A 5G terminal adaptive optimization method and system based on cognitive collaborative bus

Abstract

This application provides a 5G terminal adaptive optimization method and system based on a cognitive collaborative bus, belonging to the field of mobile communication and terminal performance management technology. The method includes: collecting multimodal state data and user scenario data of the 5G terminal; obtaining historical strategy execution success rates; adjusting the dynamic weights of multi-objective optimization based on the historical strategy execution success rates to obtain preliminary dynamic weights; determining the final dynamic weights of multi-objective optimization based on the multimodal state data, user scenario data, and the preliminary dynamic weights; generating optimization strategies for power consumption, signal strength, and stability based on the final dynamic weights; executing the optimization strategies in parallel and performing feedforward hardware pre-adjustment based on the multimodal state data; recording the execution effect of the optimization strategies, generating feedback data, and updating the historical strategy execution success rates to complete closed-loop optimization. This application dynamically adjusts strategies based on a cognitive bus, introduces hardware lifetime protection, and solves the problem of power consumption and signal balance in 5G terminals.

Description

Technical Field

[0001] This application belongs to the field of mobile communication and terminal performance management technology, specifically relating to a 5G terminal adaptive optimization method and system based on cognitive collaborative bus. Background Technology

[0002] With the large-scale deployment of 5G communication, terminal devices face severe challenges in scenarios involving high data throughput, low latency response, and multiple connections. Existing 5G terminals generally suffer from the following technical deficiencies in performance optimization: First, the optimization strategies are simplistic and have fixed weights. Most existing terminals balance power consumption and performance based on preset rules or static weights, failing to dynamically adjust according to real-time network environment changes, user behavior patterns, and historical optimization results. This leads to poor optimization performance in complex scenarios (such as frequent network switching during high-speed movement), and can even trigger a ping-pong effect, exacerbating power consumption. Second, the perception and execution stages are disconnected. In the traditional perception-decision-execution architecture, there is a delay in data flow between modules. The decision-making unit must wait for the perception unit to complete complete data collection and analysis before generating a strategy, which is then executed by the execution unit. This results in delayed response to sudden signal attenuation or scene switching, impacting user experience (such as video stuttering and call interruptions). Finally, there is a lack of long-term, hardware-friendly optimization evaluation mechanisms. Existing technologies primarily focus on the immediate effects of optimization strategies (such as power saving percentage), neglecting the cumulative wear and tear on hardware (such as antenna switches and CPU voltage regulators) caused by frequent strategy switching. This can lead to accelerated aging of terminal hardware and reduced overall device lifespan.

[0003] Therefore, there is an urgent need for a 5G terminal optimization method that can achieve real-time perception, dynamic decision-making, rapid execution, long-term performance evaluation, and self-evolution. Summary of the Invention

[0004] In a first aspect, embodiments of this application provide a 5G terminal adaptive optimization method based on a cognitive collaborative bus, comprising the following steps: S1. Collect multimodal status data and user scenario data of 5G terminals; S2. Obtain the historical strategy execution success rate through the cognitive collaboration bus, and adjust the dynamic weights of multi-objective optimization based on the historical strategy execution success rate to obtain the preliminary dynamic weights; S3. Based on multimodal state data, user scenario data and the preliminary dynamic weights, determine the final dynamic weights for multi-objective optimization, and generate optimization strategies for power consumption, signal strength and stability based on the final dynamic weights; S4. Execute the optimization strategy in parallel and perform feedforward hardware pre-adjustment based on multimodal state data; S5. Record the execution effect of the optimization strategy, generate feedback data and update the historical strategy execution success rate to complete the closed-loop optimization.

[0005] Furthermore, the multimodal state data in step S1 includes signal strength data of the 5G communication frequency band, terminal motion data, terminal battery status data, terminal hardware status data, and network quality data; the terminal motion data includes terminal location data and terminal acceleration data. The specific steps of step S1 are as follows: S11. The signal strength data of the 5G communication frequency band is monitored synchronously through the radio frequency front-end and baseband processor of the 5G terminal, wherein the 5G communication frequency band includes the Sub-6GHz frequency band and the millimeter wave frequency band; S12. Collect terminal location data via GPS and terminal acceleration data via accelerometer; S13. Collect terminal power status data through the power management chip of the 5G terminal, including remaining battery power, battery temperature and charge / discharge cycle count; S14. Collect terminal hardware status data through the system kernel interface of the 5G terminal, including CPU frequency, CPU temperature, screen status and memory usage. S15. Collect network quality data, including network latency, packet loss rate, and available bandwidth, through the protocol stack statistics interface of the 5G terminal; S16. Perform preprocessing and fusion analysis on the multimodal state data collected in steps S11-S15 to generate user scenario data.

[0006] Furthermore, step S16 is detailed as follows: S161. The collected GPS location data and acceleration data are fused together, and the accelerometer jitter error is eliminated through velocity variance analysis to calculate the terminal's moving speed; S162. Based on the calculated terminal moving speed and combined with the changing trend of the collected signal strength data, the motion scene of the terminal is identified by a classification neural network. The motion scene includes a stationary scene, a walking scene, a vehicle-mounted moving scene, and a high-speed rail moving scene. S163. Based on the collected signal strength data and combined with the collected terminal location data, identify the environmental scene in which the terminal is located. The environmental scene includes indoor scene, outdoor scene, elevator scene and basement scene. S164. Obtain the type of application currently running on the terminal through the application process monitoring interface, and identify the user's application scenario, which includes game scenario, video playback scenario, call scenario and standby scenario; S165. The identified motion scene, environmental scene, and application scene currently running on the terminal are fused together to generate user scene data; The user scenario data includes scenario type identifier and scenario priority attribute; The scenario type identifier is used to characterize the composite scenario in which the terminal is currently located. The composite scenario includes at least a weak signal mobile scenario, a high power consumption interaction scenario, and a low power consumption standby scenario. The scenario priority attribute is used to characterize the priority order of each optimization objective in the current scenario. The optimization objectives include latency, network speed, power consumption, and heat generation.

