High-efficiency embedded encoder dynamic monitoring and optimizing system

Abstract

The invention relates to the technical field of information processing, in particular to a high-efficiency embedded encoder dynamic monitoring and optimizing system which comprises a data collection and preprocessing module, a model construction and performance evaluation module, a performance optimization strategy making module and an implementation and feedback circulation module. By collecting and preprocessing time sequence data of key performance indexes, the accuracy and reliability of the data are improved; whether the performance of the current encoder is superior to the historical optimal performance or not can be objectively and accurately evaluated by constructing a prediction model and establishing a dynamic reference; by formulating a targeted performance optimization strategy and setting a strategy pool, an optimal performance strategy is generated, and diversified candidate strategy options can be provided for an encoder; by implementing the optimal performance strategy and quantitatively analyzing the strategy implementation effect, more accurate and efficient performance optimization is realized. The method is used for solving the technical problem that the performance monitoring and optimization efficiency of the current encoder is not high.

Description

Technical Field

[0001] The present invention relates to the field of information processing technology, and in particular to a high-performance embedded encoder dynamic monitoring and optimization system. Background Art

[0002] Information processing technology refers to the technology of analyzing and processing the information reflected in the production and management process using scientific methods and means. It involves multiple links such as information collection, processing, and storage, and aims to organize and process large amounts of original information into the information people need.

[0003] In the current monitoring and optimization management of encoders, the following problems still exist: performance monitoring is not real-time and accurate. In current technologies, the performance monitoring of high-performance embedded encoders may not be real-time enough, or the collected data may not be accurate enough, resulting in the inability to timely reflect the true performance status of the encoder; performance evaluation lacks objective standards. Existing performance evaluation methods may lack objective and unified standards, making it difficult to accurately judge the gap between the current encoder's performance and historical performance; performance optimization strategies lack pertinence and intelligence. Current performance optimization strategies may be too general, lacking optimization solutions for specific performance indicators, and low optimization efficiency and intelligence. Therefore, it is necessary to use information processing technology to solve these problems in order to improve the level of monitoring and optimization management of encoders. Summary of the invention

[0004] The purpose of the present invention is to solve the problems in the background technology and to propose a high-performance embedded encoder dynamic monitoring and optimization system.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A high-performance embedded encoder dynamic monitoring and optimization system, including: a data collection and preprocessing module, a model building and performance evaluation module, a performance optimization strategy formulation module, and an implementation and feedback loop module;

[0007] Data collection and preprocessing module: collects the time series data of key performance indicators in the encoding process of the high-performance embedded encoder in real time, and uses statistical methods to identify abnormal values ​​of the time series data of key performance indicators; at the same time, configures adaptive sampling rate for sampling and collection;

[0008] Model building and performance evaluation module: Build a model to predict the time series data of key performance indicators of the high-performance embedded encoder at future moments; By establishing a dynamic benchmark, evaluate whether the performance of the current high-performance embedded encoder is better than the best historical performance, so as to obtain the performance evaluation results of the high-performance embedded encoder;

[0009] Performance optimization strategy formulation module: formulates performance optimization strategies based on the performance evaluation results of the high-performance embedded encoder, and generates the optimal performance strategy by setting the strategy pool;

[0010] Implementation and feedback loop module: Implement the optimal performance strategy, compare the time series data changes of key performance indicators before and after implementation, quantify the effect of the optimal performance strategy, and create a self-learning mechanism for feedback loop.

[0011] It should be noted that the application object of a high-efficiency embedded encoder dynamic monitoring and optimization system in an embodiment of the present invention can be the video encoding and optimization link in the field of multimedia processing. Specifically, it can be to comprehensively analyze and process the key performance indicator data generated in real time during the encoding process through time series analysis technology, aiming to accurately evaluate and optimize the performance of the high-efficiency embedded encoder; through multiple core steps such as data collection and preprocessing, model construction and performance evaluation, performance optimization strategy formulation, implementation and feedback loop, the encoding efficiency is improved, the waste of resources is reduced, and the dual improvement of encoding quality and stability is achieved. At the same time, the system also promotes the intelligence and automation of the encoder performance optimization strategy, so that the encoder can adaptively adjust the encoding parameters according to different encoding environments and requirements, so as to achieve the best encoding effect.

[0012] Furthermore, the data collection and preprocessing module collects the time series data of the key performance indicators in the encoding process of the high-efficiency embedded encoder in real time, and uses statistical methods to identify the abnormal values ​​of the time series data of the key performance indicators, including:

[0013] The key performance indicators of the high-performance embedded encoder are set as follows:

[0014] K={k1,k2,...,k n}; where n represents the number of key performance indicators; k1, k2, ..., k n Indicates the value of the first key performance indicator, the value of the second key performance indicator, ..., the value of the nth key performance indicator; the key performance indicators include encoding speed, bit rate, compression ratio, etc.;

[0015] For each key performance indicator k b , in the encoding process of the high-efficiency embedded encoder, data is collected at time intervals t to form time series data k b (t); where b represents the index of the collected key performance indicators;

[0016] Use statistical methods to analyze the time series data k of each key performance indicator b (t) Perform outlier detection to identify outliers: Get the mean of key performance indicators and standard deviation For each key performance indicator time series data k b (t), if Then the time series data k of the key performance indicator b (t) Marking as an outlier and replacing the identified outlier with a preset baseline value; wherein the identified outlier includes incomplete or abnormal values ​​such as missing values ​​and duplicate values.

