Communication channel intelligent selection method and system

Abstract

The invention discloses a communication channel intelligent selection method and system, and relates to the technical field of wireless communication, and the method comprises the steps: collecting background information of communication conditions, collecting historical data of the background information of the communication conditions, and packaging the historical data into a historical database; calculating communication channel efficiency corresponding to the background information of the current communication condition; according to the method, the reason of channel selection is perfected, the channels are selected from multiple angles, the nonlinear regression prediction model is improved, and a dynamic adaptive coding method of segmented adaptive punching of grouping sorting is added in the analysis of the nonlinear regression prediction model. Therefore, the machine learning model is adapted to channel selection of a communication channel, the communication channel efficiency of communication transmission is maximized on the premise of ensuring the overall transmission quality, and a coding scheme is dynamically adjusted according to the real-time channel condition so as to adapt to different channel conditions and transmission requirements.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless communications, and in particular to a communication channel intelligent selection method and system. Background Art

[0002] The intelligent selection method of communication channels refers to the dynamic selection of the best communication channels or spectrum resources in wireless communication systems according to different channel conditions, network requirements and service quality requirements to improve system performance, optimize resource utilization and reduce interference. This method relies on factors such as channel state information, channel quality, network topology, user needs, etc., combined with intelligent algorithms to make intelligent decisions. Wireless communication systems face many challenges, such as limited spectrum resources, fluctuating channel quality, interference management, system complexity and high latency. With the development of mobile Internet, Internet of Things, 5G and future 6G technologies, communication needs are growing. How to efficiently and intelligently select appropriate communication channels has become one of the key factors to improve system performance and user experience.

[0003] At present, the invention patent with application number CN201080014559.X discloses a method for selecting a working channel in a wireless communication network, which relates to a method for selecting a working channel with frequency parameters for a network that transmits data through a shared medium. The network is configured to communicate within a frequency range, each channel defines a predetermined frequency parameter, and the frequency range is scanned regularly to determine the interference frequency within the frequency range generated by the interference network operation in the shared medium, and based on the determined interference frequency, the channel fk is allocated to each position SPi in the sequence FS, thereby limiting the use of the interference frequency; however, the reason why the prior art cannot achieve perfect channel selection is that the channel is selected from multiple angles, the nonlinear regression prediction model cannot be improved, and the dynamic adaptive coding method of segmented adaptive puncturing with group sorting is added to the machine learning model analysis, so that the machine learning model cannot be adapted to the channel selection of the communication channel, and the communication channel efficiency of the communication transmission is maximized under the premise of ensuring the overall transmission quality. It is impossible to dynamically adjust the coding scheme according to the real-time channel situation by using adaptive coding technology, thereby failing to adapt to different channel conditions and transmission requirements, and failing to improve the system transmission rate and increase the throughput while ensuring the error performance. Summary of the invention

[0004] The technical problem solved by the present invention is: the reason why the existing technology cannot perfect channel selection, selects channels from multiple angles, cannot improve the nonlinear regression prediction model, and adds a dynamic adaptive coding method of segmented adaptive perforation with group sorting in the machine learning model analysis, cannot make the machine learning model adapt to the channel selection of the communication channel, and maximize the communication channel efficiency of the communication transmission under the premise of ensuring the overall transmission quality. It is impossible to dynamically adjust the coding scheme according to the real-time channel conditions by using adaptive coding technology, thereby failing to adapt to different channel conditions and transmission requirements, and failing to improve the system transmission rate and increase the throughput while ensuring the error performance.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a communication channel intelligent selection method, comprising the following steps:

[0006] Step S1: Collecting background information of communication conditions, and collecting historical data of the background information of communication conditions;

[0007] Step S2: Encapsulating the historical data into a historical database, and searching the historical database for corresponding historical data according to the background information;

[0008] Step S3: Calculate the communication channel efficiency corresponding to the background information of the current communication situation;

[0009] Step S4: Allocate the communication channel using the improved machine learning model.