[0007] Furthermore, the specific steps of step S2 are as follows: S21. Obtain the historical policy execution success rate from the policy execution records stored locally on the 5G terminal; S23. When the historical strategy execution success rate is lower than the preset threshold, increase the weight factor related to signal stability in the dynamic weight of multi-objective optimization.

[0008] Furthermore, in step S22, increasing the weight factors related to signal stability in the dynamic weights of the multi-objective optimization specifically includes: When the historical strategy execution success rate When the signal is less than a preset first threshold, the weighting factor related to signal stability is increased by one step length, while the weighting factor related to power consumption is decreased by one step length. The preset first threshold is calculated by a preset function with the current battery level and signal strength as input, and the first step length is determined by a preset mapping table indexed by the signal stability sensitivity level of the current scene.

[0009] Furthermore, the specific steps of step S3 are as follows: S31. Based on the scenario type identifier and scenario priority attribute in the user scenario data, determine the dynamic weight allocation strategy and obtain the dynamic weight. , , and : If the scenario type is identified as a weak signal mobile scenario, dynamic weights are assigned according to the priority order of latency first, network speed second, power consumption third, and lowest heat generation. Highest, Secondly, Secondly, lowest; If the scenario type is identified as a high-power interaction scenario, then dynamic weights are assigned according to the priority order of network speed first, latency second, heat generation third, and lowest power consumption. Highest, Secondly, Secondly, lowest; If the scenario type is identified as a low-power standby scenario, then dynamic weights are assigned according to the priority order of power consumption first, heat generation second, network speed third, and lowest latency. Highest, Secondly, Secondly, lowest; in, , , , These represent the network speed weight, latency weight, power saving weight, and heat generation penalty weight, respectively, and satisfy the following conditions: ; S32. Based on the dynamic weights allocated in step S31, calculate the optimal optimization strategy using the following multi-objective optimization formula:

[0010] in, To normalize the estimated network speed, For normalized delay scoring, To save power consumption by normalizing Normalized calorific value; Normalized estimated network speed Calculated using the following formula:

[0011] in, To estimate network speed, calculate using the following formula: Estimated Internet Speed Obtained through the following methods:

[0012] in, This is the historical average internet speed. The current signal strength, This is a reference value for signal strength. Currently available bandwidth, This is a bandwidth reference value. For the current network latency, This is a time delay reference value. , , These are preset coefficients; S33. Use the calculated optimal strategy as the generated optimization strategy.

[0013] Furthermore, the specific steps of step S4 are as follows: S41. Execute optimization strategies to automatically switch the optimal signal reception direction when holding the device in landscape or portrait mode, and switch between at least three beamforming modes; S42. Adjust the power consumption of the display screen, processor and communication module of the 5G terminal respectively, and cut off the power supply of sensors unrelated to the current scene when the screen is off. The sensors unrelated to the current scene include at least one of accelerometer, gyroscope and magnetometer. S43. Acquire the raw sensing data. When a signal strength instantaneously drops below a preset threshold, activate the signal enhancement mode before generating an optimization strategy. The signal enhancement mode includes at least one of switching antennas, enabling beamforming, or increasing the receiving gain.

[0014] Furthermore, the determination condition for the instantaneous drop in signal strength exceeding a preset threshold in step S43 is as follows:

[0015] in, For the preset time window, The change in signal strength within, This is the preset descent rate threshold.

[0016] Furthermore, the specific steps of step S5 are as follows: S51. Record the execution effect of each optimization strategy and establish the correlation data between strategy switching frequency and hardware consumption: Record the cumulative number of times key hardware components are switched. Based on hardware tolerance threshold Calculate remaining lifetime :

[0017] When the cumulative number of switches When the average value within the preset time window exceeds the maximum tolerable frequency of the hardware components, or when the remaining lifespan... When the value is below the safety threshold, it is determined to be frequent adjustment, and a limit signal is sent to step S3 to avoid generating an optimization strategy that leads to frequent switching. S52. Compare the actual results with the expected goals, generate feedback data, and automatically mark strategies with a failure frequency higher than a preset frequency threshold and lower their decision priority; the feedback data includes the strategy execution success rate. Strategy execution latency, hardware wear and tear increments, and user satisfaction ratings; The strategy execution success rate in the feedback data Update the local policy execution record for use in step S2.

