Dual-mode communication method and system based on intelligent decision

By constructing a reinforcement learning model to optimize the dual-mode communication parameters, the problem of insufficient dynamic adaptability in existing technologies is solved, achieving efficient and low-energy communication and simplifying system deployment.

CN121841553APending Publication Date: 2026-04-10SUZHOU FUTURE ELECTRICAL APP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202511832196.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-05
Publication Date
2026-04-10

AI Technical Summary

Technical Problem

Existing dual-mode communication technology lacks dynamic adaptability in channel state judgment and mode switching, resulting in poor power consumption control, frequent transmission interruptions, high system deployment costs, low reliability, and insufficient throughput.

Method used

A dual-mode communication method based on a reinforcement learning model is constructed. By acquiring historical and actual operational data from HPLC and HRF, a state space, action space, and reward function are created. Iterative training is then performed to optimize communication parameters in order to achieve adaptive mode switching and parameter adjustment.

Benefits of technology

It improves communication efficiency, reduces energy consumption and the risk of data loss, simplifies system architecture, and reduces deployment costs and response latency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121841553A_ABST
    Figure CN121841553A_ABST
Patent Text Reader

Abstract

The invention discloses a dual-mode communication method and system based on intelligent decision making. The method comprises the following steps: acquiring historical operation data and actual operation data of HPLC (High Performance Liquid Chromatography) and HRF (High Performance Framework); creating a reinforcement learning model which comprises a state space, an action space and a reward function; the state space is a combined vector of a dual-mode channel quality index, the action space comprises HPLC priority, HRF priority and dual-mode self-adaption, and the reward function is a weighted calculation formula of throughput, power consumption and packet loss rate; and performing iterative training on the reinforcement learning model based on the historical operation data to obtain an intelligent optimization model, inputting the actual operation data into the intelligent optimization model to obtain an optimal action, and adjusting the dual-mode communication parameters according to a corresponding preset parameter adjustment rule. According to the invention, communication mode selection and parameter adjustment are more suitable for actual environment requirements, the communication efficiency is improved, the energy consumption and data loss risk are effectively reduced, the system architecture is simplified, and the deployment cost and response delay are reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of communication technology, and specifically to a dual-mode communication method and system based on intelligent decision-making. Background Technology

[0002] Dual-mode communication technology, by combining the wired transmission stability of high-speed power line communication (HPLC) with the advantages of high-frequency wireless communication (HRF), has been widely applied in fields such as smart grids and the Industrial Internet of Things (IIoT). However, existing dual-mode communication relies on fixed thresholds or specific calculation methods to determine channel status and achieves mode switching by rigidly adjusting signal output. This lack of adaptability to dynamic environments leads to poor power consumption control and susceptibility to transmission interruptions. Alternatively, a centralized detection architecture, relying on external control equipment and specific communication protocols for network detection, not only increases system deployment costs and latency but also introduces single-point-of-failure risks, affecting communication reliability. Furthermore, traditional rate adjustment is mostly based on predefined discrete sequences, unable to continuously optimize according to real-time channel status, making it difficult to fully utilize channel resources and resulting in insufficient throughput and high packet loss rates. These shortcomings limit the application effectiveness of dual-mode communication technology in demanding scenarios.

[0003] Therefore, how to overcome the shortcomings of existing technologies using effective methods has become an urgent technical challenge. Summary of the Invention

[0004] The purpose of this invention is to address the above-mentioned problems by providing a dual-mode communication method and system based on intelligent decision-making.

