A universal mobile internet of things multi-modal data dynamic transmission system and method

Abstract

The application relates to a universal mobile Internet of Things multi-modal data dynamic transmission system and method, a motion-channel attenuation mapping model is constructed, electromagnetic frequency band state data of the surrounding environment of an Internet of Things device is acquired, short-term channel quality is predicted in combination with the model, a multi-protocol score array is generated, an optimal transmission path is selected, the clock of the device is calibrated to synchronization, data is reorganized, and transmission is realized; the system integrates IMU and passive sensing devices in a terminal device, and real-time collection of device motion states and electromagnetic environment data is realized; an edge terminal including a distributed edge computing node receives information transmitted by a device terminal and realizes dynamic protocol switching and resource scheduling, and timing calibration and repair of multi-protocol data are realized at a gateway terminal; and data is output after reorganization and security inspection. The application significantly improves the channel quality prediction accuracy and transmission reliability in a mobile scene, guarantees the whole-link priority occupation of emergency medical data, synchronizes errors, and meets the application requirements of high reliability and low time delay.

Description

Technical Field

[0001] This invention relates to the technical field of digital information transmission, such as telegraph communication transmission, and particularly to a universal mobile Internet of Things (IoT) multimodal data dynamic transmission system and method. Background Technology

[0002] Mobile Internet of Things (IoT) systems enable connectivity and data transmission between IoT devices via mobile communication networks. This allows IoT devices to connect to the internet anytime, anywhere, enabling real-time data transmission and processing. These systems have been widely applied in areas such as digital city construction, intelligent manufacturing, smart transportation, smart healthcare, and mobile payment.

[0003] Taking smart healthcare as an example, the Internet of Things (IoT) in smart healthcare, as a core technological infrastructure for digital healthcare transformation, has formed a new technical architecture combining multiple protocols and intelligent optimization. Current mainstream systems, through heterogeneous networks such as LoRa, Wi-Fi, or ZigBee, combined with dynamic spectrum allocation and Quality of Service (QoS) priority scheduling mechanisms, have achieved large-scale deployment in scenarios such as mobile ward rounds and remote monitoring. Regarding device interconnectivity, the ISO / IEEE 11073 series of standards has built a standardized data interaction framework, achieving seamless data integration of multimodal medical devices (such as ECG monitors and ultrasound diagnostic equipment) through layered protocol stack design, providing structured data support for clinical diagnosis and treatment. In terms of data security, the combination of lightweight encryption protocols and dynamic key negotiation mechanisms enables medical data to meet HIPAA security benchmark requirements in static transmission scenarios.

[0004] However, existing smart healthcare IoT systems face systemic challenges in dynamic healthcare scenarios:

[0005] (1) Dilemma of multimodal data collaboration: The transmission delay difference of heterogeneous protocol devices leads to time axis misalignment of multi-parameter vital sign data, such as electrocardiogram, blood oxygen, and respiratory waveforms, when they are aggregated. This time sequence distortion is particularly significant in mobile diagnosis and treatment scenarios, directly affecting the real-time performance and accuracy of clinical decision-making.

[0006] (2) Dynamic environment clock synchronization failure: Traditional clock synchronization mechanisms generate phase deviation in the scenario of device movement. When the device acceleration increases, the synchronization error grows exponentially, causing applications such as arrhythmia detection based on time-series characteristics to fail.

[0007] (3) Dynamic channel access performance bottleneck: In complex electromagnetic environments, such as MRI rooms and elevator shafts, network switching algorithms based on signal strength indication (RSSI) have serious response lag and prediction deviation. Frequent protocol switching can easily cause unstable data transmission and affect the reliability of real-time applications such as remote consultation.

