordinator by sending a preset combination of pulse signals.

[0039] Secondly, the present invention also provides a computer terminal device, comprising:

[0040] One or more processors;

[0041] A memory, coupled to the processor, for storing one or more programs;

[0042] When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the prediction-based energy harvesting wireless body area network emergency data resource allocation method in the first aspect described above.

[0043] Thirdly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the steps of the prediction-based energy harvesting wireless body area network emergency data resource allocation method in the first aspect described above.

[0044] Compared with the prior art, the present invention has the following advantages and technical effects:

[0045] This invention provides a prediction-based emergency data resource allocation method for energy harvesting wireless body area networks (WBAs), which can reduce emergency data transmission latency, improve the reliability and reachability of emergency data transmission, and enhance system throughput efficiency and long-term operational stability under uncertain energy conditions. By constructing a low-latency resource allocation framework for emergency data, this invention overcomes the shortcomings of existing technologies that focus on average throughput or fairness and lack a mechanism to guarantee low latency for emergency data, enabling priority low-latency transmission of emergency data under energy-constrained conditions. The node adaptive working mode selection mechanism proposed in this invention allows nodes to dynamically switch between energy harvesting mode, transmission mode, and energy harvesting-transmission switching mode, thereby overcoming the problem that existing fixed workflows cannot adapt to dynamic changes in energy and data, effectively reducing emergency data waiting latency. This invention introduces a channel prediction method based on temporal convolutional networks to drive resource allocation based on the prediction of future time slot channel quality, and adaptively calculates the minimum transmission power and energy harvesting-transmission switching parameters accordingly, overcoming the problems of frequent retransmissions and energy waste caused by existing methods relying on instantaneous or statistical information, and improving the effective transmission capability of a single time slot. The prediction-driven low-latency time slot scheduling method proposed in this invention introduces segmented minimum service constraints to limit the longest waiting interval of nodes on the basis of satisfying the node time slot quota, and combines the minimum cost flow algorithm to achieve global scheduling optimization. It overcomes the problem that existing scheduling strategies lack an emergency data access latency control mechanism, and can improve the overall throughput performance of the system while ensuring low latency of emergency data. Attached Figure Description

[0046] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0047] Figure 1 A topology diagram of a wireless volume area network for energy harvesting;

[0048] Figure 2 This is a superframe structure diagram of the communication between the node and the coordinator in this invention;

[0049] Figure 3 This is a structural diagram of the three working modes of the time slot of the present invention;

[0050] Figure 4This is a diagram illustrating the collaborative architecture of the mobile cloud computing platform and the mobile edge computing platform of this invention.

[0051] Figure 5 This is a flowchart of a resource allocation method for low-latency transmission of emergency data in a predictive energy harvesting wireless body area network according to the present invention.

[0052] Figure 6 This is a structural diagram of the temporal convolutional network prediction model of the present invention;

[0053] Figure 7 This is a schematic diagram of time slot allocation in this invention;

[0054] Figure 8 This is a flowchart illustrating the node selection process based on time slot adaptive working mode in this invention.

[0055] Figure 9 A comparison chart of emergency data throughput for three resource allocation methods under different numbers of sensor nodes;

[0056] Figure 10 A comparison chart of emergency data throughput for three resource allocation methods under different retransmission counts;

[0057] Figure 11 A comparison chart of the average latency of emergency data under different numbers of sensor nodes for three resource allocation methods;

[0058] Figure 12 A comparison chart showing the average latency of emergency data under different retransmission counts for three resource allocation methods. Detailed Implementation

[0059] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0060] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0061] This embodiment provides a prediction-based method for emergency data resource allocation in energy harvesting wireless body area networks, including:

[0062] Each sensor node reports node status information to the coordinator, which includes at least energy status, data status, and link channel information.

[0063] The coordinator sends the node status information to the mobile edge computing platform;

[0064] The mobile edge computing platform predicts the channel quality of each time slot during the data transmission phase based on a prediction model.

[0065] Based on the prediction results, the mobile edge computing platform calculates the minimum transmission power and maximum transmittable data volume of each node in each time slot to meet the packet loss rate threshold constraint.

[0066] The mobile edge computing platform allocates the number of data transmission time slots to each node based on the node's data requirements, energy status, and transmission capabilities.

[0067] The mobile edge computing platform uses an optimization algorithm to allocate specific time slot sets and scheduling order to each node under the condition of meeting the preset minimum service constraints.

[0068] The coordinator broadcasts the resource allocation results to each node, and the resource allocation results include at least the set of time slots and the corresponding transmission parameters.

[0069] Each node adaptively selects the energy harvesting mode, transmission mode, or energy harvesting-transmission switching mode before the arrival of the allocated time slot based on its own real-time data status and energy status, and performs the corresponding energy harvesting and data transmission operations within the time slot.

[0070] Furthermore, the process by which each sensor node reports its node status information to the coordinator includes:

[0071] During the polling phase of the superframe, polling time slots are allocated to each sensor node;

[0072] Each sensor node adopts an energy harvesting-transmission switching mode within its polling time slot, dividing the polling time slot into an energy harvesting phase and a data transmission phase.

[0073] The node collects energy during the energy harvesting phase and sends link channel quality, node data status, and remaining energy to the coordinator during the data transmission phase.

[0074] Furthermore, the process by which the mobile edge computing platform predicts the channel quality for each time slot during the data transmission phase based on a predictive model includes:

[0075] Based on the historical link channel information reported by each node, an input sequence for channel prediction is constructed.

[0076] The input sequence is fed into a temporal convolutional network prediction model, which outputs a prediction sequence for the channel quality of each time slot in the next superframe.

[0077] Furthermore, the process by which the mobile edge computing platform allocates data transmission time slots to each node based on the node's data requirements, energy status, and transmission capabilities includes:

[0078] Calculate the normalized data demand factor based on the node's data cache size, regular data arrival rate, and urgent data volume;

[0079] Calculate the normalized energy factor based on the node's remaining energy and predicted collectable energy;

[0080] Calculate the normalized transmission capacity factor based on the maximum amount of data that a node can transmit in each time slot.

[0081] The number of time slots is allocated to each node based on the node score factor composed of the data demand factor, energy factor, and transmission capacity factor.

