Adaptive signal transmission path optimization system of chip

Abstract

The invention belongs to the technical field of electronic information and communication, and discloses a self-adaptive signal transmission path optimization system of a chip, which comprises a data acquisition module used for acquiring monitoring video data, chip internal parameter data and chip transmission environment data; the data processing module is used for preprocessing the acquired monitoring video data, chip internal parameter data and chip transmission environment data to obtain a video feature data set, a chip internal parameter feature data set and a chip transmission environment feature data set; analyzing the influence of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set to obtain a transmission influence feature data set; the transmission error analysis module is used for training and obtaining a transmission error prediction model according to the transmission influence feature data set, and obtaining errors generated in the chip internal transmission process of the monitoring video signals based on the transmission error prediction model; the reliability of signal transmission is improved, and the chip resource allocation is more efficient.

Description

Technical Field

[0001] The present invention relates to the field of electronic information and communication technology, and more specifically, to a chip adaptive signal transmission path optimization system. Background Art

[0002] The patent application with the publication number CN118509042A discloses a dual-path redundant backup method based on optical fiber transmission, which relates to the field of video transmission technology, including receiving, encoding and redundant processing of video signals through a dual FPGA chip architecture; applying an adaptive coding system based on machine learning to automatically adjust coding parameters according to network conditions; using blockchain technology to verify the integrity and sequence management of each data packet; and using an intelligent fault prediction system to monitor and predict network faults in real time. The method of the present invention uses machine learning technology to automatically adjust coding parameters according to real-time network conditions, optimize the processing and transmission strategies of data packets, and adapt to network congestion and other dynamic changes, thereby improving transmission efficiency and reducing data loss.

[0003] The existing signal transmission path optimization system has the following main problems: Failure to consider the use of resampling formulas may result in failure to effectively handle the relationship between pixels, further exacerbating the degradation of image quality; lack of resolution scaling factor restrictions may cause image distortion when zooming in or out, resulting in blur, jagged or pixelated images; unreasonable frame rate adjustment may destroy the coherence of video content; lack of linear interpolation formulas and interpolation factor adjustment models, frame insertion may not be smooth and uniform, resulting in stuttering, jumping or incoherent images during video playback; failure to unify the color depth of video frames may result in redundancy of color information, increasing the burden of data processing and storage; lack of effective data compression and optimization methods may result in large video files, which is not conducive to transmission and storage; Failure to provide quantitative indicators for the evaluation of video clarity makes the discovery of video quality problems subjective and difficult to measure; without the support of quantitative indicators, quality problems such as blur and distortion in the video may be difficult to detect and deal with in a timely manner; lack of comprehensive monitoring of chip bandwidth, chip transmission delay, packet loss rate and bus congestion makes chip fault troubleshooting complicated and time-consuming; inefficient data processing may lead to waste of computing resources and storage resources, reducing resource utilization;

[0004] Time series data often has the characteristics of dynamic changes, and a fixed window length may not be able to effectively capture it. A window length that is too long contains too much historical data that is irrelevant to the current analysis, thereby diluting the weight of important information; a window length that is too short ignores the long-term trend and periodic characteristics of the data; failure to balance the weights of historical data and current data by adjusting the smoothing factor may result in unsatisfactory smoothing effects when the data fluctuates greatly, and may not accurately reflect the true trend of the data; failure to adjust the weight factor of the bus throughput may cause the system to be unable to accurately evaluate the actual load conditions of the transmission path in different time periods, thereby affecting the accuracy of resource allocation decisions. Summary of the invention

[0005] In view of this, the present invention proposes a chip adaptive signal transmission path optimization system to solve the above problem.

[0006] The present invention provides the following technical solutions: A chip adaptive signal transmission path optimization system, comprising: Data acquisition module, used to collect monitoring video data, chip internal parameter data and chip transmission environment data; The data processing module is used to pre-process the acquired monitoring video data, chip internal parameter data and chip transmission environment data to obtain a video feature data set, a chip internal parameter feature data set and a chip transmission environment feature data set; analyze the impact of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set to obtain a transmission impact feature data set; A transmission error analysis module is used to obtain a transmission error prediction model based on the transmission impact feature data set training, and obtain the error generated by the monitoring video signal during the transmission process inside the chip based on the transmission error prediction model; The transmission optimization evaluation module is used to compare the error generated by the predicted monitoring video signal during the transmission process inside the chip with the preset error threshold of the monitoring video signal during the transmission process inside the chip to determine whether signal transmission optimization is needed; Error type analysis module: if signal transmission optimization is not required, the chip signal transmission monitoring terminal issues a transmission instruction and continues to transmit the monitoring video signal; if signal transmission optimization is required, the chip signal transmission monitoring terminal issues an optimization instruction and collects signal transmission error data; The signal transmission optimization module is used to obtain a transmission error diagnosis model based on signal transmission error data training, and obtain the signal transmission error type based on the transmission error diagnosis model; the chip signal transmission monitoring terminal issues an error processing instruction according to the signal transmission error type and generates a corresponding error optimization strategy.

[0007] Furthermore, the monitoring video data includes video resolution, video frame rate, video clarity, color depth and video duration; the chip internal parameter data includes chip bandwidth, chip transmission delay, packet loss rate, bus congestion level and bus throughput; the chip transmission environment data includes the temperature of the transmission environment, the electromagnetic interference level and the transmission path load rate.