[0008] Existing research often focuses on optimizing a single technical dimension. This has led to situations where some dynamic bandwidth allocation schemes, while improving spectral efficiency, have failed to address the issue of latency differences between multiple protocols. Some machine learning-based handover prediction models, while reducing latency, are insufficient in compensating for the dynamic multipath effects of motion trajectories. Furthermore, existing systems lack a correlation model between device dynamics and channel characteristics, making it difficult for existing solutions to achieve system-level optimization. Summary of the Invention

[0009] This invention addresses the problems existing in the prior art and provides a universal mobile IoT multimodal data dynamic transmission system and method. Through the synergistic effect of spatiotemporal coupling modeling of device motion state and electromagnetic environment, heuristic protocol optimization decision-making, and biomimetic clock synchronization technology, it provides highly reliable and low-latency data transmission support. It is particularly suitable for smart medical application scenarios such as emergency transport and mobile monitoring, and solves key problems such as disordered multi-protocol data transmission, delayed response to channel mutations, and failure of emergency service priority preemption in mobile medical scenarios.

[0010] The technical solution adopted in this invention is a universal mobile Internet of Things (IoT) multimodal data dynamic transmission method, which includes the following steps:

[0011] S1 Constructs a motion-channel attenuation mapping model;

[0012] S2 acquires electromagnetic frequency band status data of the environment surrounding the IoT device, and combines the motion-channel attenuation mapping model and the Attention-LSTM model to predict short-term channel quality and generate a multi-protocol score array; the array here can also be a list or other data structure.

[0013] S3 selects the optimal transmission path;

[0014] S4 calibrates the clocks of IoT devices to synchronize them, reassembles the data, and enables transmission.

[0015] Preferably, in S1, the motion-channel attenuation mapping model satisfies ,in, For channel attenuation, For motion vectors, For Doppler shift, Let α be the multipath delay spread variance, and β and γ be the corresponding weighting coefficients.

[0016] Preferably, in S1, an IMU is integrated into the IoT device to output the motion vector at time t at a preset frequency, such as 100Hz. Deploy several edge computing nodes, add differential privacy noise to their motion vectors, and calculate... and The parameters of the motion-channel attenuation mapping model are obtained through training, including the weight coefficients α, β, and γ.

[0017] Preferably, in step S2, passive sensing devices, including but not limited to SDR receivers, are deployed to acquire electromagnetic frequency band status and construct a dynamic channel matrix; PathLoss is calculated by combining the coefficients output by the motion vector and the motion-channel attenuation mapping model; based on the Attention-LSTM model, PathLoss and the dynamic channel matrix are fused into an LSTM to predict the channel quality scores of all protocols within a preset time period, and the score array S=[S1, ..., S...] is output. n ], S n The quality score corresponding to channel n.

[0018] Preferably, S3 includes the following steps:

[0019] S3.1 Define Discrete Spin Protocol Variables N is the number of IoT devices;

[0020] S3.2 Construct a weight J that is associated with the throughput, energy consumption, and latency of the i-th IoT device. i Construct the coupling coefficient J of the i-th IoT device connecting to the j-th signal source. ij Based on J i and J ij Establish the Hamiltonian E;

[0021] S3.3 Deploy FPGA-accelerated simulated annealing algorithm on edge computing nodes to iteratively reduce Hamiltonian E and find the optimal protocol combination.

[0022] Preferably, in S3.3, the annealing parameters are dynamically adjusted based on the predicted short-term channel score. During the simulated annealing solution process, a spin state is randomly transformed in each iteration and the energy difference ΔE is calculated. According to the Metropolis criterion, the new state is accepted with probability.

[0023] When the simulated annealing temperature parameter T < 1 or the maximum number of iterations is reached, the optimal protocol combination is output.

[0024] Preferably, in S4, an oscillator is integrated into the IoT device terminal to generate a synchronization pulse signal, and a pulse neural network is deployed at the gateway to dynamically adjust the clock offset Δω according to the pulse arrival time difference;

[0025] Each data packet is assigned a four-dimensional spatiotemporal coordinate. When a data packet is detected to be lost or incomplete, a nearby device is selected based on the correlation between signal strength and motion trajectory. The lost data packet is then repaired based on the motion correlation of the nearby device.

[0026] Preferably, the method further includes an emergency service priority preemption mechanism.