[0082] Furthermore, the mobile edge computing platform employs an optimization algorithm to allocate specific time slot sets and scheduling orders to each node under preset minimum service constraints. This process includes:

[0083] The total time slots during the data transmission phase are divided into multiple consecutive segments;

[0084] Set a minimum service constraint for each segment, requiring each node to be allocated at least a preset number of time slots in each segment;

[0085] A time-slot scheduling problem is constructed with the goal of maximizing the total amount of data that can be transmitted. This problem is then transformed into a minimum-cost flow problem for solution, resulting in a specific set of time slots for each node.

[0086] Furthermore, the process by which the coordinator broadcasts the resource allocation results to each node includes:

[0087] During the algorithm execution and energy broadcasting phase of the superframe, the coordinator sends radio frequency energy signals to each sensor node;

[0088] During the beacon phase of the superframe, the coordinator broadcasts to each node the resource allocation results, which include the time slot set, minimum transmission power, time switching factor, average maximum transmittable data volume, and emergency data reserve energy.

[0089] Furthermore, the process by which each node adaptively selects its working mode based on its own real-time data and energy status includes:

[0090] The node calculates the transmission pressure factor based on the remaining data volume, the remaining number of scheduling time slots, and the average maximum transmittable data volume.

[0091] Based on the transmission pressure factor, the presence of emergency data, and the comparison between remaining energy and emergency data reserve energy, the node selects one of the following: energy harvesting mode, transmission mode, or energy harvesting-transmission switching mode.

[0092] Furthermore, before each node performs the corresponding energy harvesting and data transmission operations within the time slot, a mode indication process is also included, comprising:

[0093] At the beginning of the allocated time slot, the node indicates the selected operating mode to the coordinator by sending a preset combination of pulse signals.

[0094] Specifically, the implementation process of this embodiment includes:

[0095] Example 1

[0096] This embodiment discloses a resource allocation method for low-latency transmission of emergency data in a predictive energy harvesting wireless body area network. The energy harvesting wireless body area network includes a coordinator and multiple sensor nodes. Each sensor node has an energy conversion circuit to harvest radio frequency energy and a supercapacitor to store the harvested energy. The coordinator aggregates data from each node and interacts with an external computing platform. The network topology is as follows: Figure 1 As shown. The coordinator and sensor nodes interact using superframes as the basic communication cycle. A superframe includes a beacon phase one (B1 phase), a polling phase (PP phase), an algorithm execution and energy broadcasting phase (CEP phase), a beacon phase two (B2 phase), and a time slot scheduling phase (MAP phase). The superframe structure is as follows: Figure 2 As shown; wherein, during the time slot scheduling phase, nodes can select energy harvesting mode, transmission mode, or energy harvesting-transmission switching mode based on data status, transmission pressure factor, and energy status. The time slot structure for the three working modes is as follows: Figure 3 As shown. Further, during the algorithm execution and energy broadcasting phases, the coordinator sends node status and link information to the mobile cloud computing platform and the mobile edge computing platform. The mobile cloud computing platform trains and updates the channel prediction model, and the mobile edge computing platform calls the latest model to complete online inference and resource allocation calculations. The system collaborative architecture is as follows: Figure 4 As shown. The overall process of the resource allocation method described in this embodiment is as follows: Figure 5 As shown, the main steps include node information reporting, channel prediction and parameter calculation, time slot allocation, time slot scheduling and broadcasting, and node selection and execution of time slot adaptive mode. The resource allocation method includes the following steps:

[0097] S1. During the superframe polling phase, nodes report their node status and link channel information to the coordinator. The polling phase is as follows: Figure 2 As shown:

[0098] S1.1 In the polling phase of the superframe, a set of polling time slots is set. The number of polling time slots is consistent with the number of sensor nodes. Each sensor node corresponds to one polling time slot, which is used to complete the reporting of node information.

[0099] S1.2 Each sensor node only performs information reporting operations within its corresponding polling time slot, and completes energy acquisition and information reporting in the order of "energy acquisition first, information transmission later" within the polling time slot; the sensor node adopts the energy acquisition-transmission switching mode for energy acquisition and information reporting, and the time switching factor is a preset fixed value, such as 0.5;

[0100] S1.3 The node status reported by each sensor node to the coordinator shall include at least the node's remaining energy and node data status, wherein the node's remaining energy is the currently available energy in the node's energy storage unit;

[0101] S1.4 The node data status includes at least the node data cache size, the regular data arrival rate, and the amount of emergency data generated in the previous superframe, wherein the regular data arrival rate is used to characterize the average data generation rate of the node's regular data per unit time.

[0102] S1.5 The link channel information reported by each sensor node to the coordinator shall include at least the channel quality information of the link from the node to the coordinator in the previous superframe; the coordinator receives and aggregates the information reported by each node as input for subsequent prediction calculation and time slot allocation and scheduling, so as to realize the periodic update of node energy status, data status and link status.

[0103] S2. The coordinator sends the information gathered in step S1 to the mobile edge computing platform and the mobile cloud computing platform. The mobile edge computing platform then performs channel prediction and transmission parameter calculation based on the prediction model. The channel prediction and parameter calculation process is as follows: Figure 4 As shown:

[0104] S2.1 The coordinator sends the node status and link channel information received in step S1 to the mobile edge computing platform and the mobile cloud computing platform. The mobile cloud computing platform is used to store historical channel information and train and update the channel prediction model, while the mobile edge computing platform is used to call the latest prediction model to perform online inference.

[0105] S2.2 The mobile cloud computing platform trains and updates the temporal convolutional network prediction model based on historical path loss data, and then sends the updated model parameters to the mobile edge computing platform to ensure prediction accuracy and real-time online computation. In this embodiment, the structure of the temporal convolutional network prediction model is as follows: Figure 6 As shown, it can be updated periodically; the temporal convolutional network prediction model consists of multiple residual convolutional modules, with the number of convolutional channels of each residual module being 32, 64 and 128 respectively, and the dilation factors being 1, 2 and 4 respectively. Finally, the path loss prediction value is output through a fully connected layer.