[0005] The technical solution of this invention is: a dual-mode communication method based on intelligent decision-making, wherein the dual-mode communication is high-speed power line communication (HPLC) and high-frequency wireless communication (HRF), comprising the following steps: Historical and actual operating data of HPLC and HRF were obtained, including signal-to-noise ratio, throughput, power consumption, and packet loss rate. A reinforcement learning model is created, comprising a state space, an action space, and a reward function; wherein the state space is a combination vector of the channel quality index of HPLC and the channel quality index of HRF, the action space includes HPLC priority, HRF priority, and dual-mode adaptive, and the reward function is a weighted calculation formula of throughput, power consumption, and packet loss rate; The reinforcement learning model is iteratively trained based on the historical running data. When the preset number of iterations is reached or the fluctuation of the reward value is less than a preset threshold, the training stops and an intelligent optimization model is obtained. The actual running data is used as the input of the intelligent optimization model to obtain the optimal action. The HPLC and / or HRF communication parameters are adjusted according to the preset parameter adjustment rules corresponding to the optimal action.

[0006] As an improvement to this embodiment of the invention, the Channel Quality Index (CQI) is a multi-level index obtained by quantizing the signal-to-noise ratio (SNR); the quantization method of the Channel Quality Index (CQI) is as follows: Where SNR is the signal-to-noise ratio. For the lowest signal-to-noise ratio, For the optimal signal-to-noise ratio. This is the floor function.

[0007] As an improvement of this embodiment of the invention, the quantization level of the channel quality index is 5-10 levels; when the quantization level is K, the value range of the channel quality index is [1, K], where K is any positive integer between 5 and 10.

[0008] As an improvement to this embodiment of the invention, the weighted calculation formula of the reward function is: ; Where R is the reward value, The normalized throughput The power consumption is the normalized value. The normalized packet loss rate. , , These are the weights corresponding to throughput, power consumption, and packet loss rate, respectively.

[0009] As an improvement to this embodiment of the invention, the iterative training process of the reinforcement learning model includes: S1: Initialize model parameters, including learning rate, discount factor, and exploration rate; S2: Extract batch samples from historical running data and construct a training dataset based on the batch samples; S3: Input the channel quality index from the training dataset into the reinforcement learning model, output the action probability distribution, and select an action based on the action probability distribution according to a greedy strategy; S4. Calculate the reward function value and update the model parameters based on the reward function value; S5. Repeat S2-S4 until the preset number of iterations is reached or the fluctuation range of the reward value in several consecutive iterations is less than the preset threshold, then stop training and obtain the intelligent optimization model.

[0010] As an improvement to this embodiment of the invention, the reinforcement learning model is constructed using a deep Q-network, which includes an input layer, a hidden layer, and an output layer; the dimension of the input layer is the same as the number of quantization levels of the channel quality index, and the dimension of the output layer is the same as the number of actions in the action space.

[0011] As an improvement to this embodiment of the invention, the communication parameters include transmission rate, modulation scheme, transmission power, and channel switching threshold.

[0012] As an improvement to this embodiment of the invention, the preset parameter adjustment rule is as follows: If the optimal action is HPLC priority, the transmission rate is the maximum stable rate of HPLC, a high-order modulation method is adopted, the transmission power of HPLC is adjusted to the value corresponding to the lowest transmission loss, and the channel switching threshold is the first preset threshold. If the optimal action is HRF priority, the transmission rate is dynamically adjusted according to the HRF channel quality index, a medium- or low-order modulation method is adopted, the HRF transmit power is controlled in the low power consumption range, and the channel switching threshold is a second preset threshold. If the optimal action is dual-mode adaptive, the transmission rate is allocated according to the resource allocation mode, the modulation method is determined according to the channel quality index, the transmit power is modulated according to the dual-channel requirements, and the channel switching threshold is the third preset threshold and the fourth preset threshold.

[0013] As an improvement to this embodiment of the invention, the execution logic of the dual-mode adaptive action is as follows: When the channel quality index of either HPLC or HRF is higher than the third preset threshold, the channel is preferentially used for data transmission. When the channel quality indices of both HPLC and HRF are between the third and fourth preset thresholds, time-division multiplexing is used to allocate transmission resources. When the channel quality index of both HPLC and HRF is lower than the fourth preset threshold, the same data is transmitted synchronously through dual channels; wherein the third preset threshold is greater than the fourth preset threshold.