[0008] Furthermore, the method of preprocessing the acquired monitoring video data, chip internal parameter data, and chip transmission environment data to obtain a video feature data set, a chip internal parameter feature data set, and a chip transmission environment feature data set includes: S31. For surveillance video data: adjusting the video resolution at the pixel level by a resampling formula, converting each frame of image from the original resolution to the standard resolution; the resampling formula is: ;in, For the adjusted image in position The pixel value of For the original image at position The pixel value of The resolution scaling factor for converting each frame from the original resolution to the standard resolution; The resolution scaling factor for converting each frame image from the original resolution to the standard resolution is calculated using the scaling factor limiting formula. impose restrictions; The scaling factor limiting formula is: ;in, is the width of the original image; is the height of the original image; is the width of the standard image; is the height of the standard image; is the number of frames of the video; There are two adjacent frames by default and For each pair of adjacent frames, a linear interpolation formula is used to calculate and insert a and New frames between ; The linear interpolation formula is: ;in, For the Frame at position The pixel value of For the Frame at position The pixel value of is the interpolation factor; Adjust the model by interpolation factor Adjustment is performed, and the interpolation factor adjustment model is: ;in, For two adjacent frames and The number of new frames to insert between is the sequence number of the newly inserted frame; By adjusting , the position ratio of each inserted frame can be dynamically defined to make the distribution of frames uniform and smooth; For a frame rate The frame rate of the video is reduced by dropping frames. The preset video is reduced to Frame rate, calculate the required frame retention interval, the calculation formula is: ;in, Reserve intervals for required frames; is the floor rounding function; Use OpenCV to extract images from the video frame by frame, and calculate the sharpness of the video image for each frame using the Laplace operator formula, which reflects the video clarity. The Laplace operator formula is: ;in, For a single pixel at position The output value of the Laplace operator at is the sharpness of the video image; For images at position The pixel value of For images in The second derivative in the axis direction; For images in The second derivative in the axis direction; The different original image color depths in the video frame are uniformly converted into a standard color depth through a color quantization formula; the color quantization formula is: ;in, The quantized image is at position The pixel value of is the color depth of the original image in the video frame; is the standard color depth; The entire security surveillance video is divided into v time windows through the FFmpeg multimedia processing tool. For each time window, the time points within the time window are calculated. The video feature values ​​on the image are obtained, and the features of each time window are summarized by the time window feature summary formula; the time window feature summary formula is: ;in, It is the time window feature summary value of the entire security surveillance video; For time point The characteristic value of is the length of the time window; are continuous time points; Adjust the length of the time window for the model by using the window length Adjustment is performed, and the window length adjustment model is: ;in, For the A weight factor related to the time window feature; For the The number of characteristics; For time point Related characteristic functions; Types of characteristics; is the index of the feature type; An adjustment factor for adjusting the length of the time window in the window length adjustment model; Time The relevant characteristic function expression is: ;in, For time point The characteristic value of is the time interval between time points; S32. For chip internal parameter data: smooth the chip bandwidth using an exponential smoothing formula; The exponential smoothing formula is: ;in, For time The smoothing bandwidth value at the moment; For time The smoothing bandwidth value at the moment; For time The original bandwidth value at the time; is the smoothing factor; The smoothing factor is adjusted by a smoothing factor adjustment formula, and the smoothing factor adjustment formula is: ;in, For in time The number of smoothing bandwidth data points within ; The time required to smooth the chip bandwidth; An adjustment coefficient for controlling the influence of the time required for smoothing chip bandwidth on the smoothing factor; For chip transmission delay, quantile analysis is used to identify abnormal delays; for packet loss rate, weighted average method is used to Summarize the packet loss in each time period and find The packet loss rate in each time period; for the bus congestion level, the sliding window method is used to calculate The average bus congestion level in a time period; For bus throughput, use the cumulative throughput formula to calculate arrive The cumulative bus throughput in the time period; the cumulative throughput formula is: ;in, For arrive The total throughput of the bus during the time period; The instantaneous throughput at a given moment; For the moment The weight factor affecting bus throughput; The weight factor of the bus throughput is adjusted by a weight factor adjustment model, and the weight factor adjustment model is: ;in, For the moment The weight factor affecting bus throughput; is the total number of moments in the time period; is the standard deviation of the bus throughput during the time period; S33. For chip transmission environment data: For the temperature of the transmission environment, Fourier transform is used to identify the periodic changes in the temperature of the transmission environment; for the electromagnetic interference level, a wavelet transform formula is used to perform multi-scale analysis on the electromagnetic interference level to identify interference characteristics on different time scales; the wavelet transform formula is: ;in, For scale and location The wavelet transform result under ; For each time point The level of electromagnetic interference on is the mother wavelet function; For the transmission path load rate, the transmission path load rate characteristic value in each time window segmented by the FFmpeg multimedia processing tool is calculated, and the calculation formula is: ;in, is the transmission path load rate characteristic summary value; For time point The transmission path load factor characteristic value; The LOF algorithm is used on the processed monitoring video data, chip internal parameter data and chip transmission environment data to identify the abnormal values ​​in the monitoring video data, chip internal parameter data and chip transmission environment data, and the identified abnormal values ​​are removed to obtain the processed video feature data set, chip internal parameter feature data set and chip transmission environment feature data set; The acquired video feature dataset, chip internal parameter feature dataset and chip transmission environment feature dataset are subjected to standard deviation normalization and converted into standard normal distribution, and finally the normalized video feature dataset, chip internal parameter feature dataset and chip transmission environment feature dataset are obtained.

[0009] Furthermore, the method of analyzing the influence of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set to obtain the transmission influence feature data set includes: A joint regression model is constructed to quantitatively analyze the impact of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set, and a transmission impact feature data set is obtained; the joint regression model is: ;in, To transmit comprehensive features of the feature dataset; For the Characteristics of the chip transmission environment; An index of environmental characteristics transmitted to the chip; The number of environmental characteristics transmitted to the chip; For the Video features; is the index of video features; is the number of video features; For the Internal parameter characteristics of each chip; is the index of the internal parameter characteristics of the chip; is the number of internal parameter features of the chip; is the regression coefficient of chip transmission environment characteristics; is the regression coefficient of the video feature; is the regression coefficient of the chip internal parameter characteristics; is the regression coefficient of the interaction term between the video feature and the chip internal parameter feature; is the error term of the model; is the constant term of the joint regression model.

[0010] Furthermore, the training method of the transmission error prediction model includes: The data set is divided into a training set, a validation set, and a test set to train the model and evaluate the model performance; the sample set is a subset of the data set, and each sample set includes a historical transmission impact feature data set and the errors generated by the corresponding monitoring video signal during the transmission process inside the chip; a transmission error prediction model is constructed, including an input layer, a GRU layer, a fully connected layer, and an output layer; The input layer of the transmission error prediction model is used to input the historical transmission impact feature data set. The number of neurons in the input layer should match the number of features in the historical transmission impact feature data set. The GRU layer is used to process the historical comprehensive feature data set. The number of GRU layers and the number of neurons are adjusted according to the complexity of the task. The fully connected layer is used to provide additional nonlinear transformation. The output layer of the model is used to output the error generated by the monitoring video signal during the transmission process inside the chip. The number of neurons in the output layer corresponds to the number of prediction targets. The prediction value is output through one neuron, and the identity function is used as the activation function. The transmission error prediction model is a gated recurrent unit model. Define the loss function of the model, and use the L2 regularized mean square error loss function to measure the difference between the error generated during the predicted security surveillance video transmission process and the error generated during the actual security surveillance video transmission process; The L2 regularized mean square error loss function is: ;in, is the sample size; For the The predicted value of samples; For the The actual value of the samples, is the regularization coefficient; is the number of training samples, is the sum of squares of all trainable weights; The transmission error prediction model is trained using the training set, and the model parameters are updated through the back propagation algorithm to minimize the loss function. The performance of the transmission error prediction model is evaluated by calculating the accuracy index using the validation set. The Adam optimization algorithm is used as the optimizer. The model is tuned according to the performance feedback of the validation set. The model parameters are adjusted until the performance is no longer significantly improved or the preset number of iterations is reached. The test set is used to evaluate the performance of the model in the prediction task, and the trained transmission error prediction model is used to predict the current transmission impact feature data set to obtain the error generated by the monitoring video signal during the transmission process inside the chip.