[0106] S2.3 The mobile edge computing platform performs sampling at preset time intervals. The path loss information is serialized, and an input sequence for channel prediction is constructed. Let the superframe duration be... The number of sampling points in each superframe is:

[0107] ;

[0108] node The path loss sequence obtained from sampling in the previous superframe is represented as follows:

[0109] ;

[0110] in Represents a node The The path loss obtained from sampling at each sampling point ;

[0111] For ease of understanding, the input sampled values ​​and path loss observations can be equivalently represented as follows:

[0112] ;

[0113] in Represents a node In the Path loss observations at each sampling time;

[0114] S2.4, The mobile edge computing platform will input the sequence Input the path loss prediction model of the temporal convolutional network to obtain the nodes. The predicted path loss sequence within this superframe. The output predicted sequence is:

[0115] ;

[0116] in Represents a node The Predicted path loss for each sampling point , The number of sampling points within each superframe;

[0117] The predictive relationship is represented as:

[0118] ;

[0119] in, For nodes The path loss sequence obtained by sampling in the previous superframe. Here is the path loss prediction function based on a temporal convolutional network; for ease of understanding, the prediction output is equivalent to the predicted path loss as follows:

[0120] ;

[0121] in, Represents a node In the The predicted path loss value at each sampling time.

[0122] Furthermore, to facilitate correspondence with subsequent time slot scheduling stages, the sampling point index is... Mapped to slot index Assume the time slot scheduling phase includes There are 1 basic time slot, and the length of each basic time slot is 1. Then the first The set of sampling points corresponding to each time slot is represented as follows:

[0123] ;

[0124] in, This is the preset sampling time interval.

[0125] In this embodiment, to reduce online computational overhead and facilitate engineering implementation, it is assumed that the path loss remains constant within the same time slot; that is, all sampling points within the time slot use the same predicted path loss value to represent the channel state of that time slot. Specifically, nodes In the time slot The predicted path loss is defined as:

[0126] ;

[0127] in, Represents a node In the The predicted path loss value at each sampling time. Represents a node Time slots in the time slot scheduling phase The predicted path loss value, , This represents the total number of time slots during the time slot scheduling phase. This leads to the node... Slot-level predicted path loss sequence during the slot scheduling phase ;

[0128] S2.5 The mobile edge computing platform calculates the minimum transmission power required to meet the packet error rate threshold based on predicted path loss. In transmission mode, this ensures that the packet error rate of nodes does not exceed the preset threshold. This requires the receiver to meet a minimum received power constraint. Specifically, in this embodiment, the node uses differential binary phase shift keying modulation and demodulation, and does not employ channel coding and error correction mechanisms. Therefore, the minimum received power of the receiver can be derived based on the packet error rate constraint of differential binary phase shift keying modulation and demodulation in an additive white Gaussian noise channel:

[0129] ;

[0130] in, For nodes Data rate, For noise power, For bandwidth, For data packet size, For nodes In the time slot The receiving power, For nodes In the time slot Satisfying the packet error rate threshold The minimum received power. Since the transmit power, receive power, and path loss satisfy:

[0131] ;

[0132] in, For nodes In the time slot Transmission power, For nodes In the time slot The receiving power, Represents a node In the time slot The predicted path loss value.

[0133] Therefore, node In the time slot Meets the packet error rate threshold The minimum transmission power is:

[0134] ;

[0135] in, For nodes In the time slot The minimum transmission power is the lower bound constraint of the transmission power in the energy sampling-transmission switching mode in step S2.5.

[0136] S2.6 The mobile edge computing platform determines the time switching factor for the energy harvesting-transmission switching mode based on energy causal constraints and calculates the maximum amount of data that a node can transmit within a time slot. In the energy harvesting-transmission switching mode, the node... In the time slot Energy extraction time within the area is Data transmission time is Then the energy consumption of the node in transmitting data within this time slot is:

[0137] ;

[0138] in, The time switching factor for the energy harvesting-transmission switching mode. The time slot length, For nodes In the time slot The transmission power.

[0139] The collectable energy of the node in this time slot is:

[0140] ;

[0141] in, For energy conversion efficiency, The power of the coordinator's broadcast energy. For nodes In the time slot The channel gain. The channel gain can be obtained from the predicted path loss:

[0142] ;

[0143] in For nodes In the time slot The predicted path loss value.

[0144] In the worst-case scenario, assuming no initial residual energy at the nodes, energy causality constraints must be satisfied. ,have to:

[0145] ;

[0146] because The smaller the value, the longer the data transmission time and the larger the amount of data that can be transmitted. Therefore, the minimum time switching factor can be obtained as:

[0147] ;

[0148] in, For nodes In the time slot Meets the packet error rate threshold Minimum transmission power, For nodes In the time slot The minimum time switching factor.

[0149] Then node In the time slot The maximum amount of data that can be transferred within is:

[0150] ;

[0151] in, For nodes Data rate, This represents the time slot length.

[0152] This yields the node. Transmittable capacity sequence during time slot scheduling phase This is used for subsequent resource allocation and time slot scheduling decisions;

[0153] S2.7, The mobile edge computing platform calculates the collectable energy of nodes during the algorithm execution and energy broadcasting phases based on predicted path loss. Assume the algorithm execution and energy broadcasting phases include... There are 1 time slot, and the length of each time slot is also 1. The coordinator's transmission power during the algorithm execution and energy broadcasting phases is To facilitate correspondence with the prediction results at the sampling points, the time slot index of the algorithm execution and the energy broadcasting phase is denoted as... and define nodes During the time slots of algorithm execution and energy broadcasting phases The predicted path loss is .in, Predicted path loss with sampling points The relationship is consistent with the time slot mapping in step S2.4, that is, the set of sampling points belonging to this time slot. The predicted values ​​within the range are aggregated to obtain:

[0154] ;

[0155] Then node In the time slot The predicted received power satisfies:

[0156] ;

[0157] Therefore, we can obtain the node. The predicted collectable energy during the algorithm execution and energy broadcasting phases is:

[0158] ;

[0159] in, Energy conversion efficiency.