[0027] Preferably, the emergency service priority preemption mechanism includes the following steps:

[0028] S5.1 Perform a security check on the aforementioned emergency service;

[0029] S5.2 Agreed-upon signature for emergency business data that meets security requirements;

[0030] After the edge computing node verifies the legitimacy of the IoT device, it adjusts the weight coefficient J of the corresponding node in the Hamiltonian. i It is also added to the simulated annealing real-time queue to force the transmission of critical data to be prioritized.

[0031] A dynamic transmission system employing the aforementioned universal mobile Internet of Things (IoT) multimodal data dynamic transmission method, the system comprising:

[0032] Several terminal devices, integrating IMU and passive sensing devices, including but not limited to SDR receivers, are used to collect real-time data on device motion status and electromagnetic environment.

[0033] The edge, including distributed edge computing nodes, is used to receive information transmitted from the device and to implement dynamic protocol switching and resource scheduling.

[0034] At the gateway end, it is used for timing calibration and repair of multi-protocol data;

[0035] The data at the gateway is output after being reassembled and undergoes security checks.

[0036] This invention relates to a universal mobile Internet of Things (IoT) multimodal data dynamic transmission system and method. It constructs a motion-channel attenuation mapping model, acquires electromagnetic frequency band status data of the environment surrounding IoT devices, combines the model to predict short-term channel quality and generate a multi-protocol score array, selects the optimal transmission path, calibrates the IoT device's clock to synchronization, reassembles the data, and achieves transmission. The system integrates an IMU and passive sensing devices in the terminal device, including but not limited to SDR receivers, to collect real-time data on device motion status and electromagnetic environment. An edge device, including distributed edge computing nodes, receives information transmitted from the device and performs dynamic protocol switching and resource scheduling. A gateway performs timing calibration and repair of the multi-protocol data. The data is output after reassembly and security verification.

[0037] The beneficial effects of this invention are as follows:

[0038] (1) Breaking through the limitations of traditional single-dimensional perception, significantly improving the channel quality prediction accuracy and transmission reliability in mobile scenarios;

[0039] (2) A hardware-accelerated heuristic model is used to achieve millisecond-level protocol switching decisions, ensuring priority preemption of emergency medical data across the entire link;

[0040] (3) The use of biomimetic clock reconstruction technology ensures the spatiotemporal consistency of multimodal data and the synchronization error meets the equipment standards.

[0041] (4) It is particularly suitable for smart healthcare systems, meeting the application requirements of high reliability and low latency. Attached Figure Description

[0042] Figure 1 This is a flowchart of the method of the present invention;

[0043] Figure 2 This is a schematic diagram of the system architecture of the present invention;

[0044] Figure 3 This is a schematic diagram of the multimodal spatiotemporal sensing module structure of the present invention;

[0045] Figure 4 This is a flowchart illustrating the heuristic transmission protocol decision-making process of the present invention.

[0046] Figure 5 This is a flowchart illustrating the emergency service interruption process of this invention. Detailed Implementation

[0047] The present invention will be further described in detail below with reference to embodiments, but the scope of protection of the present invention is not limited thereto.

[0048] This invention relates to a universal method for dynamic transmission of multimodal data in mobile Internet of Things (IoT), the method comprising the following steps:

[0049] (1) Construct a motion-channel attenuation mapping model;

[0050] (2) Obtain electromagnetic frequency band status data of the environment around the IoT device, and combine the motion-channel attenuation mapping model and the Attention-LSTM model to predict short-term channel quality and generate a multi-protocol score array.

[0051] (3) Select the optimal transmission path;

[0052] (4) Calibrate the clock of the IoT device to synchronize, reassemble the data, and realize transmission.

[0053] The following section uses mobile medical devices as IoT devices and provides a detailed explanation of the steps involved.