[0011] Furthermore, the method of comparing the predicted error generated by the monitoring video signal during the internal transmission of the chip with a preset error threshold of the monitoring video signal during the internal transmission of the chip to determine whether signal transmission optimization is required includes: If the error generated by the predicted monitoring video signal during the transmission process inside the chip is less than the preset error threshold of the monitoring video signal during the transmission process inside the chip, it is determined that signal transmission optimization is not required; If the error generated by the predicted monitoring video signal during the transmission inside the chip is greater than or equal to the error threshold generated by the monitoring video signal during the transmission inside the chip, it is determined that signal transmission optimization is required.

[0012] Furthermore, the signal transmission error data includes a signal packet loss rate, a signal retransmission rate, a bit error rate, a signal loss rate and a data packet arrival sequence error rate during chip transmission.

[0013] Furthermore, the training method of the transmission error diagnosis model includes: The data set is divided into a training set, a validation set, and a test set to train the model and evaluate the model performance; the sample set is a subset of the data set, and each sample set includes historical signal transmission error data and the corresponding signal transmission error type; Construct a transmission error diagnosis model, which includes an input layer, a hidden layer and an output layer; the input layer is used to input historical signal transmission error data, and the output layer is used to output the signal transmission error type; the output layer is provided with neurons equal to the number of signal transmission error types, and each neuron corresponds to the predicted probability of a signal transmission error type; a softmax function is used as an activation function; the transmission error diagnosis model is a multi-layer perceptron MLP model; Multi-classification cross entropy is used as the loss function of the model to measure the difference between the predicted value and the actual value of the model; the multi-classification cross entropy loss function is: ;in, is the average loss of the dataset; is the total number of samples in the data set; is the number of signal transmission error types; For the For the sample The true labels of each type; The model predicts The samples belong to The probability of a type; The transmission error diagnosis model is trained using the training set, and the model parameters are updated through the back propagation algorithm to minimize the loss function. The performance of the transmission error diagnosis model is evaluated by calculating the accuracy index using the validation set. The SGD optimization algorithm is selected as the optimizer, and the model is tuned according to the performance feedback of the validation set. The model parameters are adjusted until the performance is no longer significantly improved or the preset stopping condition is reached. The test set is used to evaluate the performance of the model in the prediction task, and the trained transmission error diagnosis model is used to predict the current signal transmission error data to obtain the signal transmission error type.

[0014] Furthermore, the chip signal transmission monitoring terminal issues an error processing instruction according to the signal transmission error type, and the method for generating a corresponding error optimization strategy includes: The error severity analysis of the signal transmission error type is performed using an error severity evaluation model, and the error severity evaluation model is: ;in, is the severity of the error; is the signal packet loss rate during chip transmission; is the signal retransmission rate; is the bit error rate during signal transmission; is the signal loss rate; is the packet arrival sequence error rate; Preset error severity threshold, further divide the preset error severity threshold into low error severity threshold and high error severity threshold ; The severity of the error The lower error severity threshold and high error severity threshold For comparison, if , then the signal transmission error type is determined to be a low error type; if , then the signal transmission error type is determined to be a moderate error type; if , then the signal transmission error type is determined to be a height error type; For low error types, the chip signal transmission monitoring terminal issues a third-level error processing instruction to generate a mild error optimization strategy; for moderate error types, the chip signal transmission monitoring terminal issues a second-level error processing instruction to generate a moderate error optimization strategy; for high error types, the chip signal transmission monitoring terminal issues a first-level error processing instruction to generate a high error optimization strategy.

[0015] Furthermore, the mild error optimization strategy includes adjusting the signal buffer settings, optimizing the signal encoding settings and monitoring and adjusting the bus load; the moderate error optimization strategy includes enhancing the signal retransmission mechanism, adjusting the signal frequency and implementing error correction codes; the severe error optimization strategy includes increasing the number of signal data packet retransmissions, increasing the signal buffer capacity and switching the signal transmission path.

[0016] The present invention provides a chip adaptive signal transmission path optimization system, which has the following beneficial effects: The resampling formula and resolution scaling factor limit ensure the conversion quality of video frames at different resolutions, avoiding image blur or distortion caused by resolution mismatch; the scaling factor limit formula ensures that the image will not be distorted due to excessive enlargement or reduction during the resolution conversion process; the most appropriate scaling factor is dynamically calculated according to the size ratio of the original image and the standard image to maintain the clarity and details of the image; the linear interpolation formula and interpolation factor adjustment model make the insertion of frames smoother and more uniform, improving the fluency of the video; at the same time, by discarding frames to reduce the frame rate, the amount of data can be reduced while maintaining the consistency of the video content; the color depth of the video frame is unified, the redundancy of color information is reduced, and the data processing efficiency and accuracy are improved; The Laplace operator is used to calculate the sharpness of the video image, which provides a quantitative indicator for the evaluation of video clarity and helps to detect video quality problems in a timely manner. By smoothing the chip bandwidth, analyzing the chip transmission delay, packet loss rate and bus congestion, as well as the bus network throughput, the internal state of the chip is fully monitored, providing data support for chip optimization and troubleshooting. Periodic and multi-scale analysis of the temperature and electromagnetic interference level of the transmission environment helps to identify potential environmental interference factors and improve the stability and reliability of data transmission. FFmpeg is used to split the video into time windows and summarize the features in each window, which reduces the complexity and time cost of data processing. The weight factors and characteristic functions in the window length adjustment model can be adjusted and optimized according to the needs of specific application scenarios; the model can adapt to different data characteristics and analysis objectives, and provide support for time series data analysis in various complex scenarios; by adjusting the smoothing factor, the weights of historical data and current data in the smoothed value can be balanced. When the chip bandwidth fluctuates greatly, the value of the smoothing factor can be appropriately increased to consider more historical data, thereby reducing the impact of fluctuations on the smoothed value; by adjusting the weight factor of the bus throughput, the system can more accurately evaluate the actual load of the transmission path in different time periods, thereby making more reasonable resource allocation decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of the structure of an adaptive signal transmission path optimization system of a chip of the present invention; Figure 2 The present invention is a schematic flow chart of a method for optimizing a chip's adaptive signal transmission path. DETAILED DESCRIPTION

[0018] 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 described embodiments 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.