[0160] S3, The total number of time slots for the mobile edge computing platform during the time slot scheduling phase is: Under the constraints of node data requirements, available energy, and transmission capacity, the number of data transmission time slots available for each node during the time slot scheduling phase is allocated as follows:

[0161] S3.1, The total number of time slots in the time slot scheduling phase of the mobile edge computing platform is: Under these conditions, in order to meet the minimum service constraints of nodes, pre-determine the nodes. The number of allocated time slots is And calculate the number of remaining allocable time slots. , represented as:

[0162] ;

[0163] in, This refers to the number of sensor nodes in the body area network. Furthermore, to satisfy the segmented minimum service constraint in step S4, the guaranteed number of time slots... satisfy:

[0164] ;

[0165] in, The number of segments in the time slot scheduling phase. This refers to the minimum number of service slots within each segment. In this embodiment... ;

[0166] S3.2, The mobile edge computing platform is based on nodes Remaining data cache size Regular data arrival rate And the amount of urgent data generated by the previous superframe To predict the amount of data that needs to be transmitted in this superframe, the node is obtained. Data demand metrics , represented as:

[0167] ;

[0168] in, This refers to the superframe duration. The mobile edge computing platform normalizes the data demand metrics to obtain the node... Data demand factors Represented as:

[0169] ;

[0170] in, For nodes Data requirement metrics and These are the minimum and maximum values ​​of the data demand indicators for all nodes, respectively.

[0171] S3.3, Mobile edge computing platform based on nodes Remaining energy The nodes predicted in step S2.7 can collect energy during the time slot scheduling phase. Construct nodes Initial available energy index , represented as:

[0172] ;

[0173] The mobile edge computing platform normalizes the initial available energy index to obtain the node. energy factor , represented as:

[0174] ;

[0175] in, For nodes The initial available energy index, and These are the minimum and maximum values ​​of the initial available energy index for all nodes, respectively;

[0176] S3.4 The mobile edge computing platform is based on the nodes obtained in step S2.6 Maximum transmissible data volume in each time slot during the time slot scheduling phase compute nodes Average maximum transmissible data volume across all time slots during the time slot scheduling phase , represented as:

[0177] ;

[0178] in This represents the total number of time slots during the time slot scheduling phase.

[0179] The mobile edge computing platform will... As a node The transmission capacity index is obtained and normalized to obtain the node. Transmission capacity factor , represented as:

[0180] ;

[0181] in, and These represent the maximum and minimum values ​​of the transmission capacity indicators for all nodes, respectively.

[0182] S3.5 The mobile edge computing platform utilizes the nodes obtained in steps S3.2, S3.3, and S3.4. Data demand factors Energy factors With transmission capacity factor Construct nodes Scoring factors Based on the scoring factor, the remaining allocable time slots are allocated to each node proportionally. , represented as:

[0183] ;

[0184] ;

[0185] in, These are the weighting coefficients. For nodes The theoretical number of allocated time slots, Obtained from step S3.1, This represents the number of sensor nodes in the body area network.

[0186] S3.6 The mobile edge computing platform will allocate the theoretical number of time slots obtained in step S3.5. Rounding up yields the node. Basic allocation of time slots And calculate the corresponding decimal remainder. , represented as:

[0187] ;

[0188] ;

[0189] The mobile edge computing platform further calculates the number of unallocated remaining time slots:

[0190] ;

[0191] in, This represents the number of sensor nodes in the body area network. The remaining number of allocable time slots obtained in step S3.1;

[0192] when At that time, the mobile edge computing platform will sort all nodes by fractional remainder. Sort from largest to smallest, and then use the sorting results as the top choices. Each node is allocated an additional time slot, thus obtaining the number of additional time slots allocated. If there are multiple nodes with decimal remainders If the nodes are equal, they are allocated priority according to their node numbers in ascending order. Ultimately, the nodes... Time slot allocation results Represented as:

[0193] ;

[0194] in, The number of guaranteed time slots in step S3.1, Based on the allocation of time slots, To allocate additional time slots.

[0195] S4. The mobile edge computing platform obtains the time slot allocation results for each node output in step S3. Subsequently, the time slots in the time slot scheduling phase are further segmented, and a specific set of time slot numbers is allocated to each node under the premise of satisfying the minimum service constraint of each segment, so as to achieve time slot scheduling with maximum throughput. The time slot allocation process can be found in [reference needed]. Figure 7 :

[0196] S4.1, the mobile edge computing platform will schedule the time slots during the time slot phase. Each time slot is divided into A set of segments ,in Indicates the first The set of time slots within each segment, and satisfying as well as ;when Unable to be When dividing by a whole number, the remaining time slots are merged into the last segment. ;

[0197] S4.2 The mobile edge computing platform satisfies the node allocation time slot number obtained in step 3.6. Based on this, a segmented minimum service constraint is introduced to limit the maximum waiting interval of a node during the time slot scheduling phase, that is, to constrain each node. In each segment At least one allocation Each time slot is represented as:

[0198] ;

[0199] in, Indicates time slot Is it assigned to a node? , Indicates the time slot Assigned to node ,otherwise ; The minimum number of service slots within each segment, typically set as follows: ;

[0200] S4.3, The mobile edge computing platform is constructed based on the throughput maximization time slot scheduling optimization problem under the condition of satisfying the segmented minimum service constraint, and uses binary variables. This indicates the time slot allocation relationship, where Indicates the time slot Assigned to node ,otherwise The optimization problem is expressed as:

[0201] ;

[0202] And satisfy the following constraints:

[0203] ;

[0204] ;

[0205] ;

[0206] ;

[0207] in The nodes calculated in step S2.6 In the time slot The maximum amount of data that can be transmitted. Assigned to the node in step 3.6 The number of time slots, This represents the number of sensor nodes in the body area network. This represents the total number of time slots during the time slot scheduling phase. The minimum number of service slots within each segment. The number of segments in the time slot scheduling phase. For the first The set of time slots within each segment;

[0208] S4.4 The mobile edge computing platform transforms the throughput maximization time slot scheduling optimization problem described in step S4.3 into a minimum cost maximum flow problem, and achieves a joint solution for constraint satisfaction and throughput maximization by constructing a flow network with a "segmented constraint layer and a free allocation layer". Specifically, the mobile edge computing platform constructs a flow network including a source node... Exchange Point Piecewise constraint layer vertex set Freely allocate layer vertex set and the set of time slot layer vertices Given a directed flow network, define the following edges and their capacities and costs, where... This represents the number of sensor nodes in the body area network. This represents the total number of time slots during the time slot scheduling phase. The number of segments in the time slot scheduling phase. Minimum number of service slots within each segment:

[0209] (1) Source point to piecewise constraint layer edge: from source point To the vertex of the segmented constraint layer Establish directed edges Its capacity setting node In segments The internal requirement is to meet the minimum number of slots needed for protection. This is used to enforce the minimum service constraint in segments; where, when a node Total number of allocated time slots satisfy season:

[0210] ;

[0211] when To avoid scheduling problems becoming infeasible, the mobile edge computing platform adaptively adjusts the minimum number of service time slots in each segment to a lower bound of "as evenly as possible distribution," that is:

[0212] or ;

[0213] Make And prioritize allocating the extra 1 to segments with smaller segment indexes;

[0214] (2) Source point to free allocation layer edge: from source point To the free allocation layer vertex Establish directed edges Its capacity is set as a node The remaining number of allocable time slots is expressed as This is used to perform a flexible allocation of the remaining time slots to maximize throughput after satisfying the minimum service constraint of the segment;

[0215] (3) Edge from segmented constraint layer to time slot layer: from the vertex of segmented constraint layer To belong to the segmentation The top of the time slot layer inside Establish directed edges Its capacity is set to 1, indicating that a time slot can be allocated at most once; and its cost is set to the negative of the maximum amount of data a node can transmit in that time slot to maximize throughput, expressed as:

[0216] ;

[0217] in, The nodes calculated in step S2.6 In the time slot The maximum amount of data that can be transmitted;

[0218] (4) From the edge of the free allocation layer to the time slot layer: from the vertex of the free allocation layer To any time slot layer vertex Establish directed edges Its capacity is also set to 1, and the cost is set to:

[0219] ;

[0220] in The nodes calculated in step S2.6 In the time slot The maximum amount of data that can be transmitted;

[0221] (5) From the time slot layer to the sink edge: from the vertex of the time slot layer To the remittance point Establish directed edges Its capacity is set to 1 to limit each time slot to a maximum of one node. The mobile edge computing platform performs a minimum-cost maximum-flow solution on the flow network to maximize global throughput while satisfying the segmented minimum service constraint and the node time slot number constraint, thereby obtaining the time slot allocation result. Based on the allocation results, the time slot set for each node during the time slot scheduling phase is determined:

[0222] ;

[0223] Furthermore, the mobile edge computing platform can utilize the time slot set The time slot indexes within the node are sorted in ascending order to determine the transmission timing of the node;

[0224] S4.5 The mobile edge computing platform is based on the node time slot set obtained in step S4.4. compute nodes Average maximum transmissible data volume in scheduled time slots , represented as:

[0225] ;

[0226] in, , Assigned to the node in step S3 The number of time slots, The nodes calculated in step S2.6 In the time slot The maximum amount of data that can be transmitted;

[0227] S4.6, Mobile edge computing platform obtains nodes The set of allocated time slots Subsequently, to ensure timely transmission of emergency data, the minimum transmission power of the node in each allocated time slot is determined based on step S2.5. Reserve energy required for emergency data transmission of computing nodes , represented as:

[0228] ;

[0229] in, The node determined in step S4.4 The set of allocated time slots, This represents the time slot length.

[0230] S5. After completing the path loss prediction and transmission parameter calculation in step S2, the time slot allocation in step S3, and the time slot scheduling in step S4, the coordinator sends the resource allocation results to the sensor nodes, and each sensor node completes energy acquisition and data transmission according to the allocation results. The process specifically includes the following steps:

[0231] S5.1 During the algorithm execution and energy broadcasting phase, the coordinator sends radio frequency energy signals to each sensor node, enabling each sensor node to collect energy from the radio frequency energy signals through the energy conversion circuit and store them in the supercapacitor.

[0232] S5.2 After completing the time slot allocation in step S3 and the time slot scheduling in step S4, the mobile edge computing platform sends the resource allocation results to the coordinator; the resource allocation results include at least the time slot set of each node during the time slot scheduling phase. Minimum transmission power corresponding to each time slot Time switching factor in energy harvesting-transmission switching mode The average maximum amount of data that a node can transmit on scheduled time slots. and emergency data reserve capacity ;

[0233] S5.3. During the second phase of beacon operation, the coordinator broadcasts the resource allocation results to each sensor node, enabling each sensor node to allocate resources according to the set of allocated time slots. The corresponding parameter configurations are used to perform energy acquisition and data transmission in the subsequent time slot scheduling phase. Each sensor node only receives and parses the resource allocation information corresponding to its own node.

[0234] S6. After receiving the resource allocation results broadcast in step S5, each sensor node adaptively selects its operating mode within each allocated time slot based on its node data status, remaining energy status, and the resource parameters. It then feeds back the operating mode for that time slot to the coordinator via an indication phase. The node's adaptive operating mode selection process according to time slot is as follows: Figure 8 As shown, the specific steps include:

[0235] S6.1 Before each allocated time slot arrives, each node determines the amount of data remaining on its node. Number of remaining scheduling time slots and the average maximum transmittable data amount obtained in step S5.2 Calculate the transmission pressure factor , represented as:

[0236] ;

[0237] And With preset pressure threshold The comparison is used to characterize the urgency of a node's time slot resources. In this embodiment, the preset pressure threshold... The preferred setting is 1, which characterizes the critical state where the node's average transmittable capacity within the remaining scheduling time slots is equivalent to the amount of data to be transmitted. This indicates that the transmission pressure is high, and the node prioritizes increasing its data sending capacity.

[0238] S6.2. The node determines its data status. If the node has no data to be transmitted, then the following condition is met. ,in For nodes The remaining data cache size. The node selects the energy harvesting mode and performs energy harvesting only within the corresponding time slot;

[0239] S6.3 If the data to be transmitted by a node contains urgent data, then the node... Emergency data volume satisfy .node According to the allocated time slot minimum transmission power Calculate time slots in transmission mode Transmittable time And based on the time switching factor obtained in step S5.2 Time slots in the energy harvesting-transmission switching mode Transmittable time , respectively represented as:

[0240] ;

[0241] ;

[0242] in, For nodes The remaining energy, This represents the time slot length. Node comparison. and ,like If the condition is met, select the transmission mode; otherwise, select the energy acquisition-transmission switching mode to improve the timely transmission capability of emergency data.