[0014] To achieve one of the above-mentioned objectives, one embodiment of the present invention provides a dual-mode communication system based on intelligent decision-making. The dual-mode communication system is used to implement the dual-mode communication method as described in any of the above claims. The dual-mode communication system includes a perception layer, a decision layer, and an execution layer. The sensing layer is used to acquire historical and actual operating data of HPLC and HRF, including signal-to-noise ratio, throughput, power consumption, and packet loss rate. The decision layer is used to create a reinforcement learning model, which includes a state space, an action space, and a reward function. The state space is a combination vector of the channel quality index of HPLC and the channel quality index of HRF. The action space includes HPLC-first, HRF-first, and dual-mode adaptive approaches. The reward function is a weighted calculation formula of throughput, power consumption, and packet loss rate. The reinforcement learning model is iteratively trained based on the historical operating data. Training stops when a preset number of iterations is reached or the reward value fluctuation is less than a preset threshold, and an intelligent optimization model is obtained. The actual operating data is used as input to the intelligent optimization model to obtain the optimal action. The execution layer is used to adjust the HPLC and / or HRF communication parameters according to the preset parameter adjustment rules corresponding to the optimal action.

[0015] The dual-mode communication method and system based on intelligent decision-making provided in this invention have the following advantages: This invention constructs a reinforcement learning model by integrating multi-dimensional operational data from HPLC and HRF, defining the state space with a channel quality index combination vector. This allows for the capture of dynamic channel changes, making communication mode selection and parameter adjustment more aligned with actual environmental requirements. The reward function achieves multi-objective optimization balance by rationally allocating weights to throughput, power consumption, and packet loss rate, effectively reducing energy consumption and data loss risks while improving communication efficiency. The reinforcement learning model possesses autonomous learning and adaptive adjustment capabilities, enabling local decision-making without relying on external control equipment, simplifying the system architecture and reducing deployment costs and response latency. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the dual-mode communication method based on intelligent decision-making described in this invention. Figure 2 This is a schematic diagram of the dual-mode communication system based on intelligent decision-making described in this invention. Detailed Implementation

[0017] The present invention will now be described in detail with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the scope of protection of the present invention.

[0018] The scope of the embodiments described herein includes the entire scope of the claims and all available equivalents thereof. Throughout this document, the terms “first,” “second,” etc., are used only to distinguish one element from another without requiring or implying any actual relationship or order between the elements. Indeed, a first element can also be referred to as a second element, and vice versa. Furthermore, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a structure, apparatus, or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a structure, apparatus, or device. Without further limitations, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the structure, apparatus, or device that includes said element. The various embodiments described herein are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably.

[0019] The terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer" used in this document to indicate orientation or positional relationships are based on the orientation or positional relationships shown in the accompanying drawings and are used only for the convenience of describing this document and simplifying the description. They do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention. In the description herein, unless otherwise specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two elements, or direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0020] This invention provides a dual-mode communication method based on intelligent decision-making, wherein the dual-mode communication comprises high-speed power line communication (HPLC) and high-frequency wireless communication (HRF), such as... Figure 1 As shown, it includes the following steps: Step 101: Obtain historical and actual operating data of HPLC and HRF, including signal-to-noise ratio, throughput, power consumption, and packet loss rate. In practice, the historical operational data is a long-term accumulation of actual operational data, with an accumulation period of no less than 30 days. It needs to cover operational scenarios with different time periods, different interference environments, and different data volumes, and is stored in a structured data format. Each data record includes a timestamp, communication module type, signal-to-noise ratio, throughput, transmit power consumption, receive power consumption, standby power consumption, packet loss rate, and operational scenario identifier. The operational scenario identifier can be manually configured or automatically identified through the system input interface. Both the collected historical operational data and actual operational data need to undergo preprocessing. Preprocessing includes at least outlier removal, data completion, and data standardization. Outlier removal can use the 3σ criterion or box plot method to remove obvious abnormal data caused by sensor failure. Data completion addresses minor data loss due to communication interruptions. If the data loss is continuous for no more than three acquisition cycles, linear interpolation is used for completion. If the data loss is continuous for more than three acquisition cycles, it is marked as a data failure segment and will not participate in subsequent decision calculations. Data standardization converts all parameters to the range of 0 to 1, obtained by subtracting the minimum value from the original value and dividing by the difference between the maximum value and the minimum value. The minimum and maximum values ​​are the extreme values ​​from the historical data.