[0010] Preferably, the step S1 comprises:

[0011] The background information of the communication situation includes channel quality, interference situation, spectrum utilization, network load, time-varying characteristics, geographic space changes, user needs and behaviors, communication protocols and channel utilization requirements;

[0012] The channel quality includes signal-to-noise ratio, channel gain, signal transmission distance, channel delay, signal transmission bandwidth, signal transmission rate, bit error rate and received signal strength;

[0013] The interference conditions include co-channel interference, adjacent channel interference and external interference;

[0014] The spectrum utilization includes idle spectrum detection and channel spectrum utilization rate;

[0015] The network load includes the current channel load and the network topology connection status, and the network topology connection status is obtained from the routing forwarding table of the nearest neighbor router;

[0016] The time-varying characteristics include the regularity of the channel changing with time;

[0017] The geographic space changes include the shift in areas where communication equipment signals are strong or weak;

[0018] The user requirements and behaviors include data rate requirements, user mobility requirements for signal quality, real-time requirements, and reliability requirements;

[0019] The communication protocols include PPP protocol, TCP protocol, IP protocol, UDP protocol, HTTP protocol, FTP protocol, SMTP protocol, DNS protocol and wireless LAN protocol;

[0020] The channel utilization requirements include quality of service QoS requirements and priority mechanisms. The quality of service QoS requirements include different types of communication requirements, including voice, video and text data transmission. The selected channel is adjusted according to the QoS requirements.

[0021] Preferably, the historical data searches the corresponding background information through big data to obtain the historical channel selection information of each background information, and the historical channel selection information includes selecting the ath channel quality channel, the bth anti-interference method, the cth spectrum utilization channel, the dth network congestion relief method, the time-varying characteristic law corresponding to the eth channel quality selection channel, the fth spatial change selection channel, the gth user demand selection channel, the hth communication protocol, and the ith channel utilization requirement corresponding channel according to the background information, and the a, b, c, d, e, f, g, h, and i are natural numbers respectively;

[0022] The anti-interference methods include frequency hopping, power control, spectrum management and dynamic spectrum access, channel coding and error correction, frequency hopping and code division multiple access technology, time division multiple access, spatial diversity and antenna technology beamforming, anti-interference filtering, modulation and demodulation technology, software defined radio, anti-interference waveforms, spread spectrum technology, noise suppression technology and multipath propagation;

[0023] The network congestion relief method includes flow control, congestion control, load balancing, traffic shaping, data compression, multi-path transmission, content distribution network CDN and network topology optimization;

[0024] The spatially varying channel selection includes selecting a channel with a preset weak channel quality interval threshold when the communication device has a strong signal in the current area;

[0025] When the communication device has a weak signal in the current area, a channel with a preset strong channel quality interval threshold is selected;

[0026] The user demand channel selection includes classifying the historical channel quality of user demand and behavior demand of big data search by using k-means clustering algorithm, and obtaining a first user demand channel, a second user demand channel, ... and a Gth user demand channel, where G is the total number of user demand channel classifications;

[0027] The channel utilization requirement corresponding channel includes using a k-means clustering algorithm to classify the historical channel requirements for voice, video and text data transmission searched by big data, and obtaining a first channel utilization requirement corresponding channel, a second channel utilization requirement corresponding channel, ... and an Ith channel utilization requirement corresponding channel, where I is the total number of channel utilization requirement corresponding channel categories.

[0028] Preferably, step S2 comprises:

[0029] All historical channel selection information corresponding to each background information is encapsulated into a historical database, the historical database includes the background information and all historical channel selection information corresponding to the background information, the historical database is stored in the memory in the form of a historical channel selection B-tree, and the historical channel selection B-tree is used as an update list to update the stored channel selection.

[0030] Preferably, the step S3 comprises:

[0031] Detect the background information of the current communication situation, retrieve and obtain the ath channel quality channel, bth anti-interference method, cth spectrum utilization channel, dth network congestion relief method, time-varying characteristic law corresponding to the eth channel quality selection channel, fth spatial variation selection channel, gth user demand selection channel, hth communication protocol and ith channel utilization requirement corresponding channel required by the current communication situation from the historical channel selection B tree, and calculate the channel power, channel efficiency, user fairness, overall network throughput and total delay of the communication network under the communication situation, and its mathematical expression is:

[0032]

[0033] C = B log2 (1 + SNR);

[0034]

[0035] T network =C·η ch ;

[0036] D total =D trans +D prop +D queue +D process ;

[0037] D trans =LR;

[0038] D prop =dv;

[0039] Among them, P ch is the channel power, Es is the symbol energy, T S is the symbol duration, η ch is the channel efficiency, R data is the effective data transmission rate, C is the channel capacity, B is the bandwidth, SNR is the signal-to-noise ratio, F Jain is Jain's fairness index, which is used to measure user fairness, T j is the channel throughput of the jth user, N is the number of users, the value range of the Jain's fairness index is [0,1]. In the range of [0,1], the larger the value of the Jain's fairness index is, the higher the fairness is. network is the network throughput, D total is the total delay of the communication network, D trans is the transmission delay, L is the size of the data packet, R is the link bandwidth, and D prop is the propagation delay, d is the distance the signal propagates, and v is the signal propagation speed. Signal propagation includes light propagation and electromagnetic wave propagation. D queue is the queuing delay, D process To deal with delays;

[0040] The channel power, channel efficiency, user fairness, overall network throughput and total delay of the communication network in the communication situation are calculated as the communication channel efficiency through weighted average.