[0019] Example 1, please refer to Figure 1 In this embodiment, a chip adaptive signal transmission path optimization system includes: Data acquisition module, used to collect monitoring video data, chip internal parameter data and chip transmission environment data; The data processing module is used to pre-process the acquired monitoring video data, chip internal parameter data and chip transmission environment data to obtain a video feature data set, a chip internal parameter feature data set and a chip transmission environment feature data set; analyze the impact of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set to obtain a transmission impact feature data set; A transmission error analysis module is used to obtain a transmission error prediction model based on the transmission impact feature data set training, and obtain the error generated by the monitoring video signal during the transmission process inside the chip based on the transmission error prediction model; The transmission optimization evaluation module is used to compare the error generated by the predicted monitoring video signal during the transmission process inside the chip with the preset error threshold of the monitoring video signal during the transmission process inside the chip to determine whether signal transmission optimization is needed; Error type analysis module: if signal transmission optimization is not required, the chip signal transmission monitoring terminal issues a transmission instruction and continues to transmit the monitoring video signal; if signal transmission optimization is required, the chip signal transmission monitoring terminal issues an optimization instruction and collects signal transmission error data; The signal transmission optimization module is used to obtain a transmission error diagnosis model based on signal transmission error data training, and obtain the signal transmission error type based on the transmission error diagnosis model; the chip signal transmission monitoring terminal issues an error processing instruction according to the signal transmission error type and generates a corresponding error optimization strategy.

[0020] Monitoring video data includes video resolution, video frame rate, video clarity, color depth and video duration; chip internal parameter data includes chip bandwidth, chip transmission delay, packet loss rate, bus congestion level and bus throughput; chip transmission environment data includes transmission environment temperature, electromagnetic interference level and transmission path load rate.

[0021] The method of preprocessing the acquired monitoring video data, chip internal parameter data and chip transmission environment data to obtain a video feature data set, a chip internal parameter feature data set and a chip transmission environment feature data set includes: S31. For surveillance video data: adjusting the video resolution at the pixel level by a resampling formula, converting each frame of image from the original resolution to the standard resolution; the resampling formula is: ;in, For the adjusted image in position The pixel value of For the original image at position The pixel value of The resolution scaling factor for converting each frame from the original resolution to the standard resolution; The resolution scaling factor for converting each frame image from the original resolution to the standard resolution is calculated using the scaling factor limiting formula. impose restrictions; The scaling factor limiting formula is: ;in, is the width of the original image; is the height of the original image; is the width of the standard image; is the height of the standard image; is the number of frames of the video; For example, the width of the original image 6, the height of the original image 5, the width of the standard image 10, the width of the standard image is 10, the number of video frames is 3, then the resolution scaling factor is 0.9; There are two adjacent frames by default and For each pair of adjacent frames, a linear interpolation formula is used to calculate and insert a and New frames between ; The linear interpolation formula is: ;in, For the Frame at position The pixel value of For the Frame at position The pixel value of is the interpolation factor; Adjust the model by interpolation factor Adjustment is performed, and the interpolation factor adjustment model is: ;in, For two adjacent frames and The number of new frames to insert between is the sequence number of the newly inserted frame; For example, two adjacent frames and The number of new frames to insert between 4 frames, the sequence number of the newly inserted frame 6, the number of video frames For 5 frames, the interpolation factor is 0.3; By adjusting , the position ratio of each inserted frame can be dynamically defined, so that the distribution of frames is uniform and smooth; For a frame rate The frame rate of the video is reduced by dropping frames. The preset video is reduced to Frame rate, calculate the required frame retention interval, the calculation formula is: ;in, Reserve intervals for required frames; is the floor rounding function; Use OpenCV to extract images from the video frame by frame, and calculate the sharpness of the video image for each frame using the Laplace operator formula, which reflects the video clarity. The Laplace operator formula is: ;in, For a single pixel at position The output value of the Laplace operator at is the sharpness of the video image; For images at position The pixel value of For the image The second derivative in the axis direction; For the image The second derivative in the axis direction; The different original image color depths in the video frame are uniformly converted to a standard color depth through a color quantization formula; the color quantization formula is: ;in, The quantized image is at position The pixel value of is the color depth of the original image in the video frame; is the standard color depth; The entire security surveillance video is divided into v time windows through the FFmpeg multimedia processing tool. For each time window, the time points within the time window are calculated. The video feature values ​​on the image are obtained, and the features of each time window are summarized by the time window feature summary formula; the time window feature summary formula is: ;in, It is the time window feature summary value of the entire security surveillance video; For time point The characteristic value of is the length of the time window; are continuous time points; Adjust the length of the time window for the model by using the window length Adjustment is performed, and the window length adjustment model is: ;in, For the A weight factor related to the time window feature; For the The number of characteristics; For time point Related characteristic functions; Types of characteristics; is the index of the feature type; An adjustment factor for adjusting the length of the time window in the window length adjustment model; Time The relevant characteristic function expression is: ;in, For time point The characteristic value of is the time interval between time points; For example, if the feature type is 3 and the time point is 10. There are three weight factors related to time window features. The weight factor of video clarity is 0.5, and the number of video clarity features is 8. The weight factor of frame rate is 0.3, and the number of frame rate features is 20. The weight factor of key frame number is 0.2, and the number of key frame features is 5. The adjustment coefficient of the time window length in the window length adjustment model is 1.2. Suppose a simple time point The relevant characteristic function expression is , indicating the changing trend of the feature over time; The sum of the weighted number of features ; Assume that the time point is selected is 5, , substitute into the model to calculate the time window length ; Supplementary explanation: The signal transmission path optimization process is based on video surveillance data. Video surveillance data is collected, transmitted and processed through the chip. The signal transmission path optimization process in the chip is aimed at improving the efficiency and stability of video surveillance data transmission; Specifically, when the video surveillance data passes through the chip, it will be used as a signal input and will be subject to the constraints and control of the signal transmission path optimization; the optimization module will dynamically adjust the chip signal transmission according to the characteristics of the video surveillance data (such as resolution, frame rate, clarity, etc.) to ensure the optimal signal quality and transmission effect during the monitoring data transmission process, thereby ensuring the real-time and image quality of the video surveillance system.