[0243] S6.4 If a node contains only regular data, then the following conditions must be met: When the transmission pressure factor satisfies This indicates that the transmission pressure is relatively high, so the node prioritizes increasing its data transmission capacity and compares according to step S6.3. and ,like If yes, select the transmission mode; otherwise, select the energy acquisition-transmission switching mode.

[0244] S6.5, When the transmission pressure factor satisfies This indicates relatively low transmission pressure, and the node further determines its transmission based on remaining energy. Emergency data reserve energy obtained in step S4.6 To make a judgment: If If so, then select the energy harvesting mode; if Then, surplus energy is defined as follows:

[0245] ;

[0246] And calculate the time slots of surplus energy in the transmission mode. Transmittable time , represented as:

[0247] ;

[0248] in, For nodes In the time slot Minimum transmission power, The time slot length;

[0249] Node comparison The result obtained from step S6.3 ,like If yes, select the transmission mode; otherwise, select the energy acquisition-transmission switching mode.

[0250] S6.6 After determining the working mode of this time slot, the node sets two consecutive indication phases at the beginning of the allocated time slot, which are respectively denoted as the first indication phase. With the second instruction phase Each indication phase lasts for tens of microseconds. During each indication phase, the node feeds back its operating mode for the current time slot to the coordinator by transmitting or not transmitting a pulse signal. Transmitting a pulse signal represents "1", and not transmitting a pulse signal represents "0". The pulse signal is a pulse detection signal that requires no modulation or demodulation. Thus, the node... Selected working mode for combined encoding: When When, it indicates that the node selects the energy harvesting mode; when When, it indicates that the node selects the energy sampling-transmission switching mode; when When the time slot is selected, it indicates that the node has chosen a transmission mode. After the coordinator receives and parses the indication signal, the node performs energy harvesting and data transmission according to the selected mode in the subsequent stages of the time slot. When the energy harvesting mode is selected, only energy harvesting is performed, and when the transmission mode is selected, the transmission power obtained in step S5.2 is used. When sending data and selecting the energy acquisition-transmission switching mode, the time switching factor obtained in step S5.2 is used. The time slot is divided into an energy harvesting phase and a data transmission phase. Nodes collect energy during the energy harvesting phase and transmit the energy using the power obtained in step S5.2 during the data transmission phase. Send data. Furthermore, to ensure the indicated phase... and Timing synchronization can be achieved. In this example, the crystal oscillator frequency error between the sensor nodes and the coordinator in the body area network is typically in the range of 10~30ppm. Calculated using the worst-case error of 30ppm, when the superframe length is 300 milliseconds, the cumulative time offset introduced by the crystal oscillator error is approximately 9 microseconds, which is negligible compared to the 50-microsecond duration of the indication signal. Therefore, it can support the coordinator's reliable synchronous detection of the indication signal.

[0251] Example 2

[0252] Referring to the prediction-based low-latency transmission resource allocation method for emergency data in energy harvesting wireless body area networks disclosed in Example 1 (steps S1 to S6), this example further evaluates the performance of the proposed scheme using Monte Carlo simulation based on channel data measured in a real wireless body area network. Specifically, this example builds a simulation platform using Python and simulates the network operation process on an actual measured channel. The network operation and resource allocation process is consistent with that of Example 1. Based on this, this example focuses on testing and analyzing the changing trend of emergency data throughput in energy harvesting body area networks under different numbers of sensor nodes and different retransmission counts, and compares and verifies the proposed scheme with two comparative schemes. The key parameters used in this simulation are as follows: superframe length is 300ms, time slot length is 2ms, algorithm execution and energy broadcast phase length is 45ms, beacon second phase length is 10ms; coordinator transmission power is fixed at 0dBm, node maximum transmission power is 0dBm; data packet size is 256 bits, data rate is 1024kbps, system bandwidth is 1MHz, noise power is −110dBm, and energy conversion efficiency is 0.8. To reduce the influence of random factors and improve statistical reliability, all simulation results are obtained by averaging multiple independent simulation runs. Figure 9 and Figure 10 The results of emergency data throughput comparisons between the method of this invention (Prediction-driven Emergency Low-latency Dynamic Resource Allocation Strategy, PELDRA), the first comparative method (PASA) proposed by Hu et al., and the second comparative method (ATMAC) proposed by Qi et al., under different numbers of sensor nodes and different retransmission times are presented.

[0253] Figure 9The emergency data throughput of PELDRA, PASA, and ATMAC is compared under different numbers of sensor nodes. As shown in the figure, the emergency data throughput of all three schemes generally increases as the number of nodes increases from 6 to 10. This is because the probability of emergency data generation and the amount of data that can be transmitted increase simultaneously with the increase in node size, thus increasing the overall emergency data transmission capacity of the system. Based on this, the proposed PELDRA consistently achieves the highest throughput across all node sizes. Its key lies in its prediction-driven dynamic resource allocation mechanism, which can more finely adapt transmission resources and parameters to changes in channel state. Simultaneously, by ensuring that node transmission opportunities cover the entire data transmission phase, it guarantees that emergency data generated within a superframe can obtain available time slots more promptly, increasing the probability of successful transmission and significantly improving emergency data throughput. In contrast, ATMAC uses fixed transmission power and continuously aggregates time slots, while PASA, although possessing superframe-scale phase length and power adjustment capabilities, struggles to track rapid dynamic changes in the channel. Both of these limitations restrict emergency data transmission efficiency, resulting in lower throughput than PELDRA.