[0041] Preferably, step S4 comprises:

[0042] The calculated communication channel efficiency under the current communication situation is used as the ideal target output, and the background information corresponding to the current communication situation is input into the embedding layer for feature extraction to obtain channel features and user features respectively. The channel features are obtained by extracting channel quality, interference, spectrum utilization, network load, communication protocol and channel utilization requirements through the embedding layer, and the user features are obtained by extracting time-varying characteristics, geographic space changes and user needs and behaviors through the embedding layer;

[0043] In the process of real-time analysis of the machine learning model of communication channel efficiency, the channel coding rate is adjusted and changed through the dynamic adaptive coding method of segmented adaptive perforation with group sorting, the machine learning model is optimized, and the optimized machine learning model is trained to predict the optimal communication channel selection.

[0044] Preferably, optimizing the machine learning model includes:

[0045] Divide the channel into multiple sub-areas, wherein the sub-areas include the ath channel quality channel, the bth anti-interference method, the cth spectrum utilization channel, the dth network congestion relief method, the eth channel quality selection channel corresponding to the time-varying characteristic law, the fth spatial variation selection channel, the gth user demand selection channel, the hth communication protocol and the ith channel utilization requirement corresponding channel stored in the historical channel selection B-tree, dynamically select the best coding and modulation mode for each sub-area, the channels in each group are different in coding and modulation strategies, and use puncturing technology for each channel segment to optimize the transmission of data packets. The puncturing technology adjusts the transmission rate according to the feedback of the current channel, and the feedback is the communication channel efficiency calculated;

[0046] The dynamic selection includes using the dynamic adjustment modulation mode 64-QAM. The dynamic adjustment logic includes:

[0047] When the channel quality of the channel is higher than a preset high quality threshold, high order modulation is used;

[0048] When the channel quality of the channel is lower than a preset high quality threshold, low order modulation is used;

[0049] The channel characteristic signal and user characteristic signal of the input channel are encoded, modulated and resource mapped through the dynamically adjusted modulation mode 64-QAM, and the dynamically adjusted modulated signal is sent to the receiving end. After the receiving end demodulates and decodes the data, the current signal-to-noise ratio is obtained through channel estimation. The receiving end feeds back the demodulated and decoded data and the signal-to-noise ratio to the transmitting end. The receiving end and the transmitting end calculate the current channel condition through an iterative segmented adaptive puncturing method according to the feedback and switch to the corresponding modulation and coding scheme. The modulation and coding scheme is used in the next transmission, and matrices of different code rates are used as the main code for puncturing.

[0050] Preferably, the channel characteristics and user characteristics are used as input, the communication channel efficiency is used as the target output, and the random forest regression model is trained to predict the optimal channel selection. The training process includes:

[0051] The input channel characteristics and user characteristics are standardized, the historical channel selection B-tree is randomly divided into 75% training set and 25% validation set, the training set is used to train the random forest regression model, the mean square error is used to measure the prediction error of the random forest regression model using the validation set, the random forest regression model is continuously trained and adjusted until the mean square error of the random forest regression model is lower than a preset error threshold, and a channel selection model for simultaneous channel prediction and coding adjustment is obtained.

[0052] Preferably, the background information of the current channel is collected and input into a trained channel selection model to obtain a predicted optimal channel selection, and the coding rate and modulation mode are dynamically adjusted through an iterative segmented adaptive puncturing method according to the predicted optimal channel selection.

[0053] Preferably, a communication channel intelligent selection system includes a collection module, a database module, a calculation module and a channel allocation module:

[0054] The acquisition module is used to collect background information of communication conditions and collect historical data of background information of communication conditions;

[0055] The database module is used to encapsulate the historical data into a historical database, and search the historical database for corresponding historical data according to the background information;

[0056] The calculation module is used to calculate the communication channel efficiency corresponding to the background information of the current communication situation;

[0057] The channel allocation module is used to allocate communication channels using an improved machine learning model.