[0022] S32. For chip internal parameter data: smooth the chip bandwidth using an exponential smoothing formula; The exponential smoothing formula is: ;in, For time The smoothing bandwidth value at the moment; For time The smoothing bandwidth value at the moment; For time The original bandwidth value at the time; is the smoothing factor; The smoothing factor is adjusted by a smoothing factor adjustment formula, and the smoothing factor adjustment formula is: ;in, For in time The number of smoothing bandwidth data points within ; The time required to smooth the chip bandwidth; An adjustment coefficient for controlling the influence of the time required for smoothing chip bandwidth on the smoothing factor; For example, at time The number of smoothing bandwidth data points within =15, the time required to smooth the chip bandwidth The adjustment coefficient of 10 seconds is used to control the influence of the time required for smoothing chip bandwidth on the smoothing factor. is 0.2, then the adjusted smoothing factor is approximately 0.88; By adjusting, you can control the chip bandwidth smoothing degree; the larger the value, the more emphasis is placed on current data, and the smaller the value, the more historical data is considered; For chip transmission delay, quantile analysis is used to identify abnormal delays; for packet loss rate, weighted average method is used to Summarize the packet loss in each time period and find The packet loss rate in each time period; for the bus congestion level, the sliding window method is used to calculate The average bus congestion level in a time period; For bus throughput, use the cumulative throughput formula to calculate arrive The cumulative bus throughput in the time period; the cumulative throughput formula is: ;in, For arrive The total throughput of the bus during the time period; The instantaneous throughput at a given moment; For the moment The weight factor affecting bus throughput; The weight factor of the bus throughput is adjusted by a weight factor adjustment model, and the weight factor adjustment model is: ;in, For the moment The weight factor affecting bus throughput; is the total number of moments in the time period; is the standard deviation of the bus throughput during the time period; For example, the total number of moments in a time period is 5, the standard deviation of the bus throughput in the time period is 2, then the weight factor of the bus throughput is 0.4; S33. For chip transmission environment data: For the temperature of the transmission environment, Fourier transform is used to identify the periodic changes in the temperature of the transmission environment; for the electromagnetic interference level, a wavelet transform formula is used to perform multi-scale analysis on the electromagnetic interference level to identify interference characteristics on different time scales; the wavelet transform formula is: ;in, For scale and location The wavelet transform result under ; For each time point The level of electromagnetic interference on is the mother wavelet function; For the transmission path load rate, the transmission path load rate characteristic value in each time window segmented by the FFmpeg multimedia processing tool is calculated, and the calculation formula is: ;in, is the transmission path load rate characteristic summary value; For time point The transmission path load factor characteristic value; The LOF algorithm is used on the processed monitoring video data, chip internal parameter data and chip transmission environment data to identify the abnormal values ​​in the monitoring video data, chip internal parameter data and chip transmission environment data, and the identified abnormal values ​​are removed to obtain the processed video feature data set, chip internal parameter feature data set and chip transmission environment feature data set; The acquired video feature dataset, chip internal parameter feature dataset and chip transmission environment feature dataset are subjected to standard deviation normalization and converted into standard normal distribution, and finally the normalized video feature dataset, chip internal parameter feature dataset and chip transmission environment feature dataset are obtained.

[0023] The method of analyzing the influence of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set to obtain the transmission influence feature data set includes: A joint regression model is constructed to quantitatively analyze the impact of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set, and a transmission impact feature data set is obtained; the joint regression model is: ;in, To transmit comprehensive features of the feature dataset; For the Characteristics of the chip transmission environment; An index of environmental characteristics transmitted to the chip; The number of environmental characteristics transmitted to the chip; For the Video features; is the index of video features; is the number of video features; For the Internal parameter characteristics of each chip; is the index of the internal parameter characteristics of the chip; is the number of internal parameter features of the chip; is the regression coefficient of chip transmission environment characteristics; is the regression coefficient of the video feature; is the regression coefficient of the chip internal parameter characteristics; is the regression coefficient of the interaction term between the video feature and the chip internal parameter feature; is the error term of the model; is the constant term of the joint regression model.

[0024] The training method of the transmission error prediction model includes: The data set is divided into a training set, a validation set, and a test set to train the model and evaluate the model performance; the sample set is a subset of the data set, and each sample set includes a historical transmission impact feature data set and the errors generated by the corresponding monitoring video signal during the transmission process inside the chip; a transmission error prediction model is constructed, including an input layer, a GRU layer, a fully connected layer, and an output layer; The input layer of the transmission error prediction model is used to input the historical transmission impact feature data set. The number of neurons in the input layer should match the number of features in the historical transmission impact feature data set. The GRU layer is used to process the historical comprehensive feature data set. The number of GRU layers and the number of neurons are adjusted according to the complexity of the task. The fully connected layer is used to provide additional nonlinear transformation. The output layer of the model is used to output the error generated by the monitoring video signal during the transmission process inside the chip. The number of neurons in the output layer corresponds to the number of prediction targets. The prediction value is output through one neuron, and the identity function is used as the activation function. The transmission error prediction model is a gated recurrent unit model. Define the loss function of the model, and use the L2 regularized mean square error loss function to measure the difference between the error generated during the predicted security surveillance video transmission process and the error generated during the actual security surveillance video transmission process; The L2 regularized mean square error loss function is: ;in, is the sample size; For the The predicted value of samples; For the The actual value of the samples, is the regularization coefficient; is the number of training samples, is the sum of squares of all trainable weights; The transmission error prediction model is trained using the training set, and the model parameters are updated through the back propagation algorithm to minimize the loss function. The performance of the transmission error prediction model is evaluated by calculating the accuracy index using the validation set. The Adam optimization algorithm is used as the optimizer. The model is tuned according to the performance feedback of the validation set. The model parameters are adjusted until the performance is no longer significantly improved or the preset number of iterations is reached. The test set is used to evaluate the performance of the model in the prediction task, and the trained transmission error prediction model is used to predict the current transmission impact feature data set to obtain the error generated by the monitoring video signal during the transmission process inside the chip.

[0025] The method of comparing the predicted error generated by the monitoring video signal during the transmission process inside the chip with the preset error threshold generated by the monitoring video signal during the transmission process inside the chip to determine whether signal transmission optimization is required includes: If the error generated by the predicted monitoring video signal during the transmission process inside the chip is less than the preset error threshold of the monitoring video signal during the transmission process inside the chip, it is determined that signal transmission optimization is not required; If the error generated by the predicted monitoring video signal during the transmission inside the chip is greater than or equal to the error threshold generated by the monitoring video signal during the transmission inside the chip, it is determined that signal transmission optimization is required.

[0026] The signal transmission error data includes the signal packet loss rate, signal retransmission rate, bit error rate, signal loss rate and data packet arrival sequence error rate during chip transmission.