[0018] Secondly, embodiments of this application also provide a 5G terminal adaptive optimization system based on a cognitive collaborative bus, comprising: The intelligent sensing module is used to collect multimodal status data of 5G terminals and user scenario data; The preliminary dynamic weight determination module is used to obtain the historical strategy execution success rate through the cognitive collaboration bus, and adjust the dynamic weights of multi-objective optimization based on the historical strategy execution success rate to obtain the preliminary dynamic weights. The cognitive decision-making module is used to determine the final dynamic weights for multi-objective optimization based on multimodal state data, user scenario data and the preliminary dynamic weights, and to generate optimization strategies for power consumption, signal and stability based on the final dynamic weights. A heterogeneous execution module is used to execute the optimization strategy in parallel and perform feedforward hardware pre-adjustment based on multimodal state data; The effect evaluation and feedback module is used to record the execution effect of optimization strategies, generate feedback data, update the historical strategy execution success rate, and complete closed-loop optimization.

[0019] As can be seen from the above technical solutions, this application has the following advantages: The adaptive optimization method and system for 5G terminals based on cognitive collaborative bus provided in this application accurately identifies user complex scenarios by integrating GPS, accelerometer and radio frequency data; balances network speed, latency, power consumption and heat generation by using dynamic weight algorithm; eliminates response delay by introducing feedforward mechanism; and protects the lifespan of key components by establishing hardware loss model. This application improves the communication stability, energy efficiency ratio and long-term reliability of 5G terminals in complex environments. Attached Figure Description

[0020] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating the adaptive optimization method for 5G terminals based on cognitive collaborative bus of the present invention.

[0022] Figure 2 This is a schematic diagram of the 5G terminal adaptive optimization system based on cognitive collaborative bus of the present invention. Detailed Implementation

[0023] The various embodiments of this disclosure will be described more fully in the following detailed description of the specific steps of the 5G terminal adaptive optimization method based on cognitive cooperative bus. This disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of this disclosure to the specific embodiments disclosed herein, but rather this disclosure should be understood to cover all adjustments, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of this disclosure.

[0024] This embodiment provides a 5G terminal adaptive optimization method based on cognitive collaborative bus, which integrates multimodal perception and scene recognition. Through dynamic weight allocation and feedforward adjustment, it balances network speed, power consumption and heat generation, thereby extending the terminal's battery life and hardware lifespan.

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Please see Figure 1 The diagram shows a flowchart of a 5G terminal adaptive optimization method based on a cognitive collaborative bus in a specific embodiment. The method includes the following steps: S1. Collect multimodal status data and user scenario data of 5G terminals; It should be noted that by simultaneously monitoring Sub-6GHz and millimeter-wave frequency bands and fusing GPS and accelerometer data, the error of a single sensor can be eliminated, accurately constructing a real-time digital profile of the terminal, ensuring that subsequent decisions are not based on one-sided information, and providing a high-dimensional and high-precision input source for the entire optimization system; S2. Obtain the historical strategy execution success rate through the cognitive collaboration bus, and adjust the dynamic weights of multi-objective optimization based on the historical strategy execution success rate to obtain the preliminary dynamic weights; It should be noted that the dynamic weights in this step are different from static configurations. Historical success rate data is used to dynamically adjust the current optimization direction. If the historical strategy has a high failure rate in a certain environment, the system will automatically increase the weight of key indicators, realizing data-driven adaptive adjustment and avoiding simplistic rule matching. S3. Based on multimodal state data, user scenario data and the preliminary dynamic weights, determine the final dynamic weights for multi-objective optimization, and generate optimization strategies for power consumption, signal strength and stability based on the final dynamic weights; It should be noted that this step uses dynamic weights and, combined with the current scenario, calculates the Pareto optimal solution under the current environment through an algorithm. This ensures that the 5G terminal can automatically select the strategy that best meets the user's needs in different scenarios, achieving a balanced calculation of multiple objectives. S4. Execute the optimization strategy in parallel and perform feedforward hardware pre-adjustment based on multimodal state data; It should be noted that parallel execution improves optimization efficiency and shortens the time for strategy implementation; through feedforward pre-adjustment, there is no need to wait for the complete decision loop to end, but instead, hardware adjustments are directly triggered based on the original sensing data. This allows for the acquisition of valuable milliseconds of time in scenarios where the signal deteriorates rapidly, preventing communication interruption and ultimately solving the latency problem and improving response speed. S5. Record the execution effect of the optimization strategy, generate feedback data and update the historical strategy execution success rate to complete the closed-loop optimization; It should be noted that by recording the success rate and updating historical data, the system can achieve self-evolution and continuously correct algorithm parameters. By introducing hardware wear and lifespan calculations, when a certain optimization strategy would lead to excessively frequent hardware switching and thus shorten the lifespan, the system will automatically limit the generation of that strategy. This protects users' hardware assets while ensuring the long-term stability of 5G terminals.

[0027] This embodiment achieves multi-objective dynamic optimization through a closed loop of perception, decision-making, execution, and verification using a cognitive collaborative bus. It not only adaptively adjusts resources according to scenario priorities but also introduces hardware wear assessment to prevent hardware aging caused by frequent switching, thereby improving user experience and device durability.