[0021] Step 102: Create a reinforcement learning model, which includes a state space, an action space, and a reward function. The state space is a combination vector of the channel quality index of HPLC and the channel quality index of HRF. The action space includes HPLC-first, HRF-first, and dual-mode adaptive approaches. The reward function is a weighted calculation formula of throughput, power consumption, and packet loss rate. Iteratively train the reinforcement learning model based on the historical operating data. When a preset number of iterations is reached or the fluctuation of the reward value is less than a preset threshold, training stops and an intelligent optimization model is obtained. The actual operating data is used as the input to the intelligent optimization model to obtain the optimal action. Here, the reinforcement learning model obtains the current state from the state space; then, it selects and executes an action from the action space, calculates the R value corresponding to the action using a reward function, judges the quality of the action, and finally adjusts the model parameters based on the R value so that the model selects the high-reward action more likely in the same state. After multiple rounds of iterative training, the model can form a mapping relationship between the state and the optimal action, ultimately obtaining an intelligent optimization model that conforms to the real-time channel environment.

[0022] Step 103: Adjust the HPLC and / or HRF communication parameters according to the preset parameter adjustment rules corresponding to the optimal action. The communication parameters include transmission rate, modulation scheme, transmit power, and channel switching threshold. The preset parameter adjustment rules are as follows: If the optimal action is HPLC priority, the transmission rate is the maximum stable rate of HPLC, a high-order modulation method is adopted, the transmission power of HPLC is adjusted to the value corresponding to the lowest transmission loss, and the channel switching threshold is the first preset threshold. In practice, the maximum stable rate is determined using historical operating data accumulated in step 101. The maximum throughput value of the HPLC under communication conditions with a packet loss rate not exceeding 1% and a signal-to-noise ratio not lower than 10 dB can be selected from historical data and used as the current HPLC transmission rate. The higher-order modulation method can specifically be 256 orthogonal amplitude modulation or 1024 orthogonal amplitude modulation under orthogonal frequency division multiplexing modulation. The transmit power is adjusted to the value corresponding to the lowest transmission loss. This can be calculated using the transmission loss data of the HPLC at different transmit powers collected in step 101. Transmission loss equals transmit power minus receive power. A transmit power-transmission loss curve is plotted, and the transmit power corresponding to the minimum transmission loss on the curve is the target value. The first preset threshold is the lower limit of the HPLC channel quality index, determined by the critical channel quality index where the HPLC transmission success rate is below 90% in historical data.

[0023] If the optimal action is HRF priority, the transmission rate is dynamically adjusted according to the HRF channel quality index, a medium- or low-order modulation method is adopted, the HRF transmit power is controlled in the low power consumption range, and the channel switching threshold is a second preset threshold. In practice, the HRF transmission rate is dynamically adjusted according to the HRF channel quality index. The adjustment gradient can be calibrated using historical data. When the HRF channel quality index is not lower than 6, the transmission rate is configured to the maximum HRF rate, such as 10 Mbps. When the HRF channel quality index is 3 or higher but lower than 6, the transmission rate is reduced to a medium rate, such as 5 Mbps. When the HRF channel quality index is lower than 3, the transmission rate is reduced to a basic rate, such as 1 Mbps. The intermediate-order modulation can be 16-quadrature amplitude modulation, and the low-order modulation can be quadrature phase shift keying. The low-power range is 30% to 50% of the HRF's rated transmit power. The second preset threshold is the lower limit of the HRF channel quality index, determined by the critical channel quality index where the HRF transmission success rate is lower than 85% in historical data. For example, if historical data shows that the HRF transmission success rate drops to 80% when the HRF channel quality index does not exceed 2, then the second preset threshold is set to 2.