[0027] The training method of the transmission error diagnosis model includes: The data set is divided into a training set, a validation set, and a test set to train the model and evaluate the model performance; the sample set is a subset of the data set, and each sample set includes historical signal transmission error data and the corresponding signal transmission error type; Construct a transmission error diagnosis model, which includes an input layer, a hidden layer and an output layer; the input layer is used to input historical signal transmission error data, and the output layer is used to output the signal transmission error type; the output layer is provided with neurons equal to the number of signal transmission error types, and each neuron corresponds to the predicted probability of a signal transmission error type; a softmax function is used as an activation function; the transmission error diagnosis model is a multi-layer perceptron MLP model; Multi-classification cross entropy is used as the loss function of the model to measure the difference between the predicted value and the actual value of the model; the multi-classification cross entropy loss function is: ;in, is the average loss of the dataset; is the total number of samples in the data set; is the number of signal transmission error types; For the The samples for The true labels of each type; The model predicts The samples belong to The probability of a type; The transmission error diagnosis model is trained using the training set, and the model parameters are updated through the back propagation algorithm to minimize the loss function. The performance of the transmission error diagnosis model is evaluated by calculating the accuracy index using the validation set. The SGD optimization algorithm is selected as the optimizer, and the model is tuned according to the performance feedback of the validation set. The model parameters are adjusted until the performance is no longer significantly improved or the preset stopping condition is reached. The test set is used to evaluate the performance of the model in the prediction task, and the trained transmission error diagnosis model is used to predict the current signal transmission error data to obtain the signal transmission error type.

[0028] The chip signal transmission monitoring terminal issues an error processing instruction according to the signal transmission error type, and the method for generating a corresponding error optimization strategy includes: The error severity analysis of the signal transmission error type is performed using an error severity evaluation model, and the error severity evaluation model is: ;in, is the severity of the error; is the signal packet loss rate during chip transmission; is the signal retransmission rate; is the bit error rate during signal transmission; is the signal loss rate; is the packet arrival sequence error rate; For example, the signal packet loss rate during chip transmission is 0.2, signal retransmission rate is 0.8, the bit error rate during signal transmission is 0.6, signal loss rate is 0.9, and the packet arrival order error rate is 0.1, then the error severity is about 1.9; Preset error severity threshold, further divide the preset error severity threshold into low error severity threshold and high error severity threshold ; The severity of the error The lower error severity threshold and high error severity threshold For comparison, if , then the signal transmission error type is determined to be a low error type; if , then the signal transmission error type is determined to be a moderate error type; if , then the signal transmission error type is determined to be a height error type; For low error types, the chip signal transmission monitoring terminal issues a third-level error processing instruction to generate a mild error optimization strategy; for moderate error types, the chip signal transmission monitoring terminal issues a second-level error processing instruction to generate a moderate error optimization strategy; for high error types, the chip signal transmission monitoring terminal issues a first-level error processing instruction to generate a high error optimization strategy.

[0029] The mild error optimization strategy includes adjusting the signal buffer settings, optimizing the signal encoding settings and monitoring and adjusting the bus load; the moderate error optimization strategy includes enhancing the signal retransmission mechanism, adjusting the signal frequency and implementing error correction codes; the severe error optimization strategy includes increasing the number of signal data packet retransmissions, increasing the signal buffer capacity and switching the signal transmission path.

[0030] The error threshold generated by the preset monitoring video signal during the transmission inside the chip is set by the staff, and the different errors generated by the monitoring video signal during the transmission inside the chip are collected by the chip signal transmission monitoring terminal, and the average value of the errors generated by multiple monitoring video signals during the transmission inside the chip is taken as the preset error threshold generated by the monitoring video signal during the transmission inside the chip.

[0031] In this embodiment, the conversion quality of video frames at different resolutions is ensured through the resampling formula and the resolution scaling factor restriction, thereby avoiding image blur or distortion caused by resolution mismatch; the scaling factor restriction formula ensures that the image will not be distorted due to excessive enlargement or reduction during the resolution conversion process; the most suitable scaling factor is dynamically calculated according to the size ratio of the original image and the standard image, thereby maintaining the clarity and details of the image; the linear interpolation formula and the interpolation factor adjustment model make the insertion of frames smoother and more uniform, thereby improving the fluency of the video; at the same time, by discarding frames to reduce the frame rate, the amount of data can be reduced while maintaining the consistency of the video content; the color depth of the video frames is unified, the redundancy of color information is reduced, and the data processing efficiency and accuracy are improved; The Laplace operator is used to calculate the sharpness of the video image, which provides a quantitative indicator for the evaluation of video clarity and helps to detect video quality problems in a timely manner. By smoothing the chip bandwidth, analyzing the chip transmission delay, packet loss rate and bus congestion, as well as the bus network throughput, the internal state of the chip is fully monitored, providing data support for chip optimization and troubleshooting. Periodic and multi-scale analysis of the temperature and electromagnetic interference level of the transmission environment helps to identify potential environmental interference factors and improve the stability and reliability of data transmission. FFmpeg is used to split the video into time windows and summarize the features in each window, which reduces the complexity and time cost of data processing. The weight factors and characteristic functions in the window length adjustment model can be adjusted and optimized according to the needs of specific application scenarios; the model can adapt to different data characteristics and analysis objectives, and provide support for time series data analysis in various complex scenarios; by adjusting the smoothing factor, the weights of historical data and current data in the smoothed value can be balanced. When the chip bandwidth fluctuates greatly, the value of the smoothing factor can be appropriately increased to consider more historical data, thereby reducing the impact of fluctuations on the smoothed value; by adjusting the weight factor of the bus throughput, the system can more accurately evaluate the actual load of the transmission path in different time periods, thereby making more reasonable resource allocation decisions.

[0032] Example 2, please refer to Figure 2 As shown, the present embodiment provides a chip adaptive signal transmission path optimization method, including: S1, collect monitoring video data, chip internal parameter data and chip transmission environment data; S2, preprocessing the acquired monitoring video data, chip internal parameter data and chip transmission environment data to obtain a video feature data set, a chip internal parameter feature data set and a chip transmission environment feature data set; analyzing the impact of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set to obtain a transmission impact feature data set; S3, training a transmission error prediction model based on the transmission impact feature data set, and obtaining the error generated by the monitoring video signal during the transmission process inside the chip based on the transmission error prediction model; S4, comparing the predicted error generated by the monitoring video signal during the transmission process inside the chip with the preset error threshold of the monitoring video signal during the transmission process inside the chip to determine whether signal transmission optimization is needed; S5. If signal transmission optimization is not required, the chip signal transmission monitoring terminal issues a transmission instruction and continues to transmit the monitoring video signal; if signal transmission optimization is required, the chip signal transmission monitoring terminal issues an optimization instruction and collects signal transmission error data;

[0033] S6. A transmission error diagnosis model is obtained according to signal transmission error data training, and a signal transmission error type is obtained based on the transmission error diagnosis model; the chip signal transmission monitoring terminal issues an error processing instruction according to the signal transmission error type and generates a corresponding error optimization strategy.