[0054] (1) Construct a motion-channel attenuation mapping model;

[0055] The motion-channel attenuation mapping model satisfies ,in, For channel attenuation, For motion vectors, For Doppler shift, For multipath delay spread variance, α, β, and γ are the corresponding weighting coefficients;

[0056] Furthermore,

[0057]

[0058]

[0059] The device's velocity is obtained by integrating the IMU data. Let θ be the carrier frequency, θ be the angle between the direction of motion and the direction of signal incidence, c be the speed of light, and τ be the carrier frequency. k Let N be the relative time delay of the k-th multipath component. p For the effective number of multipaths;

[0060] After building the model, an IMU, such as the ICM-42670-P 6-axis IMU sensor, is integrated into the IoT device, with the sampling frequency set to 100Hz, to output the motion vector at time t. To be precise, , where a x a y a z The triaxial acceleration value, ω θ ω φ ω ψ To calculate the three-axis angular velocity values, several edge computing nodes are deployed, and differential privacy noise is added to their motion vectors. and The motion-channel attenuation model is trained using mean squared error as the loss function, and the parameters α, β, and γ are iteratively optimized using the stochastic gradient descent algorithm to minimize the error between the predicted and measured values ​​of PathLoss.

[0061] During training, edge computing nodes collaboratively optimize the motion-channel attenuation model using a federated learning framework. For example, deploying five edge computing nodes within a hospital and communicating via the gRPC protocol based on TLS 1.3 encryption, differential privacy noise M is added to the motion vectors before local training, resulting in...

[0062]

[0063] Each edge computing node is trained on local data for 100 epochs with a batch size of 64 and a learning rate of 0.001. The coefficients are adjusted through local learning.

[0064] In practical applications, motion information from mobile medical devices is collected in real time and input into a pre-trained model, which outputs the motion-channel attenuation model optimization parameters α, β, and γ.

[0065] (2) Obtain electromagnetic frequency band status data of the environment surrounding the IoT device, and combine the motion-channel attenuation mapping model and the Attention-LSTM model to predict short-term channel quality and generate a multi-protocol score array; the array here can also be a list or other data structure.

[0066] Deploy passive sensing devices, including but not limited to SDR receivers, to acquire electromagnetic frequency band status and construct a dynamic channel matrix. Calculate PathLoss by combining motion vectors with coefficients output from the motion-channel attenuation mapping model. Based on an Attention-LSTM model, fuse PathLoss and the dynamic channel matrix into an LSTM to predict the channel quality score of all protocols within a preset time period, outputting a score array S=[S1, ..., S...]. n ], S n The quality score corresponding to channel n.

[0067] Specifically, an SDR receiver is used to passively acquire the environmental electromagnetic frequency band status, sensing electromagnetic signals in the 2.4GHz and 5GHz bands of the environment at a period of 5ms. The channel energy distribution is measured (resolution 1MHz) through a spectrum analysis algorithm to generate a dynamic channel matrix. , where |h ij The passively received signal strength of the i-th device to the j-th signal source is obtained through power spectral density estimation and satisfies:

[0068]

[0069] P received,ij P represents the measured SDR received power. ref =1mW is the reference power;

[0070] ∠h ij The corresponding phase offset is calculated by the SDR digital downconversion module, and t is the measurement time.

[0071] Based on Attention-LSTM, the model predicts short-term channel scores for various protocols, typically within the next 5 seconds. The model consists of a concatenated input layer, LSTM layer, Attention layer, linear layer, and Softmax output layer. The inputs are the calculated historical PathLoss for the past 5 seconds and the channel matrix, and the output is a score vector S. The calculation formula is as follows:

[0072]

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079]

[0080] in, , , These are the forget gate, input gate, and output gate of an LSTM network. For the Sigmoid function, W ∗ and b ∗ For a trainable weight matrix, c t h t These represent the memory cells and hidden states of an LSTM network, respectively. t Here, represents the attention weights used for the weighted LSTM hidden state, Softmax is the activation function, and c is the value of attention. Attention S is the context vector, and S is the multi-protocol score array output by the prediction model;

[0081] The LSTM layer includes 32 hidden units and a Dropout layer with a dropout rate of 0.2 and a time step of 5. The model uses 5,000 pre-collected samples, including those from normal environments and MRI rooms with strong interference, for parameter optimization. Edge computing nodes make predictions every 5 seconds, and the output score is dynamically updated to select the optimal transmission path during heuristic protocol decision-making.

[0082] (3) Select the optimal transmission path;

[0083] This mainly refers to the edge computing nodes, which can heuristically select the optimal transmission path based on channel scoring and simulated annealing algorithm, and reasonably balance throughput, energy consumption and latency.