[0017] Furthermore, the process of configuring the data collection and preprocessing module to perform sampling and collection with an adaptive sampling rate includes:

[0018] The thresholds for coding load and network status are defined as L th 、N th ;

[0019] According to the current coding load L(t) and network status N(t), adjust the sampling rate f(t) of the high-performance embedded encoder: If the current coding load L(t) is lower than or equal to the coding load threshold L th , and the current network status N(t) is higher than or equal to the network status threshold N th , then use a high sampling frequency f high ; If the current coding load L(t) is higher than the coding load threshold L th , or the current network status N(t) is lower than the network status threshold N th , then use a low sampling frequency f low ; In other cases, the standard sampling frequency is used to sample and collect the high-performance embedded encoder;

[0020] It is understandable that the encoding load refers to the amount of tasks or work intensity of a high-performance embedded encoder when processing video, audio or other types of data, reflecting the amount of data currently being processed by the encoder and the computing resources required to process the data. The level of the encoding load directly affects the performance and efficiency of the encoder. The value of the encoding load can be indirectly obtained by monitoring the CPU usage, memory occupancy, processing time and other parameters of the encoder during data processing. These parameters can reflect the resource consumption of the encoder when processing data, thereby inferring the size of the encoding load; the network status refers to the network environment faced by the high-performance embedded encoder during data transmission. It includes performance indicators such as network bandwidth, delay, jitter, etc. These indicators directly affect the transmission speed and stability of the encoder data. The value of the network status can be obtained through network monitoring tools or network performance testing equipment. These tools can monitor the network's bandwidth usage, delay time, data packet loss rate and other key indicators in real time, thereby comprehensively evaluating the quality of the network status. In addition, more detailed network performance data can be obtained by analyzing network logs or data packets.

[0021] Furthermore, the model building and performance evaluation module builds a model to predict the time series data of key performance indicators of the high-performance embedded encoder at future moments, including:

[0022] Select the ARIMA model to predict the time series of key performance indicators of the high-performance embedded encoder;

[0023] Extract the time series data of key performance indicators from the database as historical data for ARIMA model prediction and use it to fit the model;

[0024] Set the parameters p and q of the ARIMA model; where p is the order of the autoregressive term, which indicates the number of time series data of the historical key performance indicators used in the ARIMA model; q is the order of the moving average term, which indicates the number of forecast errors used in the ARIMA model;

[0025] The fitting process of the time series data of historical key performance indicators is represented by the difference equation of the ARIMA model:

[0026] k i (t)=φ1k i (t-1)+φ2k i (t-2)+…+φ p k i (tp)+ε t -θ1ε t-1 -θ2ε t-2 -…-θ q ε t-q ,

[0027] In the formula, k i (t) represents the time series data of the historical key performance indicators; i represents the time series data index of the historical key performance indicators; φ1, φ2, ..., φ p represents the autoregressive coefficients; θ1,θ2,...,θ q represents the moving average coefficient; ε t represents the prediction error at time t;

[0028] By fitting the time series data of historical key performance indicators k i (t) Determine the ARIMA model parameters p and q;

[0029] Based on the determined ARIMA model parameters p and q, predict the time series data of key performance indicators at future moments Where h represents the prediction step size.

[0030] Furthermore, the model building and performance evaluation module establishes a dynamic benchmark to evaluate whether the performance of the current high-efficiency embedded encoder is better than the best historical performance, so that the process of obtaining the performance evaluation result of the high-efficiency embedded encoder includes:

[0031] Based on the time series data k of each historical key performance indicator i (t) and the time series data k of key performance indicators collected in real time b (t) Establish dynamic benchmarks:

[0032] k dy (t) = α·k be (t)+(1-α)·k b (t),

[0033] In the formula, k dy (t) represents the dynamic benchmark value; α represents the balance coefficient, which is used to balance the optimal value k of the time series data of the historical key performance indicators. be (t) and the time series data k of key performance indicators collected in real time b (t), where the optimal value k be (t) is the time series data of historical key performance indicators k i The maximum value in (t) represents the highest performance level that the high-efficiency embedded encoder can achieve in the past moment;

[0034] According to the established dynamic benchmark, it is judged whether the performance of the current high-performance embedded encoder is better than the best performance in history, so as to obtain the performance evaluation result of the high-performance embedded encoder: if the time series data k of the current key performance indicator is b (t) is higher than the dynamic reference value k dy (t), then the performance of the current high-performance embedded encoder is judged to be better than the best performance in history, and a performance compliance signal is generated; if the time series data k of the current key performance indicator is b (t) not higher than the dynamic reference value k dy (t), it is determined that the performance of the current high-performance embedded encoder is inferior to the best historical performance, and a performance failure signal is generated.