[0058] Beneficial effects of the present invention: The present invention improves the reasons for channel selection, selects channels from multiple angles, improves the nonlinear regression prediction model, and adds a dynamic adaptive coding method of segmented adaptive perforation with group sorting to the nonlinear regression prediction model analysis, so that the machine learning model is adapted to the channel selection of the communication channel, and maximizes the communication channel efficiency of the communication transmission under the premise of ensuring the overall transmission quality. By using adaptive coding technology, the coding scheme is dynamically adjusted according to the real-time channel conditions, so as to adapt to different channel conditions and transmission requirements, and the transmission rate of the system is improved and the throughput is increased while ensuring the error performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A basic flow chart of a communication channel intelligent selection method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0060] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0061] Reference Figure 1 , as an embodiment of the present invention, provides a communication channel intelligent selection method, comprising the following steps:

[0062] Step S1: Collecting background information of communication conditions, and collecting historical data of the background information of communication conditions;

[0063] Step S2: Encapsulating the historical data into a historical database, and searching the historical database for corresponding historical data according to the background information;

[0064] Step S3: Calculate the communication channel efficiency corresponding to the background information of the current communication situation;

[0065] Step S4: Allocate the communication channel using the improved machine learning model.

[0066] The present invention improves the reasons for channel selection, selects channels from multiple angles, improves the nonlinear regression prediction model, makes the machine learning model adaptable to the channel selection of the communication channel, and maximizes the communication channel efficiency of the communication transmission while ensuring the overall transmission quality.

[0067] The step S1 comprises:

[0068] The background information of the communication situation includes channel quality, interference situation, spectrum utilization, network load, time-varying characteristics, geographic space changes, user needs and behaviors, communication protocols and channel utilization requirements;

[0069] The channel quality includes signal-to-noise ratio, channel gain, signal transmission distance, channel delay, signal transmission bandwidth, signal transmission rate, bit error rate and received signal strength;

[0070] The interference conditions include co-channel interference, adjacent channel interference and external interference;

[0071] The spectrum utilization includes idle spectrum detection and channel spectrum utilization rate;

[0072] The network load includes the current channel load and the network topology connection status, and the network topology connection status is obtained from the routing forwarding table of the nearest neighbor router;

[0073] The time-varying characteristics include the regularity of the channel changing with time;

[0074] The geographic space changes include the shift in areas where communication equipment signals are strong or weak;

[0075] The user requirements and behaviors include data rate requirements, user mobility requirements for signal quality, real-time requirements, and reliability requirements;

[0076] The communication protocols include PPP protocol, TCP protocol, IP protocol, UDP protocol, HTTP protocol, FTP protocol, SMTP protocol, DNS protocol and wireless LAN protocol;

[0077] The channel utilization requirements include quality of service QoS requirements and priority mechanisms. The quality of service QoS requirements include different types of communication requirements, including voice, video and text data transmission. The selected channel is adjusted according to the QoS requirements. For example, voice calls require lower latency and more stable channels, while high-definition video transmission requires higher bandwidth and stronger signal quality.

[0078] The historical data searches the corresponding background information through big data to obtain the historical channel selection information of each background information, wherein the historical channel selection information includes selecting the ath channel quality channel, the bth anti-interference method, the cth spectrum utilization channel, the dth network congestion relief method, the eth channel quality selection channel corresponding to the time-varying characteristic law, the fth spatial change selection channel, the gth user demand selection channel, the hth communication protocol, and the ith channel utilization requirement corresponding channel according to the background information, wherein a, b, c, d, e, f, g, h, and i are natural numbers respectively;

[0079] The anti-interference methods include frequency hopping, power control, spectrum management and dynamic spectrum access, channel coding and error correction, frequency hopping and code division multiple access technology, time division multiple access, spatial diversity and antenna technology beamforming, anti-interference filtering, modulation and demodulation technology, software defined radio, anti-interference waveforms, spread spectrum technology, noise suppression technology and multipath propagation;

[0080] The network congestion relief method includes flow control, congestion control, load balancing, traffic shaping, data compression, multi-path transmission, content distribution network CDN and network topology optimization;

[0081] The spatially varying channel selection includes selecting a channel with a preset weak channel quality interval threshold when the communication device has a strong signal in the current area;

[0082] When the communication device has a weak signal in the current area, a channel with a preset strong channel quality interval threshold is selected;

[0083] The user demand channel selection includes classifying the historical channel quality of user demand and behavior demand of big data search by using k-means clustering algorithm, and obtaining a first user demand channel, a second user demand channel, ... and a Gth user demand channel, where G is the total number of user demand channel classifications;

[0084] The channel utilization requirement corresponding channel includes using a k-means clustering algorithm to classify the historical channel requirements for voice, video and text data transmission searched by big data, and obtaining a first channel utilization requirement corresponding channel, a second channel utilization requirement corresponding channel, ... and an Ith channel utilization requirement corresponding channel, where I is the total number of channel utilization requirement corresponding channel categories.