[0084] Specifically, it includes the following steps:

[0085] (3-1) Define Discrete Spin Protocol Variables N represents the number of IoT devices. The protocol selection of each device is regarded as a "spin state". Different spin states correspond to the transmission characteristics of different protocols. The overall performance of the protocol combination is quantified by the energy function.

[0086] Here, the multi-protocol transmission path optimization problem is transformed into an Ising energy minimization problem. For the i-th IoT device, the discrete spin protocol variable σ is used. i z Indicates the choice of its transmission protocol;

[0087] (3-2) Construct a weight J that is associated with the throughput, energy consumption, and latency of the i-th IoT device. i Construct the coupling coefficient J of the i-th IoT device connecting to the j-th signal source. ij Based on J i and J ij Establish the Hamiltonian E;

[0088]

[0089] in, , For the predicted channel score, E cost For energy consumption, D time For time delay, ,|h ij The smaller the value, the higher the coupling strength, which can effectively suppress high-interference links;

[0090] (3-3) Deploy FPGA-accelerated simulated annealing algorithm on edge computing nodes to iteratively reduce Hamiltonian E and find the optimal protocol combination;

[0091] With the goal of minimizing the Hamiltonian E, the spin state of a device is randomly flipped in each iteration of the simulated annealing process, the energy difference ΔE is calculated, and the new state is accepted according to the Metropolis criterion with probability.

[0092] When the simulated annealing temperature parameter T < 1 or the maximum number of iterations is reached, the optimal protocol combination is output.

[0093] Specifically, the annealing parameters are dynamically adjusted based on the predicted short-term channel score. If the predicted channel conditions are poor (S < 0.3), a high initial annealing temperature is set to expand the search range. If the predicted channel conditions are good (S ≥ 0.3), a low initial annealing temperature is set to quickly focus on the local optimum. The high and low initial annealing temperatures can be set by those skilled in the art based on their needs.

[0094]

[0095] E new E is the Hamiltonian after randomly flipping the spin state of a device. old This is the Hamiltonian before this iteration;

[0096] According to the Metropolis criterion, the new state is accepted probabilistically, and the probability of acceptance is satisfied.

[0097]

[0098] After completing one iteration, the temperature decays before proceeding to the next iteration; when the simulated annealing temperature parameter T < 1 or the maximum number of iterations is reached, the optimal protocol combination {σ1} for N IoT devices is output. z , σ2 z , ..., σ N z}

[0099] (4) Calibrate the clocks of IoT devices to synchronization, reassemble the data, and enable transmission;

[0100] Used to calibrate distributed device clocks via a spiking neural network and repair data streams lost due to channel switching;

[0101] Specifically, an oscillator, such as the SiT9367 oscillator, is integrated into the IoT device terminal to generate a synchronization pulse signal with a precision of 1μs. <Device ID, t send >, where Device ID is the device identification number, t send The device sends timestamps with an accuracy of ±1μs. A spiking neural network is deployed at the gateway to dynamically adjust the clock offset Δω based on the pulse arrival time difference according to the STDP synchronization rule.

[0102]

[0103] Among them, t receive The gateway receives message timestamps. During implementation, the learning rate η = 0.1 and the time constant τ = 10ms.

[0104] If the pulse from device A is earlier than that from device B (t A <t B If ω increases the weight of A, then A's weight ω A =ω A +△ω, the weight ω that suppresses B B =ω B -△ω, conversely, suppress the weight of A and enhance the weight of B;

[0105] Each data packet is appended with four-dimensional spatiotemporal coordinates (x, y, z, t), where x, y, and z are calculated by IMU integration, and t is a microsecond-level timestamp after synchronization. When data packet loss or incompleteness is detected, based on signal strength (e.g., RSSI > -70 dBm) and motion trajectory correlation (e.g., motion trajectory covariance > 0.8), nearby devices are selected. The lost data packets are repaired based on the motion correlation of nearby devices, satisfying the following conditions.

[0106]

[0107] Where, d i and These are reference data and data to be repaired, respectively. k p is the location index of the data to be repaired. i This is the location index for the reference data.