[0024] If the optimal action is dual-mode adaptive, the transmission rate is allocated according to the resource allocation mode, the modulation method is determined according to the channel quality index, the transmit power is modulated separately according to the dual-channel requirements, and the channel switching threshold is a third preset threshold and a fourth preset threshold. Specifically, when the channel quality index of either HPLC or HRF is higher than the third preset threshold, that channel is preferentially used for data transmission; When the channel quality indices of both HPLC and HRF are between the third and fourth preset thresholds, time-division multiplexing is used to allocate transmission resources. When the channel quality index of both HPLC and HRF is lower than the fourth preset threshold, the same data is transmitted synchronously through dual channels; wherein the third preset threshold is greater than the fourth preset threshold.

[0025] In practice, the transmission rate allocation according to the resource allocation mode can be determined based on the ratio of the real-time HPLC channel quality index to the HRF channel quality index, which corresponds to the rate of dual-link data allocation. The third preset threshold is greater than the fourth preset threshold. The third preset threshold is the critical channel quality index for dual-link priority transmission, and the fourth preset threshold is the critical channel quality index for dual-link synchronous transmission. Both are determined by the channel quality index range with the best transmission efficiency in historical data.

[0026] In this embodiment, the Channel Quality Index (CQI) is a multi-level index obtained by quantizing the signal-to-noise ratio (SNR); the quantization method of the Channel Quality Index (CQI) is as follows: Where SNR is the signal-to-noise ratio. For the lowest signal-to-noise ratio, For the optimal signal-to-noise ratio. This is the floor function.

[0027] In this embodiment, the quantization level of the channel quality index is 5-10; when the quantization level is K, the value range of the channel quality index is [1, K], where K is any positive integer between 5 and 10.

[0028] In this embodiment, the weighted calculation formula for the reward function is: ; Where R is the reward value, The normalized throughput The power consumption is the normalized value. The normalized packet loss rate. , , These are the weights corresponding to throughput, power consumption, and packet loss rate, respectively.

[0029] In this embodiment, the iterative training process of the reinforcement learning model includes: S1: Initialize model parameters, including learning rate, discount factor, and exploration rate; S2: Extract batch samples from historical running data and construct a training dataset based on the batch samples; S3: Input the channel quality index from the training dataset into the reinforcement learning model, output the action probability distribution, and select an action based on the action probability distribution according to a greedy strategy; S4. Calculate the reward function value and update the model parameters based on the reward function value; S5. Repeat S2-S4 until the preset number of iterations is reached or the fluctuation range of the reward value in several consecutive iterations is less than the preset threshold, then stop training and obtain the intelligent optimization model.

[0030] In this embodiment, the reinforcement learning model is constructed using a deep Q-network, which includes an input layer, a hidden layer, and an output layer. The dimension of the input layer is the same as the number of quantization levels of the channel quality index, and the dimension of the output layer is the same as the number of actions in the action space.

[0031] Here, the deep Q-network fits the state-action value function through an input layer, hidden layer, and output layer to achieve the mapping from communication state to optimal action. The input layer is used to receive the quantized information of the reinforcement learning model's state space. Combining the channel quality index value range of [0,1] in step 101, it is divided into N equally spaced quantization levels. The total dimension of the input layer is 2×N. Each neuron corresponds to a certain quantization level of the channel quality index of a single link. The received value is a one-hot encoded vector, thereby converting the discrete channel quality index levels into a numerical signal that the network can recognize. The hidden layer is used to extract key features in the input state and achieve complex mapping through a nonlinear activation function. It can have 2-3 layers. Preferably, in this embodiment, there are 2 hidden layers. The number of neurons in the first hidden layer is 2-3 times the dimension of the input layer, and the second layer is half the dimension of the first layer. Both use the ReLU activation function, and the weight initialization uses Xavier initialization to accelerate training convergence. The output layer is used to output the value estimate Q value of each action. Each neuron corresponds to the Q value of an action. The network selects the optimal action by comparing the size of the Q values.