[0085] The step S2 comprises:

[0086] All historical channel selection information corresponding to each background information is encapsulated into a historical database, the historical database includes the background information and all historical channel selection information corresponding to the background information, the historical database is stored in the memory in the form of a historical channel selection B-tree, and the historical channel selection B-tree is used as an update list to update the stored channel selection.

[0087] The step S3 comprises:

[0088] Detect the background information of the current communication situation, retrieve and obtain the ath channel quality channel, bth anti-interference method, cth spectrum utilization channel, dth network congestion relief method, time-varying characteristic law corresponding to the eth channel quality selection channel, fth spatial variation selection channel, gth user demand selection channel, hth communication protocol and ith channel utilization requirement corresponding channel required by the current communication situation from the historical channel selection B tree, and calculate the channel power, channel efficiency, user fairness, overall network throughput and total delay of the communication network under the communication situation, and its mathematical expression is:

[0089]

[0090]

[0091] C = B log2 (1 + SNR);

[0092]

[0093] T network =C·η ch ;

[0094] D total =D trans +D prop +D queue +D process ;

[0095] D trans =LR;

[0096] D prop =dv;

[0097] Among them, P ch is the channel power, E s is the symbol energy, T S is the symbol duration, η ch is the channel efficiency, R data is the effective data transmission rate, C is the channel capacity, B is the bandwidth, SNR is the signal-to-noise ratio, F Jainis Jain's fairness index, which is used to measure user fairness, T j is the channel throughput of the jth user, N is the number of users, the value range of the Jain's fairness index is [0,1]. In the range of [0,1], the larger the value of the Jain's fairness index is, the higher the fairness is. network is the network throughput, D total is the total delay of the communication network, D trans is the transmission delay, L is the size of the data packet, R is the link bandwidth, and D prop is the propagation delay, d is the distance the signal propagates, and v is the signal propagation speed. Signal propagation includes light propagation and electromagnetic wave propagation. D queue is the queuing delay, D process To deal with delays;

[0098] The channel power, channel efficiency, user fairness, overall network throughput and total delay of the communication network in the communication situation are calculated as the communication channel efficiency through weighted average.

[0099] The step S4 comprises:

[0100] The calculated communication channel efficiency under the current communication situation is used as the ideal target output, and the background information corresponding to the current communication situation is input into the embedding layer for feature extraction to obtain channel features and user features respectively. The channel features are obtained by extracting channel quality, interference, spectrum utilization, network load, communication protocol and channel utilization requirements through the embedding layer, and the user features are obtained by extracting time-varying characteristics, geographic space changes and user needs and behaviors through the embedding layer;

[0101] In the process of real-time analysis of the machine learning model of communication channel efficiency, the channel coding code rate is adjusted and changed through the dynamic adaptive coding method of segmented adaptive perforation with group sorting, the machine learning model is optimized, and the optimized machine learning model is trained to predict the optimal communication channel channel selection, so that the machine learning model is adapted to the channel selection of the communication channel and achieves maximization of the communication channel efficiency of the communication transmission under the premise of ensuring the overall transmission quality.

[0102] Optimizing machine learning models involves:

[0103] Divide the channel into multiple sub-areas, wherein the sub-areas include the ath channel quality channel, the bth anti-interference method, the cth spectrum utilization channel, the dth network congestion relief method, the eth channel quality selection channel corresponding to the time-varying characteristic law, the fth spatial variation selection channel, the gth user demand selection channel, the hth communication protocol and the ith channel utilization requirement corresponding channel stored in the historical channel selection B-tree, dynamically select the best coding and modulation mode for each sub-area, the channels in each group are different in coding and modulation strategies, and use puncturing technology for each channel segment to optimize the transmission of data packets. The puncturing technology adjusts the transmission rate according to the feedback of the current channel. The feedback is the communication channel efficiency calculated to reduce the bit error rate and improve the throughput;