[0034] Since the electronic device introduced in this embodiment is an electronic device used to implement the adaptive signal transmission path optimization system based on a chip in the embodiment of the present application, based on the adaptive signal transmission path optimization system of a chip introduced in the embodiment of the present application, the technical personnel of the field can understand the specific implementation of the electronic device of the present embodiment and its various variations, so how the electronic device implements the method in the embodiment of the present application is not described in detail here. As long as the technical personnel of the field implement the electronic device used by the adaptive signal transmission path optimization system of a chip in the embodiment of the present application, it belongs to the scope of protection of this application.

[0035] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0036] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technical users in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A chip adaptive signal transmission path optimization system, characterized in that: include: Data acquisition module, used to collect monitoring video data, chip internal parameter data and chip transmission environment data; The data processing module is used to pre-process the acquired monitoring video data, chip internal parameter data and chip transmission environment data to obtain a video feature data set, a chip internal parameter feature data set and a chip transmission environment feature data set; analyze the impact of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set to obtain a transmission impact feature data set; A transmission error analysis module is used to obtain a transmission error prediction model based on the transmission impact feature data set training, and obtain the error generated by the monitoring video signal during the transmission process inside the chip based on the transmission error prediction model; The transmission optimization evaluation module is used to compare the error generated by the predicted monitoring video signal during the transmission process inside the chip with the preset error threshold of the monitoring video signal during the transmission process inside the chip to determine whether signal transmission optimization is needed; Error type analysis module: if signal transmission optimization is not required, the chip signal transmission monitoring terminal issues a transmission instruction and continues to transmit the monitoring video signal; if signal transmission optimization is required, the chip signal transmission monitoring terminal issues an optimization instruction and collects signal transmission error data; A signal transmission optimization module, used to obtain a transmission error diagnosis model according to signal transmission error data training, and obtain a signal transmission error type based on the transmission error diagnosis model; The chip signal transmission monitoring terminal issues error processing instructions according to the signal transmission error type and generates corresponding error optimization strategies.

2. The chip adaptive signal transmission path optimization system according to claim 1, characterized in that: The monitoring video data includes video resolution, video frame rate, video clarity, color depth and video duration; The chip internal parameter data includes chip bandwidth, chip transmission delay, packet loss rate, bus congestion level and bus throughput; The chip transmission environment data includes the temperature of the transmission environment, the electromagnetic interference level and the transmission path load rate.

3. The adaptive signal transmission path optimization system of a chip according to claim 2, characterized in that: The method of preprocessing the acquired monitoring video data, chip internal parameter data and chip transmission environment data to obtain a video feature data set, a chip internal parameter feature data set and a chip transmission environment feature data set includes: S31. For surveillance video data: adjusting the video resolution at the pixel level by a resampling formula, converting each frame of image from the original resolution to the standard resolution; the resampling formula is: ;in, For the adjusted image in position The pixel value of For the original image at position The pixel value of The resolution scaling factor for converting each frame from the original resolution to the standard resolution; The resolution scaling factor for converting each frame image from the original resolution to the standard resolution is calculated using the scaling factor limiting formula. impose restrictions; The scaling factor limiting formula is: ;in, is the width of the original image; is the height of the original image; is the width of the standard image; is the height of the standard image; is the number of frames of the video; There are two adjacent frames by default and For each pair of adjacent frames, a linear interpolation formula is used to calculate and insert a and New frames between ; The linear interpolation formula is: ;in, For the Frame at position The pixel value of For the Frame at position The pixel value of is the interpolation factor; Adjust the model by interpolation factor Adjustment is performed, and the interpolation factor adjustment model is: ;in, For two adjacent frames and The number of new frames to insert between is the sequence number of the newly inserted frame; By adjusting , the position ratio of each inserted frame can be dynamically defined, so that the distribution of frames is uniform and smooth; For a frame rate The frame rate of the video is reduced by dropping frames. The preset video is reduced to Frame rate, calculate the required frame retention interval, the calculation formula is: ;in, Reserve intervals for required frames; is the floor rounding function; Use OpenCV to extract images from the video frame by frame, and calculate the sharpness of the video image for each frame using the Laplace operator formula, which reflects the video clarity. The Laplace operator formula is: ;in, For a single pixel at position The output value of the Laplace operator at is the sharpness of the video image; For images at position The pixel value of For the image The second derivative in the axis direction; For the image The second derivative in the axis direction; The different original image color depths in the video frame are uniformly converted to a standard color depth through a color quantization formula; the color quantization formula is: ;in, The quantized image is at position The pixel value of is the color depth of the original image in the video frame; is the standard color depth; The entire security surveillance video is divided into v time windows through the FFmpeg multimedia processing tool. For each time window, the time points within the time window are calculated. The video feature values ​​on the image are obtained, and the features of each time window are summarized by the time window feature summary formula; the time window feature summary formula is: ;in, It is the time window feature summary value of the entire security surveillance video; For time point The characteristic value of is the length of the time window; are consecutive time points; Adjust the length of the time window for the model by using the window length Adjustment is performed, and the window length adjustment model is: ;in, For the A weight factor related to the time window feature; For the The number of characteristics; For time point Related characteristic functions; Types of characteristics; is the index of the feature type; An adjustment factor for adjusting the length of the time window in the window length adjustment model; Time The relevant characteristic function expression is: ;in, For time point The characteristic value of is the time interval between time points; S32. For chip internal parameter data: smooth the chip bandwidth using an exponential smoothing formula; The exponential smoothing formula is: ;in, For time The smoothing bandwidth value at the moment; For time The smoothing bandwidth value at the moment; For time The original bandwidth value at the time; is the smoothing factor; The smoothing factor is adjusted by a smoothing factor adjustment formula, and the smoothing factor adjustment formula is: ;in, For in time The number of smoothing bandwidth data points within ; The time required to smooth the chip bandwidth; An adjustment coefficient for controlling the influence of the time required for smoothing chip bandwidth on the smoothing factor; For chip transmission delay, quantile analysis is used to identify abnormal delays; for packet loss rate, weighted average method is used to Summarize the packet loss in each time period and find The packet loss rate in each time period; for the bus congestion level, the sliding window method is used to calculate The average bus congestion level in a time period; For bus throughput, use the cumulative throughput formula to calculate arrive The cumulative bus throughput in the time period; the cumulative throughput formula is: ;in, For arrive The total throughput of the bus during the time period; The instantaneous throughput at a given moment; For the moment The weight factor affecting bus throughput; The weight factor of the bus throughput is adjusted by a weight factor adjustment model, and the weight factor adjustment model is: ;in, For the moment The weight factor affecting bus throughput; is the total number of moments in the time period; is the standard deviation of the bus throughput during the time period; S33. For chip transmission environment data: For the temperature of the transmission environment, Fourier transform is used to identify the periodic changes in the temperature of the transmission environment; for the electromagnetic interference level, a wavelet transform formula is used to perform multi-scale analysis on the electromagnetic interference level to identify interference characteristics on different time scales; the wavelet transform formula is: ;in, For scale and location The wavelet transform result under ; For each time point The level of electromagnetic interference on is the mother wavelet function; For the transmission path load rate, the transmission path load rate characteristic value in each time window segmented by the FFmpeg multimedia processing tool is calculated. The calculation formula is: ;in, is the transmission path load rate characteristic summary value; For time point The transmission path load factor characteristic value; The LOF algorithm is used on the processed monitoring video data, chip internal parameter data and chip transmission environment data to identify the abnormal values ​​in the monitoring video data, chip internal parameter data and chip transmission environment data, and the identified abnormal values ​​are removed to obtain the processed video feature data set, chip internal parameter feature data set and chip transmission environment feature data set; The acquired video feature dataset, chip internal parameter feature dataset and chip transmission environment feature dataset are subjected to standard deviation normalization and converted into standard normal distribution, and finally the normalized video feature dataset, chip internal parameter feature dataset and chip transmission environment feature dataset are obtained.