[0032] This invention also provides a dual-mode communication system based on intelligent decision-making, wherein the dual-mode communication system is used to implement the dual-mode communication method as described in any of the preceding claims, such as... Figure 2 As shown, the dual-mode communication system includes a perception layer 201, a decision layer 202, and an execution layer 203; The sensing layer 201 is used to acquire historical and actual operating data of HPLC and HRF, including signal-to-noise ratio, throughput, power consumption and packet loss rate. The decision layer 202 is used to create a reinforcement learning model, which includes a state space, an action space, and a reward function. The state space is a combination vector of the channel quality index of HPLC and the channel quality index of HRF. The action space includes HPLC-first, HRF-first, and dual-mode adaptive approaches. The reward function is a weighted calculation formula of throughput, power consumption, and packet loss rate. The reinforcement learning model is iteratively trained based on the historical operating data. Training stops when a preset number of iterations is reached or the reward value fluctuation is less than a preset threshold, and an intelligent optimization model is obtained. The actual operating data is used as the input to the intelligent optimization model to obtain the optimal action. The execution layer 203 is used to adjust the HPLC and / or HRF communication parameters according to the preset parameter adjustment rules corresponding to the optimal action.

[0033] This invention can be an apparatus, method, and / or computer program product. A computer program product may include a readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of the invention.

[0034] Storage media can be tangible devices that hold and store instructions for use by instruction execution devices. Storage media can include, but are not limited to, electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination thereof.

[0035] It should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0036] The detailed descriptions listed above are merely specific descriptions of feasible embodiments of the present invention, and are not intended to limit the scope of protection of the present invention. All equivalent embodiments or modifications made without departing from the spirit of the present invention should be included within the scope of protection of the present invention.

Claims

1. A dual-mode communication method based on intelligent decision-making, wherein the dual-mode communication is high-speed power line communication (HPLC) and high-frequency wireless communication (HRF), characterized in that, Includes the following steps: Historical and actual operating data of HPLC and HRF were obtained, including signal-to-noise ratio, throughput, power consumption, and packet loss rate. A reinforcement learning model is created, comprising a state space, an action space, and a reward function; wherein the state space is a combination vector of the channel quality index of HPLC and the channel quality index of HRF, the action space includes HPLC priority, HRF priority, and dual-mode adaptive, and the reward function is a weighted calculation formula of throughput, power consumption, and packet loss rate; The reinforcement learning model is iteratively trained based on the historical running data. When the preset number of iterations is reached or the fluctuation of the reward value is less than a preset threshold, the training stops and an intelligent optimization model is obtained. The actual running data is used as the input of the intelligent optimization model to obtain the optimal action. The HPLC and / or HRF communication parameters are adjusted according to the preset parameter adjustment rules corresponding to the optimal action.

2. The dual-mode communication method based on intelligent decision-making according to claim 1, characterized in that, The Channel Quality Index (CQI) is a multi-level index obtained by quantizing the signal-to-noise ratio (SNR); the quantization method of the Channel Quality Index (CQI) is as follows: Where SNR is the signal-to-noise ratio. For the lowest signal-to-noise ratio, For the optimal signal-to-noise ratio. This is the floor function.

3. The dual-mode communication method based on intelligent decision-making according to claim 1, characterized in that, The quantization level of the channel quality index is 5-10; when the quantization level is K, the value range of the channel quality index is [1, K], where K is any positive integer between 5 and 10.