[0254] Figure 10 The results compare the emergency data throughput of the PELDRA scheme with the PASA and ATMAC schemes under different retransmission counts. As shown in the figure, as the allowed retransmission count increases from 1 to 5, the emergency data throughput of all three schemes generally increases, but the rate of increase gradually decreases and tends to level off. This is because with an increased retransmission count, data packets have more opportunities to retransmit when encountering instantaneous channel fading or collision failures, thus increasing the probability of successful transmission of emergency data. However, when the retransmission count further increases, the throughput gain brought by additional retransmissions gradually saturates due to the constraints of available transmission time slots within the superframe and available node energy. Further comparison shows that the proposed PELDRA scheme consistently maintains the highest emergency data throughput under all retransmission counts, and its advantage is more significant under low retransmission counts (e.g., 1-2 times). The reason is that PELDRA, through its prediction-driven resource allocation and parameter adaptation mechanism, can proactively avoid low-quality channel conditions and fine-grained adjust transmission parameters, reducing the probability of initial transmission failure from the source. Simultaneously, PELDRA's time slot scheduling ensures higher accessibility of urgent data during transmission phases, reducing data loss and delay caused by waiting for the next superframe or insufficient time slots, thus achieving higher effective throughput even with limited retransmission opportunities. In contrast, ATMAC uses fixed transmission power, making it difficult to quickly compensate for channel fluctuations, resulting in a higher initial transmission failure rate and relying more on increasing the number of retransmissions to improve throughput. While PASA can adjust transmission power and phase length at the superframe scale, its power control update granularity is coarse, making it difficult to match rapid channel changes in a timely manner; therefore, its throughput is still lower than PELDRA. In summary, Figure 10The results show that the PELDRA scheme proposed in this embodiment can achieve higher emergency data throughput under different retransmission number constraints, verifying its robustness and effectiveness under dynamic channel and energy-constrained conditions.

[0255] Example 3

[0256] Referring to the prediction-based low-latency transmission resource allocation method for emergency data in energy harvesting wireless body area networks disclosed in Example 1 (steps S1 to S6), this example further evaluates the performance of the proposed scheme using Monte Carlo simulation based on channel data measured in a real wireless body area network. Specifically, this example uses Python to build a simulation platform and simulates the network operation process on an actual measured channel. The network operation and resource allocation process is consistent with Example 1. Based on this, this example focuses on testing and analyzing the changing trend of the average latency of emergency data in the energy harvesting body area network under different numbers of sensor nodes and different retransmission counts, and compares and verifies the proposed scheme with two other schemes. Some important parameters used in the simulation and the calculation methods of the results in this example are consistent with those in Example 2. Figure 11 and Figure 12 The average latency of emergency data under different numbers of sensor nodes and different retransmission times is compared between the method of this invention (PELDRA), the comparative method 1 (PASA) proposed by Hu et al., and the comparative method 2 (ATMAC) proposed by Qi et al.

[0257] Figure 11 The average emergency data latency comparison results are presented for the PELDRA scheme, PASA scheme, and ATMAC scheme under different sensor node numbers. Figure 11It can be seen that as the number of nodes increases from 6 to 10, the average latency of emergency data in the PASA and ATMAC schemes shows a slow upward trend, while the average latency of emergency data in the proposed PELDRA scheme remains at a low level, only slightly changing with the increase in node size. This indicates that with the increase in the number of nodes and the increase in network load, the comparative schemes are more likely to encounter situations where emergency data is waiting for transmission resources, thus leading to an increase in average latency. Further comparison shows that PELDRA is significantly better than PASA and ATMAC at all node sizes, and can reduce the average latency of emergency data to a level far lower than that of the comparative schemes. The key reason is that the proposed PELDRA, through a low-latency time slot scheduling mechanism for emergency data, allows the node's transmission opportunities to cover the entire data transmission stage. When emergency data is generated at any time within the superframe, it can still quickly obtain an available transmission time slot within the current superframe and complete the upload, thus significantly reducing the waiting time. At the same time, PELDRA combines prediction-driven dynamic resource allocation to adaptively adjust transmission parameters, reducing failure retransmissions and additional queuing due to channel fluctuations, further compressing the end-to-end latency of emergency data. In contrast, both ATMAC and PASA exhibit a "continuous clustering by node" characteristic in time slot allocation. After a node uses up its continuous time slots, it must wait for the next superframe to reallocate resources. This often results in urgent data not being transmitted in time within the generated superframe, which is the main reason for their significantly higher average latency, which increases with the number of nodes. In summary, Figure 11 The results show that the PELDRA scheme proposed in this invention can effectively reduce the average latency of emergency data under different node scales, and has good network scale adaptability and emergency data protection capabilities.

[0258] Figure 12 The average delay of emergency data is compared between the PELDRA scheme, PASA scheme, and ATMAC scheme under different retransmission counts. Figure 12It can be seen that as the allowed number of retransmissions increases from 1 to 5, the average latency of emergency data in the PASA and ATMAC schemes shows a significant upward trend, while the average latency of emergency data in the proposed PELDRA scheme remains at a low level, showing only a slight increase. This indicates that in the comparative schemes, increasing the number of retransmissions introduces more redundant transmissions and channel occupancy, thus increasing the waiting time of emergency data in the queue and the time required to complete transmission, ultimately leading to an increase in average latency. Further comparison shows that PELDRA can significantly reduce the average latency of emergency data under all retransmission counts, and its advantage remains stable under both low and high retransmission counts. The key reason is that PELDRA, through prediction-driven dynamic resource allocation and parameter adaptation mechanisms, can avoid low-quality channel conditions in advance and reduce the probability of initial transmission failure, reducing retransmission triggers from the source; at the same time, PELDRA adopts a time slot scheduling method oriented towards emergency data, enabling emergency data to quickly obtain transmission opportunities after being generated within the superframe, avoiding long waiting times caused by retransmission occupying additional time slots, thus maintaining low latency under different retransmission count constraints. In contrast, ATMAC, due to its fixed transmission power and difficulty in adapting to dynamic channel changes, has a higher probability of initial transmission failure. As the number of retransmissions increases, it is more prone to consecutive retransmissions, consuming more time slot resources and resulting in a more significant increase in the average latency of urgent data. While PASA can adjust transmission power and stage length at the superframe scale, its power control update granularity is coarse, making it difficult to match rapid channel fluctuations in a timely manner, still incurring significant retransmission overhead, resulting in a significantly higher latency level than PELDRA. In summary, Figure 12 The results show that the PELDRA scheme proposed in this invention can effectively suppress the growth of emergency data delay and significantly reduce the average delay under different retransmission conditions, verifying its low-latency guarantee capability and robustness under dynamic channel and retransmission constraints.