[0035] Furthermore, the performance optimization strategy formulation module formulates a performance optimization strategy according to the performance evaluation result of the high-efficiency embedded encoder, and generates an optimal performance strategy by setting a strategy pool, including:

[0036] Obtain the performance substandard signal of the high-efficiency embedded encoder and mark the corresponding key performance indicator as k r ; Where r represents the key performance indicator index corresponding to the high-performance embedded encoder that does not meet the performance standard;

[0037] Define a coding parameter space ψ, which contains m coding parameters, that is, ψ={ψ1,ψ2,...,ψ m}; where the encoding parameter is the variable you want to optimize;

[0038] It can be understood that the coding parameter space refers to a set of all possible coding parameter combinations, in which each spatial point represents a set of specific coding parameter values; defining the coding parameter space is conducive to searching for the optimal solution in this space, that is, the coding parameter combination that maximizes the objective function;

[0039] In the encoding parameter space ψ, the objective function F(ψ) is established:

[0040]

[0041] In the formula, w r Denotes the key performance indicator k r The weight of; R represents the number of key performance indicators corresponding to the high-performance embedded encoder that does not meet the performance standard;

[0042] Use the optimization algorithm to search for the optimal solution ψ' in the encoding parameter space ψ so that the objective function F(ψ) reaches the maximum value:

[0043]

[0044] In the formula, Θ represents the range of all possible values ​​of the encoding parameter space ψ;

[0045] At the same time, during the optimization process, define the constraint set C = {C1, C2, ..., C l}, each constraint is represented by C j (ψ)≤0,j=1,2,...,l, where l represents the number of constraints and j represents the constraint index. During the optimization process, it is necessary to ensure that the generated performance optimization strategy satisfies all constraints, including hardware limitations, network bandwidth, etc.

[0046] Maintain a strategy pool S = {ψ'1,ψ'2,...,ψ' J}, where J represents the total number of performance strategies in the strategy pool;

[0047] It can be understood that the strategy pool is a set of strategies that store multiple strategies that meet the constraints and have good performance. These strategies are obtained through the previous optimization process and can be selected and adjusted under different operating environments and key performance indicators. The role of the strategy pool is to provide a variety of candidate strategy options for subsequent adaptive selection, so as to better adapt to the needs of the current environment;

[0048] Select the best performance strategy from the strategy pool S to achieve the best performance in the current encoding environment:

[0049]

[0050] In the formula, ψ sel represents the optimal performance strategy; v r Indicates k in the current encoding environment r The current encoding environment covers all external conditions and internal states related to the operation and performance of the high-performance embedded encoder, including but not limited to: hardware resources (such as the availability and performance of CPU, memory, storage devices, etc.), software environment (version and configuration of operating system, driver, dependent library, etc.), input data (input data type, format, quality, resolution, etc. of the high-performance embedded encoder), encoding requirements (target bit rate, frame rate, encoding standard, color space, etc.), etc.

[0051] Furthermore, the implementation and feedback loop module implements the optimal performance strategy, compares the time series data changes of key performance indicators before and after implementation, and quantifies the effect of the optimal performance strategy. The process includes:

[0052] According to the selected optimal performance strategy ψ sel Implementing the key performance indicators of high-performance embedded encoders that do not meet performance requirements r Optimization plan;

[0053] Collect key performance indicators k after implementing the optimal performance strategy r , forming new time series data

[0054] Compare the time series data changes of key performance indicators before and after the implementation of the optimal performance strategy, and calculate the improvement rate Δk r : Among them, the improvement rate Δk r Used to reflect the degree of improvement in key performance indicators after implementing the optimal performance strategy.

[0055] Furthermore, the implementation and feedback loop module creates a self-learning mechanism, and the process of the feedback loop includes:

[0056] Step Z1: The improvement rate Δk r As a feedback result, update the strategy pool S: when Δk r >0, the strategy is retained; when Δk r When ≤0, the strategy is eliminated; at the same time, new strategies are introduced into the strategy pool S to maintain the diversity and updateability of the strategy pool;

[0057] Step Z2: According to the feedback results, the high-performance embedded encoder whose performance does not meet the standard is adjusted, specifically including adjusting the encoding parameters or modifying the key performance indicators and their weights, so as to comprehensively improve the performance of the encoder;

[0058] It is understandable that if the encoder provides a wealth of parameter adjustment options corresponding to the optimal performance strategy, and these adjustments can significantly improve performance, the encoding parameters are adjusted directly; if the goal is to improve the overall performance of the encoder in the long term, and multiple key performance indicators need to be weighed and optimized, the key performance indicators of the encoder and their weights need to be adjusted; in practical applications, the two need to be combined, that is, the encoding parameters need to be adjusted, and the key performance indicators and their weights need to be fine-tuned to achieve the best performance optimization effect;

[0059] Step Z3: Repeat steps Z1-Z2 to form a closed-loop iterative optimization process.

[0060] Compared with the existing technology, the present invention provides a high-efficiency embedded encoder dynamic monitoring and optimization system and method, which has the following advantages:

[0061] 1. The present invention can improve the accuracy and reliability of data by collecting the time series data of key performance indicators in the encoding process of the high-efficiency embedded encoder in real time and using statistical methods to identify abnormal values ​​of the time series data of key performance indicators. The sampling rate is dynamically adjusted according to the encoding load and network conditions, which can reduce unnecessary resource consumption while ensuring data quality and improve the overall efficiency of the system.