[0028] Furthermore, as a refinement and extension of the specific implementation of the above embodiments, in order to fully illustrate the specific implementation process in this embodiment, another 5G terminal adaptive optimization method based on cognitive collaborative bus is provided. Taking the scenario of a user watching high-definition video on a high-speed train using a 5G terminal as an example, the user is traveling from city A to city B on a high-speed train at a speed of approximately 250 km / h; the user is using a 5G terminal to watch high-definition video, the initial signal strength is -85dBm, the battery level is 65%, the current network latency is 20ms, and the available bandwidth is 80MHz. The method includes the following steps: S1. Collect multimodal status data and user scenario data of 5G terminals; The multimodal status data in step S1 includes signal strength data of the 5G communication frequency band, terminal motion data, terminal battery status data, terminal hardware status data, and network quality data; the terminal motion data includes terminal location data and terminal acceleration data. The specific steps of step S1 are as follows: S11. The signal strength data of the 5G communication frequency band is monitored synchronously through the radio frequency front-end and baseband processor of the 5G terminal. The 5G communication frequency band includes the Sub-6GHz frequency band and the millimeter wave frequency band. Among them, Sub-6GHz refers to the 5G communication frequency band with a frequency lower than 6 GHz, i.e., 450MHz-6GHz, and the millimeter wave frequency band is 24GHz-100GHz. For example, the terminal synchronously monitors the signal strength of the Sub-6GHz band and the millimeter wave band through the radio frequency front-end and the baseband processor; at this time, the signal strength of the Sub-6GHz band is -85dBm and the signal strength of the millimeter wave band is -92dBm; due to the blockage of the millimeter wave signal by the high-speed rail body, the system automatically anchors the main communication link to the Sub-6GHz band; S12. Collect terminal location data via GPS and terminal acceleration data via accelerometer; For example, GPS collects location data and detects that the moving speed gradually increases to 250 km / h; MEMS accelerometer collects acceleration data and shows periodic fluctuations, reflecting the vibration characteristics of the train operation; S13. Collect terminal power status data through the power management chip of the 5G terminal, including remaining battery power, battery temperature and charge / discharge cycle count; For example, the power management chip collects the power status: battery remaining power 65%, battery temperature 38°C, and the number of charge-discharge cycles accumulated to 320. S14. Collect terminal hardware status data through the system kernel interface of the 5G terminal, including CPU frequency, CPU temperature, screen status and memory usage. For example, the system kernel interface collects hardware status: CPU current frequency 1.8GHz, CPU temperature 42℃, screen is on, and memory usage is 52%; S15. Collect network quality data, including network latency, packet loss rate, and available bandwidth, through the protocol stack statistics interface of the 5G terminal; For example, the protocol stack statistics interface collects network quality data: network latency 20ms, packet loss rate 0.3%, and available bandwidth 80MHz; S16. Perform preprocessing and fusion analysis on the multimodal state data collected in steps S11-S15 to generate user scenario data; The specific steps of step S16 are as follows: S161. The collected GPS location data and acceleration data are fused together, and the accelerometer jitter error is eliminated through velocity variance analysis to calculate the terminal's moving speed; For example, GPS speed data and acceleration data are fused, and accelerometer jitter error is eliminated through speed variance analysis. The calculated terminal moving speed is 248 km / h, which is determined to be a high-speed rail moving scenario. S162. Based on the calculated terminal moving speed and combined with the changing trend of the collected signal strength data, the motion scene of the terminal is identified by a classification neural network. The motion scene includes a stationary scene, a walking scene, a vehicle-mounted moving scene, and a high-speed rail moving scene. For example, based on a moving speed of 248 km / h and the trend of signal strength gradually decreasing from -85dBm to -92dBm, the classification neural network identifies that the terminal is currently in a high-speed rail moving scenario. S163. Based on the collected signal strength data and combined with the collected terminal location data, identify the environmental scene in which the terminal is located. The environmental scene includes indoor scene, outdoor scene, elevator scene and basement scene. For example, based on signal strength data (e.g., -92dBm) combined with GPS location information (e.g., recognizing that the current location is 2 kilometers in front of the entrance to a mountain tunnel), the terminal is identified as about to enter a weak signal environment. S164. Obtain the type of application currently running on the terminal through the application process monitoring interface, and identify the user's application scenario, which includes game scenario, video playback scenario, call scenario and standby scenario; For example, the application process monitoring interface is used to identify the current foreground application as a video playback application, thus recognizing that the user is in a video playback scenario; S165. The identified motion scene, environmental scene, and application scene currently running on the terminal are fused together to generate user scene data; The user scenario data includes scenario type identifier and scenario priority attribute; The scenario type identifier is used to characterize the composite scenario in which the terminal is currently located. The composite scenario includes at least a weak signal mobile scenario, a high power consumption interaction scenario, and a low power consumption standby scenario. The scenario priority attribute is used to characterize the priority order of each optimization objective in the current scenario. The optimization objectives include latency, network speed, power consumption, and heat generation. For example, the above-mentioned motion scenario (i.e., high-speed rail movement), environmental scenario (approaching a weak signal area), and application scenario (i.e., video playback) are merged to generate user scenario data: the scenario type is identified as a weak signal movement scenario, and the scenario priority attribute is set to latency first, network speed second, power consumption third, and heat generation lowest. S2. Obtain the historical strategy execution success rate through the cognitive collaboration bus, and adjust the dynamic weights of multi-objective optimization based on the historical strategy execution success rate to obtain the preliminary dynamic weights; The specific steps of step S2 are as follows: S21. Obtain the historical policy execution success rate from the policy execution records stored locally on the 5G terminal; For example, the historical policy execution success rate is obtained from the policy execution records stored locally on the 5G terminal. Querying policy execution records for similar scenarios over the past 24 hours reveals a total of 120 policy executions, with 98 successful executions. The historical policy execution success rate is then calculated. =81.7%; S22. When the historical strategy execution success rate is lower than the preset threshold, increase the weight factor related to signal stability in the dynamic weight of multi-objective optimization; In step S22, increasing the weight factors related to signal stability in the dynamic weights of multi-objective optimization specifically includes: When the historical strategy execution success rate When the signal is less than a preset first threshold, the weighting factor related to signal stability is increased by one step length, while the weighting factor related to power consumption is decreased by one step length. The preset first threshold is calculated by a preset function with the current battery power and signal strength as input, for example, the value range is 70% to 90%, and the first step length is determined by a preset mapping table indexed by the signal stability sensitivity level of the current scene, for example, the value range is 0.05 to 0.15. For example, if the historical strategy execution success rate of 81.7% is determined to be lower than a preset first threshold of 85% (the preset first threshold is dynamically set based on the current battery level of 65% and signal strength of -92dBm, with a value range of 70%-90%), a weight adjustment is triggered; the network speed weight is adjusted. Increase the length of the first step by 0.05, while also weighting power saving. Reduce by 0.05 to prioritize signal stability and video smoothness; Ultimately, the current dynamic weight configuration is: network speed weight. =0.40, delay weight =0.35, power saving weight =0.15, calorie penalty weight =0.10, which satisfies ; S3. Based on multimodal state data, user scenario data and the preliminary dynamic weights, determine the final dynamic weights for multi-objective optimization, and generate optimization strategies for power consumption, signal strength and stability based on the final dynamic weights; The specific steps of step S3 are as follows: S31. Based on the scenario type identifier and scenario priority attribute in the user scenario data, determine the dynamic weight allocation strategy and obtain the dynamic weight. , , and : If the scenario type is identified as a weak signal mobile scenario, then dynamic weights are assigned according to the priority order of latency first, network speed second, power consumption third, and lowest heat generation. Highest, Secondly, Secondly, lowest; If the scenario type is identified as a high-power interaction scenario, then dynamic weights are assigned according to the priority order of network speed first, latency second, heat generation third, and lowest power consumption. Highest, Secondly, Secondly, lowest; If the scenario type is identified as a low-power standby scenario, then dynamic weights are assigned according to the priority order of power consumption first, heat generation second, network speed third, and lowest latency. Highest, Secondly, Secondly, lowest; in, , , , These represent the network speed weight, latency weight, power saving weight, and heat generation penalty weight, respectively, and satisfy the following conditions: ; For example, based on the scenario type identifier "weak signal mobile scenario" in the user scenario data, dynamic weights are assigned according to the priority order of latency first, network speed second, power consumption third, and lowest heat generation, and the final weight configuration is confirmed as follows. =0.40、 =0.35、 =0.15、 =0.10; S32. Based on the dynamic weights allocated in step S31, calculate the optimal optimization strategy using the following multi-objective optimization formula:

[0029] in, To normalize the estimated network speed, For normalized delay scoring, To save power consumption by normalizing Normalized calorific value; , , , For dynamic weights, satisfying ; Normalized estimated network speed Calculated using the following formula:

[0030] in, To estimate network speed, calculate using the following formula: Estimated Internet Speed Obtained through the following methods:

[0031] in, This is the historical average internet speed. The current signal strength, This is a reference value for signal strength. Currently available bandwidth, This is a bandwidth reference value. For the current network latency, This is a time delay reference value. , , These are preset coefficients; For example, the optimal optimization strategy is calculated using a multi-objective optimization formula based on the assigned dynamic weights; first, the normalized indices are calculated: Normalized estimated network speed Calculate the historical average network speed using the estimated network speed formula. =45Mbps, current signal strength = 92dBm, signal strength reference value = 80dBm, currently available bandwidth =80 MHz Bandwidth reference value =100MHz, current network latency =20ms, latency reference value =15ms, preset coefficient =0.6、 =0.3、 =0.4; Calculate the estimated network speed:

[0032] Normalized estimated network speed =32.3 / 100=0.323; Normalized Delay Score Calculated based on the ratio of the current latency of 20ms to the target latency of 10ms. =0.85; Normalized power saving Calculated based on the ratio of current power consumption to standby power consumption. =0.60; Normalized calorific value Calculated based on the ratio of the current temperature to the overheat threshold. =0.55; Substitute into the multi-objective optimization formula: ; S33. Use the calculated optimal optimization strategy as the generated optimization strategy; For example, the computationally generated optimization strategy includes: Pre-caching strategy: Download the next 60 seconds of video content in advance and cache it to local storage; Antenna strategy: Switch to the antenna combination with higher signal gain (antennas 2 and 4 in the antenna array), enable beamforming mode 2, and align the beam direction with the base station signal direction; CPU strategy: Temporarily increase the CPU frequency from 1.8GHz to 2.2GHz to accelerate video decoding; Power consumption compensation: Automatically reduce screen brightness by 10% to balance the increased power consumption caused by CPU frequency increase; S4. Execute the optimization strategy in parallel and perform feedforward hardware pre-adjustment based on multimodal state data; The specific steps of step S4 are as follows: S41. Execute optimization strategies to automatically switch the optimal signal reception direction when holding the device in landscape or portrait mode, and switch between at least three beamforming modes; For example, the optimization strategy is implemented as follows: the reconfigurable antenna array is switched to beamforming mode 2, increasing the antenna gain by 3dB; the CPU frequency is increased to 2.2GHz, increasing the video decoding speed by approximately 25%. S42. Adjust the power consumption of the display screen, processor and communication module of the 5G terminal respectively, and cut off the power supply of sensors unrelated to the current scene when the screen is off. The sensors unrelated to the current scene include at least one of accelerometer, gyroscope and magnetometer. For example, the tiered power controller performs power consumption regulation: the screen brightness is reduced from 80% to 72%, the processor enters performance mode, and the communication module maintains the current transmit power; since the screen is on and the user is watching a video, the power supply to the non-essential sensors (magnetometer) is maintained, and the accelerometer and gyroscope continue to work to monitor the grip posture; S43. Acquire the raw sensing data. When a signal strength instantaneously drops below a preset threshold, activate the signal enhancement mode before generating an optimization strategy. The signal enhancement mode includes at least one of switching antennas, enabling beamforming, or increasing the receiving gain. The determination condition for the instantaneous drop in signal strength exceeding the preset threshold in step S43 is as follows:

[0033] in, For the preset time window, The change in signal strength within, The preset descent rate threshold; For example, the feedforward control interface monitors the raw sensing data in real time; when the terminal enters the first 300 meters of the tunnel, the signal strength drops sharply from -92dBm to -105dBm within 0.2 seconds, with a drop rate of 65dBm / second, exceeding the preset threshold. =30dBm / sec; Before the cognitive decision module completes the next round of strategy generation (approximately 50ms), the feedforward control interface first activates the signal enhancement mode: immediately switches to the lowest frequency band (700MHz) of the Sub-6GHz band, enables antenna diversity reception, and increases the reception gain by 5dB; This feedforward operation is completed before the cognitive decision module completes the calculation, effectively avoiding instantaneous stuttering in video playback; S5. Record the execution effect of the optimization strategy, generate feedback data and update the historical strategy execution success rate to complete the closed-loop optimization; The specific steps of step S5 are as follows: S51. Record the execution effect of each optimization strategy and establish the correlation data between strategy switching frequency and hardware consumption: Record the cumulative number of times key hardware components are switched. Based on hardware tolerance threshold Calculate remaining lifetime :

[0034] When the cumulative number of switches When the average value within the preset time window exceeds the maximum tolerable frequency of the hardware components, or when the remaining lifespan... When the value is below the safety threshold, it is determined to be frequent adjustment, and a limit signal is sent to step S3 to avoid generating an optimization strategy that leads to frequent switching. For example, record the effect of this strategy execution: After the strategy was implemented, video playback was smooth and the average network speed remained at 28Mbps. The cumulative number of antenna switching counts increases by 1, and the cumulative number of switching counts for key hardware components is recorded. =156; Based on hardware tolerance threshold =1000 times, calculate remaining lifetime =1000 156 = 844 times, which is far higher than the safety threshold (200 times), so the limiting signal is not triggered; S52. Compare the actual results with the expected goals, generate feedback data, and automatically mark strategies with a failure frequency higher than a preset frequency threshold and lower their decision priority; the feedback data includes the strategy execution success rate. Strategy execution latency, hardware wear and tear increments, and user satisfaction ratings; The strategy execution success rate in the feedback data Update the local policy execution record for use in step S2; For example, compare the actual results with the expected goals: Expected targets: network speed ≥ 25Mbps, power consumption increase ≤ 5%, temperature rise ≤ 3℃; Actual results: network speed 28Mbps, power consumption increased by 3.5%, CPU temperature increased by 2℃; Upon successful strategy execution, feedback data is generated: strategy execution success rate. =1 (success), strategy execution delay 38ms, hardware consumption increment 1 time, user satisfaction score (automatically calculated based on video smoothness) is 95 points; Success rate of strategy execution =1 is updated to the local policy execution record, and the success rate of the updated 24-hour sliding window is 82.5%; this data will serve as the basis for dynamic weight adjustment in the next step S2.

[0035] In this embodiment, when the 5G terminal enters the tunnel, the signal strength drops to -115dBm, and the historical strategy execution success rate drops to 65%; the cognitive collaboration bus triggers weight adjustment again, adjusting the latency weight. The weighting was increased from 0.35 to 0.45, indicating that network speed was a key factor in the weighting. The power consumption weight was increased from 0.40 to 0.45. The weight of the heat generation penalty was reduced from 0.15 to 0.05. The value was reduced from 0.10 to 0.05; the cognitive decision-making module generated a new optimization strategy: enable emergency communication mode, disable millimeter wave band, use only the Sub-6GHz low frequency band, enable VoLTE voice priority mode, and switch video playback to a lower resolution; the heterogeneous execution module implemented the new strategy to ensure that calls and basic communications were not affected.

[0036] After the 5G terminal exited the tunnel, the signal strength recovered to -85dBm, the historical strategy execution success rate rebounded to 90%, the dynamic weight gradually returned to the normal configuration, and the 5G terminal performance returned to normal levels.

[0037] In another scenario, the cumulative number of switching operations for a critical hardware component (such as an antenna switch) It reached 950 times within 24 hours, approaching the hardware tolerance threshold. =1000 times, remaining lifespan =50 times, which is lower than the safety threshold of 200 times; after the effect evaluation and feedback module detects this situation, it sends a limiting signal to the cognitive decision module. The cognitive decision module automatically suppresses optimization strategies that may cause frequent antenna switching (such as rapid switching of beamforming mode) when generating subsequent strategies, and instead adopts a more stable antenna configuration to protect hardware components from premature aging.