[0259] In this embodiment, a computer terminal device is provided, including:

[0260] One or more processors;

[0261] A memory, coupled to the processor, for storing one or more programs;

[0262] When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the above-described method for emergency data resource allocation in a prediction-based energy harvesting wireless body area network.

[0263] In this embodiment, a computer-readable storage medium is also provided, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the above-described method for emergency data resource allocation in a prediction-based energy harvesting wireless body area network.

[0264] This invention provides a prediction-based emergency data resource allocation method for energy harvesting wireless body area networks (WBAs), which can reduce emergency data transmission latency, improve the reliability and reachability of emergency data transmission, and enhance system throughput efficiency and long-term operational stability under uncertain energy conditions. By constructing a low-latency resource allocation framework for emergency data, this invention overcomes the shortcomings of existing technologies that focus on average throughput or fairness and lack a mechanism to guarantee low latency for emergency data, enabling priority low-latency transmission of emergency data under energy-constrained conditions. The node adaptive working mode selection mechanism proposed in this invention allows nodes to dynamically switch between energy harvesting mode, transmission mode, and energy harvesting-transmission switching mode, thereby overcoming the problem that existing fixed workflows cannot adapt to dynamic changes in energy and data, effectively reducing emergency data waiting latency. This invention introduces a channel prediction method based on temporal convolutional networks to drive resource allocation based on the prediction of future time slot channel quality, and adaptively calculates the minimum transmission power and energy harvesting-transmission switching parameters accordingly, overcoming the problems of frequent retransmissions and energy waste caused by existing methods relying on instantaneous or statistical information, and improving the effective transmission capability of a single time slot. The prediction-driven low-latency time slot scheduling method proposed in this invention introduces segmented minimum service constraints to limit the longest waiting interval of nodes on the basis of satisfying the node time slot quota, and combines the minimum cost flow algorithm to achieve global scheduling optimization. It overcomes the problem that existing scheduling strategies lack an emergency data access latency control mechanism, and can improve the overall throughput performance of the system while ensuring low latency of emergency data.

[0265] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for emergency data resource allocation in a predictive energy harvesting wireless body area network, characterized in that, Includes the following steps: Each sensor node reports node status information to the coordinator, which includes at least energy status, data status, and link channel information. The coordinator sends the node status information to the mobile edge computing platform; The mobile edge computing platform predicts the channel quality of each time slot during the data transmission phase based on a prediction model. Based on the prediction results, the mobile edge computing platform calculates the minimum transmission power and maximum transmittable data volume of each node in each time slot to meet the packet loss rate threshold constraint. The mobile edge computing platform allocates the number of data transmission time slots to each node based on the node's data status, energy status, and transmission capacity. The mobile edge computing platform uses an optimization algorithm to allocate specific time slot sets and scheduling order to each node under the condition of meeting the preset minimum service constraints. The coordinator broadcasts the resource allocation results to each node, and the resource allocation results include at least the set of time slots and the corresponding transmission parameters. Each node adaptively selects the energy harvesting mode, transmission mode, or energy harvesting-transmission switching mode before the arrival of the allocated time slot based on its own real-time data status and energy status, and performs the corresponding energy harvesting and data transmission operations within the time slot. The process by which each sensor node reports its node status information to the coordinator includes: During the polling phase of the superframe, polling time slots are allocated to each sensor node; Each sensor node adopts an energy harvesting-transmission switching mode within its polling time slot, dividing the polling time slot into an energy harvesting phase and a data transmission phase. The node collects energy during the energy harvesting phase and sends the link channel quality, node data status, and remaining energy of the node to the coordinator during the data transmission phase. The process by which a mobile edge computing platform predicts the channel quality for each time slot during data transmission based on a predictive model includes: Based on the historical link channel information reported by each node, an input sequence for channel prediction is constructed. The input sequence is fed into a temporal convolutional network prediction model, which outputs a prediction sequence for the channel quality of each time slot in the next superframe. The process by which a mobile edge computing platform allocates data transmission time slots to each node based on its data status, energy status, and transmission capacity includes: Calculate the normalized data demand factor based on the node's data cache size, regular data arrival rate, and urgent data volume; Calculate the normalized energy factor based on the node's remaining energy and predicted collectable energy; Calculate the normalized transmission capacity factor based on the maximum amount of data that a node can transmit in each time slot. Based on the node score factor composed of the data demand factor, energy factor, and transmission capacity factor, the number of time slots is allocated to each node. The mobile edge computing platform employs an optimization algorithm to allocate specific time slot sets and scheduling orders to each node under preset minimum service constraints. This process includes: The total time slots during the data transmission phase are divided into multiple consecutive segments; Set a minimum service constraint for each segment, requiring each node to be allocated at least a preset number of time slots in each segment; A time-slot scheduling problem is constructed with the goal of maximizing the total amount of data that can be transmitted. This problem is then transformed into a minimum-cost flow problem for solution, resulting in a specific set of time slots for each node.

2. The method according to claim 1, characterized in that, The process by which the coordinator broadcasts the resource allocation results to each node includes: During the algorithm execution and energy broadcasting phase of the superframe, the coordinator sends radio frequency energy signals to each sensor node; During the beacon phase of the superframe, the coordinator broadcasts to each node the resource allocation results, which include the time slot set, minimum transmission power, time switching factor, average maximum transmittable data volume, and emergency data reserve energy.

3. The method according to claim 1, characterized in that, The process by which each node adaptively selects its working mode based on its real-time data and energy status includes: The node calculates the transmission pressure factor based on the remaining data volume, the remaining number of scheduling time slots, and the average maximum transmittable data volume. Based on the transmission pressure factor, the presence of emergency data, and the comparison between remaining energy and emergency data reserve energy, the node selects one of the following: energy harvesting mode, transmission mode, or energy harvesting-transmission switching mode.

4. The method according to claim 1, characterized in that, Before each node performs the corresponding energy harvesting and data transmission operations within the time slot, a mode indication process is also included, which includes: At the beginning of the allocated time slot, the node indicates the selected operating mode to the coordinator by sending a preset combination of pulse signals.

5. A computer terminal device, characterized in that, include: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors perform the steps of the method as described in any one of claims 1-4.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1-4.

Images