[0062] 2. The present invention constructs a model to predict the time series data of key performance indicators of the high-efficiency embedded encoder at future moments, and establishes a dynamic benchmark to evaluate whether the performance of the current high-efficiency embedded encoder is better than the best historical performance, thereby obtaining the performance evaluation result of the high-efficiency embedded encoder, which can accurately evaluate whether the current performance meets the standard, and provide clear goals and directions for subsequent performance optimization;

[0063] 3. The present invention formulates a performance optimization strategy based on the performance evaluation results of the high-efficiency embedded encoder, and sets a strategy pool to generate the optimal performance strategy, so as to achieve the best performance under different operating environments and key performance indicators; by implementing the optimal performance strategy, comparing the time series data changes of the key performance indicators before and after implementation, quantifying the effect of the optimal performance strategy, and creating a self-learning mechanism for feedback loop, the performance level of the high-efficiency embedded encoder can be significantly improved, and the encoding efficiency and quality can be improved.

[0064] In summary, the present invention can significantly improve the performance level of high-efficiency embedded encoders through steps such as real-time data collection, performance prediction, optimization strategy formulation and feedback loop implementation, thereby ensuring efficient and stable operation of a subsequent high-efficiency embedded encoder dynamic monitoring and optimization system. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 This is a module diagram of a high-performance embedded encoder dynamic monitoring and optimization system proposed by the present invention. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the implementation regulations described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] Reference Figure 1 , a high-performance embedded encoder dynamic monitoring and optimization system, the system includes a data collection and preprocessing module, a model building and performance evaluation module, a performance optimization strategy formulation module, and an implementation and feedback loop module;

[0068] Data collection and preprocessing module: collects the time series data of key performance indicators in the encoding process of the high-performance embedded encoder in real time, and uses statistical methods to identify abnormal values ​​of the time series data of key performance indicators; at the same time, configures adaptive sampling rate for sampling and collection;

[0069] Model building and performance evaluation module: Build a model to predict the time series data of key performance indicators of the high-performance embedded encoder at future moments; By establishing a dynamic benchmark, evaluate whether the performance of the current high-performance embedded encoder is better than the best historical performance, so as to obtain the performance evaluation results of the high-performance embedded encoder;

[0070] Performance optimization strategy formulation module: formulates performance optimization strategies based on the performance evaluation results of the high-performance embedded encoder, and generates the optimal performance strategy by setting the strategy pool;

[0071] Implementation and feedback loop module: Implement the optimal performance strategy, compare the time series data changes of key performance indicators before and after implementation, quantify the effect of the optimal performance strategy, and create a self-learning mechanism for feedback loop.

[0072] It should be noted that the application object of a high-efficiency embedded encoder dynamic monitoring and optimization system in an embodiment of the present invention can be the video encoding and optimization link in the field of multimedia processing. Specifically, it can be a comprehensive analysis and data processing of key performance indicator data generated in real time during the encoding process through time series analysis technology, aiming to accurately evaluate and optimize the performance of high-efficiency embedded encoders; through multiple core steps such as data collection and preprocessing, model construction and performance evaluation, performance optimization strategy formulation, implementation and feedback loop, the encoding efficiency is improved, resource waste is reduced, and the dual improvement of encoding quality and stability is achieved. At the same time, the system also promotes the intelligence and automation of the encoder performance optimization strategy, so that the encoder can adaptively adjust the encoding parameters according to different encoding environments and requirements, so as to achieve the best encoding effect.

[0073] The data collection and preprocessing module collects the time series data of key performance indicators in the encoding process of the high-efficiency embedded encoder in real time, and uses statistical methods to identify abnormal values ​​of the time series data of the key performance indicators; at the same time, the steps of configuring the adaptive sampling rate for sampling and collection include:

[0074] Step 101: Set the key performance indicator set of the high-performance embedded encoder as follows:

[0075] K={k1,k2,...,k n}; where n represents the number of key performance indicators; k1, k2, ..., k n Indicates the value of the first key performance indicator, the value of the second key performance indicator, ..., the value of the nth key performance indicator; the key performance indicators include encoding speed, bit rate, compression ratio, etc.;

[0076] Step 102: For each key performance indicator k b , in the encoding process of the high-efficiency embedded encoder, data is collected at time intervals t to form time series data k b (t); where b represents the index of the collected key performance indicators;

[0077] Step 103: Use statistical methods to analyze the time series data k of each key performance indicator. b (t) Perform outlier detection to identify outliers: Get the mean of key performance indicators and standard deviation For each key performance indicator time series data k b (t), if Then the time series data k of the key performance indicator b (t) Mark as an outlier and replace the identified outlier with a preset baseline value;

[0078] In step 103, the identified abnormal values ​​include incomplete or abnormal values ​​such as missing values ​​and duplicate values;

[0079] Step 104: define the thresholds of coding load and network status as L th 、N th ;

[0080] Step 105: According to the current coding load L(t) and the network status N(t), adjust the sampling rate f(t) of the high-performance embedded encoder: if the current coding load L(t) is lower than or equal to the coding load threshold L th , and the current network status N(t) is higher than or equal to the network status threshold N th , then use a high sampling frequency f high ; If the current coding load L(t) is higher than the coding load threshold L th , or the current network status N(t) is lower than the network status threshold N th , then use a low sampling frequency f low ; In other cases, the standard sampling frequency is used to sample and collect the high-performance embedded encoder;