[0104] The dynamic selection includes using the dynamic adjustment modulation mode 64-QAM. The dynamic adjustment logic includes:

[0105] When the channel quality of the channel is higher than a preset high quality threshold, high-order modulation is used to increase the data rate; when the channel quality of the channel is lower than the preset high quality threshold, low-order modulation is used to improve the anti-interference capability;

[0106] The channel characteristic signal and user characteristic signal of the input channel are encoded, modulated and resource mapped by dynamically adjusting the modulation mode 64-QAM, and the dynamically adjusted modulated signal is sent to the receiving end. After the receiving end demodulates and decodes the data, the current signal-to-noise ratio is obtained through channel estimation. The receiving end feeds back the demodulated and decoded data and the signal-to-noise ratio to the transmitting end. The receiving end and the transmitting end calculate the current channel condition according to the feedback through the iterative segmented adaptive puncturing method and switch the corresponding modulation and coding scheme. The modulation and coding scheme is used in the next transmission. By using adaptive coding technology, the coding scheme can be dynamically adjusted according to the real-time channel situation, and matrices with different code rates are used as the main code for puncturing to adapt to different channel conditions and transmission requirements. The use of segmented adaptive code rate puncturing can further improve the system transmission rate and throughput while ensuring the error performance. By layering and segmented multiplexing of data, more available subcarriers can be occupied, thereby improving the spectrum efficiency.

[0107] Taking channel characteristics and user characteristics as input and communication channel efficiency as target output, the random forest regression model is trained to predict the optimal channel selection. The training process includes:

[0108] The input channel characteristics and user characteristics are standardized, the historical channel selection B-tree is randomly divided into 75% training set and 25% validation set, the training set is used to train the random forest regression model, the mean square error is used to measure the prediction error of the random forest regression model using the validation set, the random forest regression model is continuously trained and adjusted until the mean square error of the random forest regression model is lower than a preset error threshold, and a channel selection model for simultaneous channel prediction and coding adjustment is obtained.

[0109] The background information of the current channel is collected and input into the trained channel selection model to obtain the predicted optimal channel selection. At the same time, the coding rate and modulation mode are dynamically adjusted through the iterative segmented adaptive puncturing method based on the predicted optimal channel selection. Combining segmented adaptive puncturing and dynamic adaptive coding, the channel selection model can further adjust the channel coding according to the segmented characteristics of different channels and real-time feedback.

[0110] A communication channel intelligent selection system includes a collection module, a database module, a calculation module and a channel allocation module:

[0111] The acquisition module is used to collect background information of communication conditions and collect historical data of background information of communication conditions;

[0112] The database module is used to encapsulate the historical data into a historical database, and search the historical database for corresponding historical data according to the background information;

[0113] The calculation module is used to calculate the communication channel efficiency corresponding to the background information of the current communication situation;

[0114] The channel allocation module is used to allocate communication channels using an improved machine learning model.

[0115] The present invention improves the reasons for channel selection, selects channels from multiple angles, improves the nonlinear regression prediction model, and adds a dynamic adaptive coding method of segmented adaptive punching with group sorting to the nonlinear regression prediction model analysis, so that the machine learning model is adapted to the channel selection of the communication channel, and maximizes the communication channel efficiency of the communication transmission while ensuring the overall transmission quality. By using adaptive coding technology, the coding scheme is dynamically adjusted according to the real-time channel situation, so as to adapt to different channel conditions and transmission requirements, and the transmission rate of the system is improved and the throughput is increased while ensuring the error performance.

[0116] It should be understood by those skilled in the art that the 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 complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program codes. Among them, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, disk or optical disk. These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0117] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A communication channel intelligent selection method, characterized in that: The following steps are involved: Step S1: Collecting background information of communication conditions, and collecting historical data of the background information of communication conditions; Step S2: Encapsulating the historical data into a historical database, and searching the historical database for corresponding historical data according to the background information; Step S3: Calculate the communication channel efficiency corresponding to the background information of the current communication situation; Step S4: Allocate the communication channel using the improved machine learning model.