[0108] To better implement the method of the present invention, achieve end-to-end dynamic resource reallocation in emergency rescue scenarios, and ensure critical data transmission in emergency rescue scenarios, the method further includes:

[0109] (5) Emergency business priority preemption mechanism;

[0110] The emergency service priority preemption mechanism includes the following steps:

[0111] S5.1 A security check is performed on the aforementioned emergency services to prevent malicious access.

[0112] When an emergency call is triggered, the device sends an authentication request containing the Device ID and a time-limited random number (Nonce) to the edge computing node. The edge computing node verifies the validity of the device certificate through the PKI system, confirms the binding relationship between the hardware fingerprint and the certificate, and issues a 5-minute valid digital token Token=HS256(DeviceID∥Nonce∥Time Limit Window) signed by the HMAC-SHA256 algorithm.

[0113] S5.2 Agreed-upon signature for emergency business data that meets security requirements;

[0114] When an authenticated emergency service is sent to the gateway via an edge node device, a signature based on the Elliptic Curve Digital Signature Algorithm (ECDSA) must be attached. The gateway verifies the validity of the signature using a pre-set public key. The format of the digital signature σ is σ=ECDSA-Sign(PrivateKey,Hash(data packet∥Token)).

[0115] After the edge computing node verifies the legitimacy of the IoT device, it adjusts the weight coefficient J of the corresponding node in the Hamiltonian. i To be precise, significantly increase J i In this embodiment, the value is set to 10 times. The spin state protocol selection corresponding to this node will significantly reduce the Hamiltonian E. After adding it to the simulated annealing real-time queue, the algorithm will accept this state with a higher probability. The protocol selection of this high-weight node will be written into the real-time queue transmission channel first, and key data will be transmitted first.

[0116] To allocate independent computing units for urgent services within the FPGA, the current optimization process is interrupted and restarted with a preemption delay of less than 20μs, ensuring that the end-to-end transmission latency of high-priority data is less than 50ms. Furthermore, to prevent congestion of routine service transmissions due to urgent priority data being monopolized, the system uses a weighted fair queue to schedule routine services, ensuring their minimum bandwidth guarantee. min =0.2·B total This is to avoid disruptions to daily operations due to network congestion.

[0117] The present invention also relates to a dynamic transmission system employing the aforementioned universal mobile Internet of Things multimodal data dynamic transmission method, the system comprising:

[0118] Several terminal devices, integrating IMU and passive sensing devices, including but not limited to SDR receivers, are used to collect real-time data on device motion status and electromagnetic environment.

[0119] The edge, including distributed edge computing nodes, is used to receive information transmitted from the device and to implement dynamic protocol switching and resource scheduling.

[0120] At the gateway end, it is used for timing calibration and repair of multi-protocol data;

[0121] The data at the gateway is output after being reassembled and undergoes security checks.

[0122] In this embodiment, the device is deployed on medical terminal devices, such as electrocardiogram monitors and defibrillators, integrating a 6-axis IMU sensor and passive sensing devices, including but not limited to SDR receivers, to collect real-time data on device motion status and electromagnetic environment. The edge consists of distributed edge computing nodes, running a federated learning framework and heuristic decision-making module to achieve dynamic protocol switching and resource scheduling. The gateway deploys a pulse neural network clock synchronization module and a spatiotemporal data reassembler to complete the timing calibration and repair of multi-protocol data. The raw data collected by the device is transmitted to the gateway via LoRa / Wi-Fi protocol after being decided by the edge, and finally connected to the hospital's central system after reconstruction and security verification.

[0123] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0124] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0125] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0126] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0127] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention.

[0128] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A universally applicable method for dynamic multimodal data transmission in mobile Internet of Things, characterized in that: The method includes the following steps: S1 Construct a motion-channel attenuation mapping model; the motion-channel attenuation mapping model satisfies ,in, For channel attenuation, For motion vectors, For Doppler shift, For multipath delay spread variance, α, β, and γ are the corresponding weighting coefficients; S2 acquires electromagnetic frequency band status data of the environment surrounding the IoT device, and combines the motion-channel attenuation mapping model and the Attention-LSTM model to predict short-term channel quality and generate a multi-protocol score array. S3 selects the optimal transmission path; S4 calibrates the clocks of IoT devices to synchronize, reassembles data, and enables transmission.