[0081] In step 105, the encoding load refers to the amount of tasks or work intensity of the high-efficiency embedded encoder when processing video, audio or other types of data, reflecting the amount of data currently being processed by the encoder and the computing resources required to process the data. The level of the encoding load directly affects the performance and efficiency of the encoder. The value of the encoding load can be indirectly obtained by monitoring the CPU usage, memory occupancy, processing time and other parameters of the encoder during data processing. These parameters can reflect the resource consumption of the encoder when processing data, thereby inferring the size of the encoding load; the network status refers to the network environment faced by the high-efficiency embedded encoder during data transmission, which includes performance indicators such as network bandwidth, delay, jitter, etc. These indicators directly affect the transmission speed and stability of the encoder data. The value of the network status can be obtained through a network monitoring tool or a network performance test device. These tools can monitor the key indicators of the network bandwidth usage, delay time, packet loss rate, etc. in real time, thereby comprehensively evaluating the quality of the network status. In addition, more detailed network performance data can be obtained by analyzing network logs or data packets.

[0082] The model building and performance evaluation module builds a model to predict the time series data of key performance indicators of the high-efficiency embedded encoder at future moments; by establishing a dynamic benchmark, the performance of the current high-efficiency embedded encoder is evaluated to see whether it is better than the best historical performance, thereby obtaining the performance evaluation results of the high-efficiency embedded encoder. The steps include:

[0083] Step 201, selecting an ARIMA model to predict the time series of key performance indicators of a high-efficiency embedded encoder;

[0084] Step 202: extracting time series data of key performance indicators from the database as historical data for ARIMA model prediction, for fitting the model;

[0085] Step 203: Set the parameters p and q of the ARIMA model; wherein p is the order of the autoregressive term, which indicates the number of time series data of the historical key performance indicators used in the ARIMA model; q is the order of the moving average term, which indicates the number of forecast errors used in the ARIMA model;

[0086] Step 204: The fitting process of the time series data of the historical key performance indicators is represented by the difference equation of the ARIMA model:

[0087] k i (t)=φ1k i (t-1)+φ2k i (t-2)+…+φ p k i (tp)+ε t -θ1ε t-1 -θ2ε t-2 -…-θ q ε t-q ,

[0088] In the formula, k i (t) represents the time series data of the historical key performance indicators; i represents the time series data index of the historical key performance indicators; φ1, φ2, ..., φ p represents the autoregressive coefficients; θ1,θ2,...,θ q represents the moving average coefficient; ε t represents the prediction error at time t;

[0089] Step 205: Fit the time series data k of historical key performance indicators i (t) Determine the ARIMA model parameters p and q;

[0090] Step 206: Based on the determined ARIMA model parameters p and q, predict the time series data of the key performance indicators at the future moment. Among them, h represents the prediction step length;

[0091] Step 207: Based on the time series data k of each historical key performance indicator i (t) and the time series data k of key performance indicators collected in real time b (t) Establish dynamic benchmarks:

[0092] k dy (t) = α·k be (t)+(1-α)·k b (t),

[0093] In the formula, k dy (t) represents the dynamic benchmark value; α represents the balance coefficient, which is used to balance the optimal value k of the time series data of the historical key performance indicators. be (t) and the time series data k of key performance indicators collected in real time b (t), where the optimal value k be (t) is the time series data of historical key performance indicators k i The maximum value in (t) represents the highest performance level that the high-efficiency embedded encoder can achieve in the past moment;

[0094] Step 208: Determine whether the performance of the current high-performance embedded encoder is better than the best performance in history according to the established dynamic benchmark, thereby obtaining the performance evaluation result of the high-performance embedded encoder: if the time series data k of the current key performance indicator is b (t) is higher than the dynamic reference value k dy (t), then the performance of the current high-performance embedded encoder is judged to be better than the best performance in history, and a performance compliance signal is generated; if the time series data k of the current key performance indicator is b (t) not higher than the dynamic reference value k dy (t), it is determined that the performance of the current high-performance embedded encoder is inferior to the best historical performance, and a performance failure signal is generated.

[0095] The performance optimization strategy formulation module formulates a performance optimization strategy according to the performance evaluation result of the high-efficiency embedded encoder, and generates the optimal performance strategy by setting a strategy pool, including the following steps:

[0096] Step 301: Obtain a performance failure signal of a high-efficiency embedded encoder, and mark the corresponding key performance indicator as k r ; Where r represents the key performance indicator index corresponding to the high-performance embedded encoder that does not meet the performance standard;

[0097] Step 302: define a coding parameter space ψ, where the coding parameter space ψ includes m coding parameters, i.e., ψ = {ψ1, ψ2, ..., ψ m};

[0098] In step 302, the coding parameter is a variable to be optimized; wherein the coding parameter space refers to a set of all possible coding parameter combinations, in which each space point represents a set of specific coding parameter values; defining the coding parameter space is conducive to searching for the optimal solution in this space, that is, the coding parameter combination that makes the objective function achieve the maximum value;