2. The communication channel intelligent selection method according to claim 1, characterized in that: The step S1 comprises: The background information of the communication situation includes channel quality, interference situation, spectrum utilization, network load, time-varying characteristics, geographic space changes, user needs and behaviors, communication protocols and channel utilization requirements; The channel quality includes signal-to-noise ratio, channel gain, signal transmission distance, channel delay, signal transmission bandwidth, signal transmission rate, bit error rate and received signal strength; The interference conditions include co-channel interference, adjacent channel interference and external interference; The spectrum utilization includes idle spectrum detection and channel spectrum utilization rate; The network load includes the current channel load and the network topology connection status, and the network topology connection status is obtained from the routing forwarding table of the nearest neighbor router; The time-varying characteristics include the regularity of the channel changing with time; The geographic space changes include the shift in areas where communication equipment signals are strong or weak; The user requirements and behaviors include data rate requirements, user mobility requirements for signal quality, real-time requirements, and reliability requirements; The communication protocols include PPP protocol, TCP protocol, IP protocol, UDP protocol, HTTP protocol, FTP protocol, SMTP protocol, DNS protocol and wireless LAN protocol; The channel utilization requirements include quality of service QoS requirements and priority mechanisms. The quality of service QoS requirements include different types of communication requirements, including voice, video and text data transmission. The selected channel is adjusted according to the QoS requirements.

3. The communication channel intelligent selection method according to claim 2, characterized in that: The historical data searches the corresponding background information through big data to obtain the historical channel selection information of each background information, wherein the historical channel selection information includes selecting the ath channel quality channel, the bth anti-interference method, the cth spectrum utilization channel, the dth network congestion relief method, the eth channel quality selection channel corresponding to the time-varying characteristic law, the fth spatial change selection channel, the gth user demand selection channel, the hth communication protocol, and the ith channel utilization requirement corresponding channel according to the background information, wherein a, b, c, d, e, f, g, h, and i are natural numbers respectively; The anti-interference methods include frequency hopping, power control, spectrum management and dynamic spectrum access, channel coding and error correction, frequency hopping and code division multiple access technology, time division multiple access, spatial diversity and antenna technology beamforming, anti-interference filtering, modulation and demodulation technology, software defined radio, anti-interference waveforms, spread spectrum technology, noise suppression technology and multipath propagation; The network congestion relief method includes flow control, congestion control, load balancing, traffic shaping, data compression, multi-path transmission, content distribution network CDN and network topology optimization; The spatially varying channel selection includes selecting a channel with a preset weak channel quality interval threshold when the communication device has a strong signal in the current area; When the communication device has a weak signal in the current area, a channel with a preset strong channel quality interval threshold is selected; The user demand channel selection includes classifying the historical channel quality of user demand and behavior demand of big data search by using k-means clustering algorithm, and obtaining a first user demand channel, a second user demand channel, ... and a Gth user demand channel, where G is the total number of user demand channel classifications; The channel utilization requirement corresponding channel includes using a k-means clustering algorithm to classify the historical channel requirements for voice, video and text data transmission searched by big data, and obtaining a first channel utilization requirement corresponding channel, a second channel utilization requirement corresponding channel, ... and an Ith channel utilization requirement corresponding channel, where I is the total number of channel utilization requirement corresponding channel categories.

4. The communication channel intelligent selection method according to claim 3, characterized in that: The step S2 comprises: All historical channel selection information corresponding to each background information is encapsulated into a historical database, the historical database includes the background information and all historical channel selection information corresponding to the background information, the historical database is stored in the memory in the form of a historical channel selection B-tree, and the historical channel selection B-tree is used as an update list to update the stored channel selection.

5. The communication channel intelligent selection method according to claim 4, characterized in that: The step S3 comprises: Detect the background information of the current communication situation, retrieve and obtain the ath channel quality channel, bth anti-interference method, cth spectrum utilization channel, dth network congestion relief method, time-varying characteristic law corresponding to the eth channel quality selection channel, fth spatial variation selection channel, gth user demand selection channel, hth communication protocol and ith channel utilization requirement corresponding channel required by the current communication situation from the historical channel selection B tree, and calculate the channel power, channel efficiency, user fairness, overall network throughput and total delay of the communication network under the communication situation, and its mathematical expression is: C = B log2 (1 + SNR); T network =C·h ch ; D total =D trans +D prop +D queue +D process ; D trans =LR; D prop =dv; Among them, P ch is the channel power, E s is the symbol energy, T S is the symbol duration, η ch is the channel efficiency, R data is the effective data transmission rate, C is the channel capacity, B is the bandwidth, SNR is the signal-to-noise ratio, F Jain is Jain's fairness index, which is used to measure user fairness, T j is the channel throughput of the jth user, N is the number of users, the value range of the Jain's fairness index is [0,1]. In the range of [0,1], the larger the value of the Jain's fairness index is, the higher the fairness is. network is the network throughput, D total is the total delay of the communication network, D trans is the transmission delay, L is the size of the data packet, R is the link bandwidth, and D prop is the propagation delay, d is the distance the signal propagates, and v is the signal propagation speed. Signal propagation includes light propagation and electromagnetic wave propagation. D queue is the queuing delay, D process To deal with delays; The channel power, channel efficiency, user fairness, overall network throughput and total delay of the communication network in the communication situation are calculated as the communication channel efficiency through weighted average.