2. The universal mobile Internet of Things (IoT) multimodal data dynamic transmission method according to claim 1, characterized in that: In S1, an IMU is integrated into the IoT device to output the motion vector at time t at a preset frequency. Deploy several edge computing nodes, add differential privacy noise to their motion vectors, and calculate... and The parameters of the motion-channel attenuation mapping model are obtained through training, including the weight coefficients α, β, and γ.

3. The universal mobile Internet of Things (IoT) multimodal data dynamic transmission method according to claim 1, characterized in that: In S2, passive sensing devices are deployed to acquire electromagnetic frequency band status and construct a dynamic channel matrix. PathLoss is calculated by combining the coefficients output from the motion vector and motion-channel attenuation mapping model. Based on the Attention-LSTM model, PathLoss and the dynamic channel matrix are fused into an LSTM to predict the channel quality scores of all protocols within a preset time period, outputting a score array S=[S1, ..., S...]. n ], S n The quality score corresponding to channel n.

4. The universal mobile Internet of Things (IoT) multimodal data dynamic transmission method according to claim 1, characterized in that: S3 includes the following steps: S3.1 Define Discrete Spin Protocol Variables N is the number of IoT devices; S3.2 Construct a weight J that is associated with the throughput, energy consumption, and latency of the i-th IoT device. i Construct the coupling coefficient J of the i-th IoT device connecting to the j-th signal source. ij Based on J i and J ij Establish the Hamiltonian E; S3.3 Deploys an FPGA-accelerated simulated annealing algorithm on edge computing nodes to iteratively reduce the Hamiltonian E and find the optimal protocol combination.

5. The universal mobile Internet of Things multimodal data dynamic transmission method according to claim 4, characterized in that: In S3.3, the annealing parameters are dynamically adjusted based on the predicted short-term channel score. During the simulated annealing solution process, a spin state is randomly transformed in each iteration and the energy difference ΔE is calculated. According to the Metropolis criterion, the new state is accepted with probability. When the simulated annealing temperature parameter T < 1 or the maximum number of iterations is reached, the optimal protocol combination is output.

6. The universal mobile Internet of Things (IoT) multimodal data dynamic transmission method according to claim 1, characterized in that: In S4, an oscillator is integrated into the IoT device terminal to generate a synchronization pulse signal, and a pulse neural network is deployed at the gateway to dynamically adjust the clock offset Δω based on the pulse arrival time difference. Each data packet is assigned a four-dimensional spatiotemporal coordinate. When a data packet is detected to be lost or incomplete, a nearby device is selected based on the correlation between signal strength and motion trajectory. The lost data packet is then repaired based on the motion correlation of the nearby device.

7. The universal mobile Internet of Things (IoT) multimodal data dynamic transmission method according to claim 1, characterized in that: The method also includes an emergency service priority preemption mechanism.

8. The universal mobile Internet of Things multimodal data dynamic transmission method according to claim 7, characterized in that: The emergency service priority preemption mechanism includes the following steps: S5.1 Perform a security check on the aforementioned emergency service; S5.2 Agreed-upon signature for emergency business data that meets security requirements; After the edge computing node verifies the legitimacy of the IoT device, it adjusts the weight coefficient J of the corresponding node in the Hamiltonian. i It is also added to the simulated annealing real-time queue to force the transmission of critical data to be prioritized.

9. A dynamic transmission system employing the universal mobile Internet of Things multimodal data dynamic transmission method as described in any one of claims 1 to 8, characterized in that: The system includes: Several terminal devices, integrating IMU and passive sensing devices, are used to collect data on device motion status and electromagnetic environment in real time; The edge, including distributed edge computing nodes, is used to receive information transmitted from the device and to implement dynamic protocol switching and resource scheduling. At the gateway end, it is used for timing calibration and repair of multi-protocol data; The data at the gateway is output after being reassembled and undergoes security verification.

Images