[0099] Step 303: Establish the objective function F(ψ) in the encoding parameter space ψ:

[0100]

[0101] In the formula, w r Denotes the key performance indicator k r The weight of; R represents the number of key performance indicators corresponding to the high-performance embedded encoder that does not meet the performance standard;

[0102] Step 304: Use an optimization algorithm to search for the optimal solution ψ' in the coding parameter space ψ so that the objective function F(ψ) achieves the maximum value:

[0103]

[0104] In the formula, Θ represents the range of all possible values ​​of the encoding parameter space ψ;

[0105] Step 305: During the optimization process, define a set of constraints C = {C1, C2, ..., C l}, each constraint is represented by C j (ψ)≤0,j=1,2,...,l, where l represents the number of constraints and j represents the constraint index;

[0106] In step 305, during the optimization process, it is necessary to ensure that the generated performance optimization strategy meets all constraints, including hardware limitations, network bandwidth, etc.;

[0107] Step 306: Maintain a strategy pool S = {ψ'1, ψ'2, ..., ψ' J}, where J represents the total number of performance strategies in the strategy pool;

[0108] In step 306, the strategy pool is a set of strategies that store multiple strategies that meet the constraints and have good performance. These strategies are obtained through the previous optimization process and can be selected and adjusted under different operating environments and key performance indicators. The role of the strategy pool is to provide a variety of candidate strategy options for subsequent adaptive selection, so as to better adapt to the needs of the current environment;

[0109] Step 307: Select the best performance strategy from the strategy pool S so that the performance is optimal in the current encoding environment:

[0110]

[0111] In the formula, ψ sel represents the optimal performance strategy; v r Indicates k in the current encoding environment r The weight of

[0112] In step 307, the current encoding environment covers all external conditions and internal states related to the operation and performance of the high-performance embedded encoder, including but not limited to: hardware resources (such as the availability and performance of CPU, memory, storage devices, etc.), software environment (version and configuration of operating system, driver, dependent library, etc.), input data (input data type, format, quality, resolution, etc. of the high-performance embedded encoder), encoding requirements (target bit rate, frame rate, encoding standard, color space, etc.), etc.

[0113] The implementation and feedback loop module implements the optimal performance strategy, compares the time series data changes of key performance indicators before and after implementation, quantifies the effect of the optimal performance strategy, and creates a self-learning mechanism. The steps of the feedback loop include:

[0114] Step 401: According to the selected optimal performance strategy ψ sel Implementing the key performance indicators of high-performance embedded encoders that do not meet performance requirements r Optimization plan;

[0115] Step 402: Collect the key performance indicators k after implementing the optimal performance strategy r , forming new time series data

[0116] Step 403: Compare the time series data changes of the key performance indicators before and after the implementation of the optimal performance strategy, and calculate the improvement rate Δk r : Among them, the improvement rate Δk r Used to reflect the degree of improvement of key performance indicators after implementing the optimal performance strategy;

[0117] Step 404: Create a self-learning mechanism and perform the following steps for a feedback loop:

[0118] Step Z1: The improvement rate Δk r As a feedback result, update the strategy pool S: when Δk r >0, the strategy is retained; when Δk r When ≤0, the strategy is eliminated; at the same time, new strategies are introduced into the strategy pool S to maintain the diversity and updateability of the strategy pool;

[0119] Step Z2: According to the feedback results, the high-performance embedded encoder whose performance does not meet the standard is adjusted, specifically including adjusting the encoding parameters or modifying the key performance indicators and their weights, so as to comprehensively improve the performance of the encoder;

[0120] In step Z2, if the encoder provides a wealth of parameter adjustment options corresponding to the optimal performance strategy, and these adjustments can significantly improve performance, the encoding parameters are adjusted directly; if the goal is to improve the overall performance of the encoder in the long term, and multiple key performance indicators need to be weighed and optimized, the key performance indicators of the encoder and their weights need to be adjusted; in practical applications, it is necessary to combine the two, that is, adjust the encoding parameters, and fine-tune the key performance indicators and their weights to achieve the best performance optimization effect;

[0121] Step Z3: Repeat steps Z1-Z2 to form a closed-loop iterative optimization process.

[0122] In the embodiment of the present invention, by collecting the time series data of the key performance indicators in the encoding process of the high-efficiency embedded encoder in real time, the timeliness and accuracy of the data are ensured, and the abnormal values ​​(such as missing values, duplicate values, etc.) in the time series data of the key performance indicators are identified and processed by statistical methods, so as to improve the reliability and quality of the data. At the same time, the sampling rate is adjusted according to the current encoding load and network conditions, and the data collection process is optimized, which not only ensures the integrity of the data, but also reduces unnecessary resource consumption. By building a model to predict the time series data of the key performance indicators at future moments and establishing a dynamic benchmark, it is evaluated whether the performance of the current encoder is better than the best performance in history, which provides a clear reference standard for performance optimization, and generates performance standards by comparing the current performance with the dynamic benchmark. The signal of the performance optimization strategy that is not up to standard or not up to standard is provided for the subsequent performance optimization strategy formulation. By formulating the optimization strategy for the key performance indicators that are not up to standard according to the performance evaluation results, the pertinence and effectiveness of the optimization are improved. By setting the strategy pool, a plurality of strategies that meet the constraints and have good performance are stored, and a variety of candidate strategy options are provided for the subsequent adaptive selection. By considering the constraints such as hardware limitations and network bandwidth in the optimization process, the feasibility and practicality of the generated performance optimization strategy are ensured. By comparing the changes of the key performance indicators before and after the implementation, the effect of the optimal performance strategy is quantified, and data support is provided for the subsequent optimization. By creating a self-learning mechanism for feedback loop and updating the strategy pool according to the improvement rate, the diversity and updateability of the strategy are maintained. In summary, the example of the present invention solves the problem that the current encoder performance monitoring and optimization efficiency is not high. In actual situations, more data and context information may be needed to make specific decisions and optimization plans.