6. The communication channel intelligent selection method according to claim 5, characterized in that: The step S4 comprises: The calculated communication channel efficiency under the current communication situation is used as the ideal target output, and the background information corresponding to the current communication situation is input into the embedding layer for feature extraction to obtain channel features and user features respectively. The channel features are obtained by extracting channel quality, interference, spectrum utilization, network load, communication protocol and channel utilization requirements through the embedding layer, and the user features are obtained by extracting time-varying characteristics, geographic space changes and user needs and behaviors through the embedding layer; In the process of real-time analysis of the machine learning model of communication channel efficiency, the channel coding rate is adjusted and changed through the dynamic adaptive coding method of segmented adaptive perforation with group sorting, the machine learning model is optimized, and the optimized machine learning model is trained to predict the optimal communication channel selection.

7. The communication channel intelligent selection method according to claim 6, characterized in that: Optimizing machine learning models involves: Divide the channel into multiple sub-areas, wherein the sub-areas include the ath channel quality channel, the bth anti-interference method, the cth spectrum utilization channel, the dth network congestion relief method, the eth channel quality selection channel corresponding to the time-varying characteristic law, the fth spatial variation selection channel, the gth user demand selection channel, the hth communication protocol and the ith channel utilization requirement corresponding channel stored in the historical channel selection B-tree, dynamically select the best coding and modulation mode for each sub-area, the channels in each group are different in coding and modulation strategies, and use puncturing technology for each channel segment to optimize the transmission of data packets. The puncturing technology adjusts the transmission rate according to the feedback of the current channel, and the feedback is the communication channel efficiency calculated; The dynamic selection includes using the dynamic adjustment modulation mode 64-QAM. The dynamic adjustment logic includes: When the channel quality of the channel is higher than a preset high quality threshold, high order modulation is used; When the channel quality of the channel is lower than a preset high quality threshold, low order modulation is used; The channel characteristic signal and user characteristic signal of the input channel are encoded, modulated and resource mapped through the dynamically adjusted modulation mode 64-QAM, and the dynamically adjusted modulated signal is sent to the receiving end. After the receiving end demodulates and decodes the data, the current signal-to-noise ratio is obtained through channel estimation. The receiving end feeds back the demodulated and decoded data and the signal-to-noise ratio to the transmitting end. The receiving end and the transmitting end calculate the current channel condition through an iterative segmented adaptive puncturing method according to the feedback and switch to the corresponding modulation and coding scheme. The modulation and coding scheme is used in the next transmission, and matrices of different code rates are used as the main code for puncturing.

8. The communication channel intelligent selection method according to claim 7, characterized in that: Taking channel characteristics and user characteristics as input and communication channel efficiency as target output, the random forest regression model is trained to predict the optimal channel selection. The training process includes: The input channel characteristics and user characteristics are standardized, the historical channel selection B-tree is randomly divided into 75% training set and 25% validation set, the training set is used to train the random forest regression model, the mean square error is used to measure the prediction error of the random forest regression model using the validation set, the random forest regression model is continuously trained and adjusted until the mean square error of the random forest regression model is lower than a preset error threshold, and a channel selection model for simultaneous channel prediction and coding adjustment is obtained.

9. The communication channel intelligent selection method according to claim 8, characterized in that: The background information of the current channel is collected and input into the trained channel selection model to obtain the predicted optimal channel selection. At the same time, the coding rate and modulation mode are dynamically adjusted through an iterative segmented adaptive puncturing method based on the predicted optimal channel selection.

10. A communication channel intelligent selection system, characterized in that: Including acquisition module, database module, calculation module and channel allocation module: The acquisition module is used to collect background information of communication conditions and collect historical data of background information of communication conditions; The database module is used to encapsulate the historical data into a historical database, and search the historical database for corresponding historical data according to the background information; The calculation module is used to calculate the communication channel efficiency corresponding to the background information of the current communication situation; The channel allocation module is used to allocate communication channels using an improved machine learning model.

Images