[0038] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0039] like Figure 2 As shown, the following are embodiments of the 5G terminal adaptive optimization system based on cognitive cooperative bus provided in this disclosure. This system and the 5G terminal adaptive optimization method based on cognitive cooperative bus in the above embodiments belong to the same inventive concept. For details not described in detail in the embodiments of the 5G terminal adaptive optimization system based on cognitive cooperative bus, please refer to the embodiments of the 5G terminal adaptive optimization method based on cognitive cooperative bus described above.

[0040] The system includes: The intelligent sensing module is used to collect multimodal status data of 5G terminals and user scenario data; The preliminary dynamic weight determination module is used to obtain the historical strategy execution success rate through the cognitive collaboration bus, and adjust the dynamic weights of multi-objective optimization based on the historical strategy execution success rate to obtain the preliminary dynamic weights. The cognitive decision-making module is used to determine the final dynamic weights for multi-objective optimization based on multimodal state data, user scenario data and the preliminary dynamic weights, and to generate optimization strategies for power consumption, signal and stability based on the final dynamic weights. A heterogeneous execution module is used to execute the optimization strategy in parallel and perform feedforward hardware pre-adjustment based on multimodal state data; The effect evaluation and feedback module is used to record the execution effect of optimization strategies, generate feedback data, update the historical strategy execution success rate, and complete closed-loop optimization.

[0041] This embodiment achieves communication continuity in weak signal environments through the interactive collaboration of the intelligent sensing module, cognitive collaboration bus, cognitive decision-making module, heterogeneous execution module, and effect evaluation and feedback module. By monitoring multimodal data in real time, using reinforcement learning to dynamically calculate the optimal weights, and combining feedforward hardware pre-adjustment, it ensures communication continuity in weak signal environments while also saving power consumption and ensuring hardware health. This solves the problems of single strategy and fragmented sensing and execution in traditional 5G terminals.

[0042] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A 5G terminal adaptive optimization method based on a cognitive collaborative bus, characterized in that, Includes the following steps: S1. Collect multimodal status data and user scenario data of 5G terminals; S2. Obtain the historical strategy execution success rate through the cognitive collaboration bus, and adjust the dynamic weights of multi-objective optimization based on the historical strategy execution success rate to obtain the preliminary dynamic weights; S3. Based on multimodal state data, user scenario data and the preliminary dynamic weights, determine the final dynamic weights for multi-objective optimization, and generate optimization strategies for power consumption, signal strength and stability based on the final dynamic weights; S4. Execute the optimization strategy in parallel and perform feedforward hardware pre-adjustment based on multimodal state data; S5. Record the execution effect of the optimization strategy, generate feedback data and update the historical strategy execution success rate to complete the closed-loop optimization.

2. The 5G terminal adaptive optimization method based on cognitive collaborative bus according to claim 1, characterized in that, The multimodal status data in step S1 includes signal strength data of the 5G communication frequency band, terminal motion data, terminal battery status data, terminal hardware status data, and network quality data; the terminal motion data includes terminal location data and terminal acceleration data. The specific steps of step S1 are as follows: S11. The signal strength data of the 5G communication frequency band is monitored synchronously through the radio frequency front-end and baseband processor of the 5G terminal, wherein the 5G communication frequency band includes the Sub-6GHz frequency band and the millimeter wave frequency band; S12. Collect terminal location data via GPS and terminal acceleration data via accelerometer; S13. Collect terminal power status data through the power management chip of the 5G terminal, including remaining battery power, battery temperature and charge / discharge cycle count; S14. Collect terminal hardware status data through the system kernel interface of the 5G terminal, including CPU frequency, CPU temperature, screen status and memory usage. S15. Collect network quality data, including network latency, packet loss rate, and available bandwidth, through the protocol stack statistics interface of the 5G terminal; S16. Perform preprocessing and fusion analysis on the multimodal state data collected in steps S11-S15 to generate user scenario data.

3. The 5G terminal adaptive optimization method based on cognitive collaborative bus according to claim 2, characterized in that, The specific steps of step S16 are as follows: S161. The collected GPS location data and acceleration data are fused together, and the accelerometer jitter error is eliminated through velocity variance analysis to calculate the terminal's moving speed; S162. Based on the calculated terminal moving speed and combined with the changing trend of the collected signal strength data, the motion scene of the terminal is identified by a classification neural network. The motion scene includes a stationary scene, a walking scene, a vehicle-mounted moving scene, and a high-speed rail moving scene. S163. Based on the collected signal strength data and combined with the collected terminal location data, identify the environmental scene in which the terminal is located. The environmental scene includes indoor scene, outdoor scene, elevator scene and basement scene. S164. Obtain the type of application currently running on the terminal through the application process monitoring interface, and identify the user's application scenario, which includes game scenario, video playback scenario, call scenario and standby scenario; S165. The identified motion scene, environmental scene, and application scene currently running on the terminal are fused together to generate user scene data; The user scenario data includes scenario type identifier and scenario priority attribute; The scenario type identifier is used to characterize the composite scenario in which the terminal is currently located. The composite scenario includes at least a weak signal mobile scenario, a high power consumption interaction scenario, and a low power consumption standby scenario. The scenario priority attribute is used to characterize the priority order of each optimization objective in the current scenario. The optimization objectives include latency, network speed, power consumption, and heat generation.

4. The 5G terminal adaptive optimization method based on cognitive collaborative bus according to claim 2, characterized in that, The specific steps of step S2 are as follows: S21. Obtain the historical policy execution success rate from the policy execution records stored locally on the 5G terminal; S22. When the historical strategy execution success rate is lower than the preset threshold, increase the weight factor related to signal stability in the dynamic weight of multi-objective optimization.