[0123] In addition, the formulas involved in the above are all calculated by removing dimensions and taking their numerical values. They are a formula that is closest to the actual situation obtained by collecting a large amount of data and performing software simulation. The proportional coefficient in the formula and the various preset thresholds in the analysis process are set by technical personnel in this field according to actual conditions or obtained by simulating a large amount of data; the size of the proportional coefficient is to quantify each parameter to obtain a specific value for subsequent comparison. The size of the proportional coefficient depends on the amount of sample data and the preliminary setting of the corresponding processing coefficient for each group of sample data by technical personnel in this field; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0124] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically based on the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0125] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0126] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0127] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0128] 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.

[0129] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0130] Secondly: In the drawings of the embodiments disclosed in the present invention, only the structures related to the embodiments disclosed in the present invention are involved, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other;

[0131] Finally: The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A high-performance embedded encoder dynamic monitoring and optimization system, characterized by: It includes data collection and preprocessing module, model building and performance evaluation module, performance optimization strategy formulation module, and implementation and feedback loop module; Data collection and preprocessing module: collects the time series data of key performance indicators in the encoding process of the high-performance embedded encoder in real time, and uses statistical methods to identify abnormal values ​​of the time series data of key performance indicators; at the same time, configures adaptive sampling rate for sampling and collection; Model building and performance evaluation module: Build models to predict time series data of key performance indicators of high-performance embedded encoders at future moments; By establishing a dynamic benchmark, the performance of the current high-performance embedded encoder is evaluated to see whether it is better than the best historical performance, thereby obtaining the performance evaluation results of the high-performance embedded encoder; Performance optimization strategy formulation module: formulates performance optimization strategies based on the performance evaluation results of the high-performance embedded encoder, and generates the optimal performance strategy by setting the strategy pool; Implementation and feedback loop module: Implement the optimal performance strategy, compare the time series data changes of key performance indicators before and after implementation, quantify the effect of the optimal performance strategy, and create a self-learning mechanism for feedback loop.

2. The high-performance embedded encoder dynamic monitoring and optimization system according to claim 1, characterized in that: The data collection and preprocessing module collects the time series data of key performance indicators in the encoding process of the high-performance embedded encoder in real time, and uses statistical methods to identify abnormal values ​​of the time series data of the key performance indicators. The process includes: The key performance indicators of the high-performance embedded encoder are set as follows: K={k1,k2,...,k n }; where n represents the number of key performance indicators; k1, k2, ..., k n represents the value of the first key performance indicator, the value of the second key performance indicator, ..., the value of the nth key performance indicator; For each key performance indicator k b , in the encoding process of the high-efficiency embedded encoder, data is collected at time intervals t to form time series data k b (t); where b represents the index of the collected key performance indicators; Use statistical methods to analyze the time series data k of each key performance indicator b (t) Perform outlier detection to identify outliers: Get the mean of key performance indicators and standard deviation For each key performance indicator time series data k b (t), if Then the time series data k of the key performance indicator b (t) Mark as an outlier and replace the identified outlier with a preset baseline value.

3. The high-performance embedded encoder dynamic monitoring and optimization system according to claim 1, characterized in that: The process of configuring the adaptive sampling rate for sampling and collection in the data collection and preprocessing module includes: The thresholds for coding load and network status are defined as L th 、N th ; According to the current coding load L(t) and network status N(t), adjust the sampling rate f(t) of the high-performance embedded encoder: If the current coding load L(t) is lower than or equal to the coding load threshold L th , and the current network status N(t) is higher than or equal to the network status threshold N th , then use a high sampling frequency f high ; If the current coding load L(t) is higher than the coding load threshold L th , or the current network status N(t) is lower than the network status threshold N th , then use a low sampling frequency f low .

4. The high-performance embedded encoder dynamic monitoring and optimization system according to claim 1, characterized in that: The process of constructing a model and performance evaluation module to predict the time series data of key performance indicators of a high-performance embedded encoder at future moments includes: Select the ARIMA model to predict the time series of key performance indicators of the high-performance embedded encoder; Extract the time series data of key performance indicators from the database as historical data for ARIMA model prediction and use it to fit the model; Set the parameters p and q of the ARIMA model; where p is the order of the autoregressive term, which indicates the number of time series data of historical key performance indicators used in the ARIMA model; q is the order of the moving average term, which indicates the number of forecast errors used in the ARIMA model; The fitting process of the time series data of historical key performance indicators is represented by the difference equation of the ARIMA model: k i (t)=φ1k i (t-1)+φ2k i (t-2)+…+φ p k i (tp)+e t -θ1ε t-1 -θ2ε t-2 -…-θ q e t-q , In the formula, k i (t) represents the time series data of the historical key performance indicators; i represents the time series data index of the historical key performance indicators; φ1, φ2, ..., φ p represents the autoregressive coefficients; θ1,θ2,...,θ q represents the moving average coefficient; ε t represents the prediction error at time t; By fitting the time series data of historical key performance indicators k i (t) Determine the ARIMA model parameters p and q; Based on the determined ARIMA model parameters p and q, predict the time series data of key performance indicators at future moments Where h represents the prediction step size.