4. The dual-mode communication method based on intelligent decision-making according to claim 1, characterized in that, The weighted calculation formula for the reward function is: ; Where R is the reward value, The normalized throughput The power consumption is the normalized value. The normalized packet loss rate. , , These are the weights corresponding to throughput, power consumption, and packet loss rate, respectively.

5. The dual-mode communication method based on intelligent decision-making according to claim 1, characterized in that, The iterative training process of the reinforcement learning model includes: S1: Initialize model parameters, including learning rate, discount factor, and exploration rate; S2: Extract batch samples from historical running data and construct a training dataset based on the batch samples; S3: Input the channel quality index from the training dataset into the reinforcement learning model, output the action probability distribution, and select an action based on the action probability distribution according to a greedy strategy; S4. Calculate the reward function value and update the model parameters based on the reward function value; S5. Repeat S2-S4 until the preset number of iterations is reached or the fluctuation range of the reward value in several consecutive iterations is less than the preset threshold, then stop training and obtain the intelligent optimization model.

6. The dual-mode communication method based on intelligent decision-making according to claim 1, characterized in that, The reinforcement learning model is constructed using a deep Q-network, which includes an input layer, a hidden layer, and an output layer. The dimension of the input layer is the same as the number of quantization levels of the channel quality index, and the dimension of the output layer is the same as the number of actions in the action space.

7. The dual-mode communication method based on intelligent decision-making according to claim 1, characterized in that, The communication parameters include transmission rate, modulation scheme, transmission power, and channel switching threshold.

8. The dual-mode communication method based on intelligent decision-making according to claim 7, characterized in that, The preset parameter adjustment rules are as follows: If the optimal action is HPLC priority, the transmission rate is the maximum stable rate of HPLC, a high-order modulation method is adopted, the transmission power of HPLC is adjusted to the value corresponding to the lowest transmission loss, and the channel switching threshold is the first preset threshold. If the optimal action is HRF priority, the transmission rate is dynamically adjusted according to the HRF channel quality index, a medium- or low-order modulation method is adopted, the HRF transmit power is controlled in the low power consumption range, and the channel switching threshold is a second preset threshold. If the optimal action is dual-mode adaptive, the transmission rate is allocated according to the resource allocation mode, the modulation method is determined according to the channel quality index, the transmit power is modulated according to the dual-channel requirements, and the channel switching threshold is the third preset threshold and the fourth preset threshold.

9. The dual-mode communication method based on intelligent decision-making according to claim 8, characterized in that, The execution logic of the dual-mode adaptive action is as follows: When the channel quality index of either HPLC or HRF is higher than the third preset threshold, the channel is preferentially used for data transmission. When the channel quality indices of both HPLC and HRF are between the third and fourth preset thresholds, time-division multiplexing is used to allocate transmission resources. When the channel quality index of both HPLC and HRF is lower than the fourth preset threshold, the same data is transmitted synchronously through dual channels; wherein the third preset threshold is greater than the fourth preset threshold.

10. A dual-mode communication system based on intelligent decision-making, the dual-mode communication system being used to implement the dual-mode communication method as described in claims 1-9, characterized in that, It includes the perception layer, decision-making layer, and execution layer; The sensing layer is used to acquire historical and actual operating data of HPLC and HRF, including signal-to-noise ratio, throughput, power consumption, and packet loss rate. The decision layer is used to create a reinforcement learning model, which includes a state space, an action space, and a reward function. The state space is a combination vector of the channel quality index of HPLC and the channel quality index of HRF. The action space includes HPLC-first, HRF-first, and dual-mode adaptive approaches. The reward function is a weighted calculation formula of throughput, power consumption, and packet loss rate. The reinforcement learning model is iteratively trained based on the historical operating data. Training stops when a preset number of iterations is reached or the reward value fluctuation is less than a preset threshold, and an intelligent optimization model is obtained. The actual operating data is used as input to the intelligent optimization model to obtain the optimal action. The execution layer is used to adjust the HPLC and / or HRF communication parameters according to the preset parameter adjustment rules corresponding to the optimal action.