5. The 5G terminal adaptive optimization method based on cognitive collaborative bus according to claim 4, characterized in that, In step S22, increasing the weight factors related to signal stability in the dynamic weights of multi-objective optimization specifically includes: When the historical strategy execution success rate When the signal is less than a preset first threshold, the weighting factor related to signal stability is increased by one step length, while the weighting factor related to power consumption is decreased by one step length. The preset first threshold is calculated by a preset function with the current battery level and signal strength as input, and the first step length is determined by a preset mapping table indexed by the signal stability sensitivity level of the current scene.

6. The 5G terminal adaptive optimization method based on cognitive collaborative bus according to claim 4, characterized in that, The specific steps of step S3 are as follows: S31. Based on the scenario type identifier and scenario priority attribute in the user scenario data, determine the dynamic weight allocation strategy and obtain the dynamic weight. , , and : If the scenario type is identified as a weak signal mobile scenario, dynamic weights are assigned according to the priority order of latency first, network speed second, power consumption third, and lowest heat generation. Highest, Secondly, Secondly, lowest; If the scenario type is identified as a high-power interaction scenario, then dynamic weights are assigned according to the priority order of network speed first, latency second, heat generation third, and lowest power consumption. Highest, Secondly, Secondly, lowest; If the scenario type is identified as a low-power standby scenario, then dynamic weights are assigned according to the priority order of power consumption first, heat generation second, network speed third, and lowest latency. Highest, Secondly, Secondly, lowest; in, , , , These represent the network speed weight, latency weight, power saving weight, and heat generation penalty weight, respectively, and satisfy the following conditions: ; S32. Based on the dynamic weights allocated in step S31, calculate the optimal optimization strategy using the following multi-objective optimization formula: in, To normalize the estimated network speed, For normalized delay scoring, To save power consumption by normalizing Normalized calorific value; Normalized estimated network speed Calculated using the following formula: in, To estimate network speed, calculate using the following formula: Estimated Internet Speed Obtained through the following methods: in, This is the historical average internet speed. The current signal strength, This is a reference value for signal strength. Currently available bandwidth, This is a bandwidth reference value. For the current network latency, This is a time delay reference value. , , These are preset coefficients; S33. Use the calculated optimal optimization strategy as the generated optimization strategy.

7. The 5G terminal adaptive optimization method based on cognitive collaborative bus according to claim 6, characterized in that, The specific steps of step S4 are as follows: S41. Execute optimization strategies to automatically switch the optimal signal reception direction when holding the device in landscape or portrait mode, and switch between at least three beamforming modes; S42. Adjust the power consumption of the display screen, processor and communication module of the 5G terminal respectively, and cut off the power supply of sensors unrelated to the current scene when the screen is off. The sensors unrelated to the current scene include at least one of accelerometer, gyroscope and magnetometer. S43. Acquire the raw sensing data. When a signal strength instantaneously drops below a preset threshold, activate the signal enhancement mode before generating an optimization strategy. The signal enhancement mode includes at least one of switching antennas, enabling beamforming, or increasing the receiving gain.

8. The 5G terminal adaptive optimization method based on cognitive collaborative bus according to claim 7, characterized in that, The determination condition for the instantaneous drop in signal strength exceeding the preset threshold in step S43 is as follows: in, For the preset time window, The change in signal strength within, This is the preset descent rate threshold.

9. The 5G terminal adaptive optimization method based on cognitive collaborative bus according to claim 7, characterized in that, The specific steps of step S5 are as follows: S51. Record the execution effect of each optimization strategy and establish correlation data between strategy switching frequency and hardware consumption: Record the cumulative number of times key hardware components are switched. Based on hardware tolerance threshold Calculate remaining lifetime : When the cumulative number of switches When the average value within the preset time window exceeds the maximum tolerable frequency of the hardware components, or when the remaining lifespan... When the value is below the safety threshold, it is determined to be frequent adjustment, and a limit signal is sent to step S3 to avoid generating an optimization strategy that leads to frequent switching. S52. Compare the actual results with the expected goals, generate feedback data, and automatically mark strategies with a failure frequency higher than a preset frequency threshold and lower their decision priority; the feedback data includes the strategy execution success rate. Strategy execution latency, hardware wear and tear increments, and user satisfaction ratings; The strategy execution success rate in the feedback data Update the local policy execution record for use in step S2.

10. A 5G terminal adaptive optimization system based on a cognitive collaborative bus, characterized in that, include: The intelligent sensing module is used to collect multimodal status data of 5G terminals and user scenario data; The preliminary dynamic weight determination module is used to obtain the historical strategy execution success rate through the cognitive collaboration bus, and adjust the dynamic weights of multi-objective optimization based on the historical strategy execution success rate to obtain the preliminary dynamic weights. The cognitive decision-making module is used to determine the final dynamic weights for multi-objective optimization based on multimodal state data, user scenario data and the preliminary dynamic weights, and to generate optimization strategies for power consumption, signal and stability based on the final dynamic weights. A heterogeneous execution module is used to execute the optimization strategy in parallel and perform feedforward hardware pre-adjustment based on multimodal state data; The effect evaluation and feedback module is used to record the execution effect of optimization strategies, generate feedback data, update the historical strategy execution success rate, and complete closed-loop optimization.

Images