5. The high-performance embedded encoder dynamic monitoring and optimization system according to claim 4, characterized in that: The model building and performance evaluation module establishes a dynamic benchmark to evaluate whether the performance of the current high-performance embedded encoder is better than the best historical performance. The process of obtaining the performance evaluation result of the high-performance embedded encoder includes: Based on the time series data k of each historical key performance indicator i (t) and the time series data k of key performance indicators collected in real time b (t) Establish dynamic benchmarks: k dy (t)=α·k be (t)+(1-α)·k b (t), In the formula, k dy (t) represents the dynamic benchmark value; α represents the balance coefficient, which is used to balance the optimal value k of the time series data of the historical key performance indicators. be (t) and the time series data k of key performance indicators collected in real time b (t), where the optimal value k be (t) is the time series data of historical key performance indicators k i The maximum value in (t) represents the highest performance level that the high-efficiency embedded encoder can achieve in the past moment; According to the established dynamic benchmark, it is judged whether the performance of the current high-performance embedded encoder is better than the best performance in history, so as to obtain the performance evaluation result of the high-performance embedded encoder: if the time series data k of the current key performance indicator is b (t) is higher than the dynamic reference value k dy (t), then the performance of the current high-performance embedded encoder is judged to be better than the best performance in history, and a performance compliance signal is generated; if the time series data k of the current key performance indicator is b (t) not higher than the dynamic reference value k dy (t), it is determined that the performance of the current high-performance embedded encoder is inferior to the best historical performance, and a performance failure signal is generated.

6. The high-performance embedded encoder dynamic monitoring and optimization system according to claim 1, characterized in that: The performance optimization strategy formulation module formulates a performance optimization strategy according to the performance evaluation results of the high-performance embedded encoder, and generates the optimal performance strategy by setting the strategy pool, including the following process: Obtain the performance substandard signal of the high-efficiency embedded encoder and mark the corresponding key performance indicator as k r ; Where r represents the key performance indicator index corresponding to the high-performance embedded encoder that does not meet the performance standard; Define a coding parameter space ψ, which contains m coding parameters, that is, ψ={ψ1,ψ2,...,ψ m }; In the encoding parameter space ψ, the objective function F(ψ) is established: In the formula, w r Denotes the key performance indicator k r The weight of; R represents the number of key performance indicators corresponding to the high-performance embedded encoder that does not meet the performance standard; Use the optimization algorithm to search for the optimal solution ψ' in the encoding parameter space ψ so that the objective function F(ψ) reaches the maximum value: In the formula, Θ represents the range of all possible values ​​of the encoding parameter space ψ; At the same time, during the optimization process, define the constraint set C = {C1, C2, ..., C l }, each constraint is represented by C j (ψ)≤0,j=1,2,...,l, where l represents the number of constraints and j represents the constraint index; Maintain a strategy pool S = {ψ'1,ψ'2,...,ψ' J }, where J represents the total number of performance strategies in the strategy pool; Select the best performance strategy from the strategy pool S to achieve the best performance in the current encoding environment: In the formula, ψ sel represents the optimal performance strategy; v r Indicates k in the current encoding environment r The weight of .

7. The high-performance embedded encoder dynamic monitoring and optimization system according to claim 1, characterized in that: The implementation and feedback loop module implements the optimal performance strategy, compares the time series data changes of key performance indicators before and after implementation, and quantifies the effect of the optimal performance strategy. The process includes: According to the selected optimal performance strategy ψ sel Implementing the key performance indicators of high-performance embedded encoders that do not meet performance requirements r Optimization plan; Collect key performance indicators k after implementing the optimal performance strategy r , forming new time series data Compare the time series data changes of key performance indicators before and after the implementation of the optimal performance strategy, and calculate the improvement rate Δk r : Among them, the improvement rate Δk r Used to reflect the degree of improvement in key performance indicators after implementing the optimal performance strategy.

8. The high-performance embedded encoder dynamic monitoring and optimization system according to claim 7, characterized in that: The implementation and feedback loop module creates a self-learning mechanism. The feedback loop process includes: Step Z1: Set the improvement rate Δk r As a feedback result, update the strategy pool S: when Δk r >0, the strategy is retained; when Δk r When ≤0, the strategy is eliminated and a new strategy is introduced into the strategy pool S; Step Z2: adjusting the high-performance embedded encoder whose performance does not meet the standard according to the feedback result, specifically including adjusting the encoding parameters or modifying the key performance indicators and their weights; Step Z3: Repeat steps Z1-Z2 to form a closed-loop iterative optimization process.

Images