4. The chip adaptive signal transmission path optimization system according to claim 3, characterized in that: The method of analyzing the influence of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set to obtain the transmission influence feature data set includes: A joint regression model is constructed to quantitatively analyze the impact of the chip transmission environment feature data set on the video feature data set and the chip internal parameter feature data set, and a transmission impact feature data set is obtained; the joint regression model is: ;in, To transmit comprehensive features of the feature dataset; For the Each chip transmits environmental characteristics; An index of environmental characteristics transmitted to the chip; The number of environmental characteristics transmitted to the chip; For the Video features; is the index of video features; is the number of video features; For the Internal parameter characteristics of each chip; is the index of the internal parameter characteristics of the chip; is the number of internal parameter features of the chip; is the regression coefficient of chip transmission environment characteristics; is the regression coefficient of the video feature; is the regression coefficient of the chip internal parameter characteristics; is the regression coefficient of the interaction term between the video feature and the chip internal parameter feature; is the error term of the model; is the constant term of the joint regression model.

5. The chip adaptive signal transmission path optimization system according to claim 4, characterized in that: The training method of the transmission error prediction model includes: The data set is divided into a training set, a validation set, and a test set to train the model and evaluate the model performance; the sample set is a subset of the data set, and each sample set includes a historical transmission impact feature data set and the errors generated by the corresponding monitoring video signal during the transmission process inside the chip; a transmission error prediction model is constructed, including an input layer, a GRU layer, a fully connected layer, and an output layer; The input layer of the transmission error prediction model is used to input the historical transmission impact feature data set. The number of neurons in the input layer should match the number of features in the historical transmission impact feature data set. The GRU layer is used to process the historical comprehensive feature data set. The number of GRU layers and the number of neurons are adjusted according to the complexity of the task. The fully connected layer is used to provide additional nonlinear transformation. The output layer of the model is used to output the error generated by the monitoring video signal during the transmission process inside the chip. The number of neurons in the output layer corresponds to the number of prediction targets. The prediction value is output through one neuron, and the identity function is used as the activation function. The transmission error prediction model is a gated recurrent unit model. Define the loss function of the model, and use the L2 regularized mean square error loss function to measure the difference between the error generated during the predicted security surveillance video transmission process and the error generated during the actual security surveillance video transmission process; The L2 regularized mean square error loss function is: ;in, is the sample size; For the The predicted value of samples; For the The actual value of the samples, is the regularization coefficient; is the number of training samples, is the sum of squares of all trainable weights; The transmission error prediction model is trained using the training set, and the model parameters are updated through the back propagation algorithm to minimize the loss function. The performance of the transmission error prediction model is evaluated by calculating the accuracy index using the validation set. The Adam optimization algorithm is used as the optimizer. The model is tuned according to the performance feedback of the validation set. The model parameters are adjusted until the performance is no longer significantly improved or the preset number of iterations is reached. The test set is used to evaluate the performance of the model in the prediction task, and the trained transmission error prediction model is used to predict the current transmission impact feature data set to obtain the error generated by the monitoring video signal during the transmission process inside the chip.

6. The chip adaptive signal transmission path optimization system according to claim 5, characterized in that: The method of comparing the predicted error generated by the monitoring video signal during the internal transmission of the chip with the preset error threshold generated by the monitoring video signal during the internal transmission of the chip to determine whether signal transmission optimization is required includes: If the error generated by the predicted monitoring video signal during the transmission process inside the chip is less than the preset error threshold of the monitoring video signal during the transmission process inside the chip, it is determined that signal transmission optimization is not required; If the error generated by the predicted monitoring video signal during the transmission inside the chip is greater than or equal to the error threshold generated by the monitoring video signal during the transmission inside the chip, it is determined that signal transmission optimization is required.

7. The chip adaptive signal transmission path optimization system according to claim 6, characterized in that: The signal transmission error data includes the signal packet loss rate, signal retransmission rate, bit error rate, signal loss rate and data packet arrival sequence error rate during chip transmission.

8. The chip adaptive signal transmission path optimization system according to claim 7, characterized in that: The training method of the transmission error diagnosis model includes: The data set is divided into a training set, a validation set, and a test set to train the model and evaluate the model performance; the sample set is a subset of the data set, and each sample set includes historical signal transmission error data and the corresponding signal transmission error type; Construct a transmission error diagnosis model, which includes an input layer, a hidden layer and an output layer; the input layer is used to input historical signal transmission error data, and the output layer is used to output the signal transmission error type; the output layer is provided with neurons equal to the number of signal transmission error types, and each neuron corresponds to the predicted probability of a signal transmission error type; a softmax function is used as an activation function; the transmission error diagnosis model is a multi-layer perceptron MLP model; Multi-classification cross entropy is used as the loss function of the model to measure the difference between the predicted value and the actual value of the model; the multi-classification cross entropy loss function is: ;in, is the average loss of the dataset; is the total number of samples in the data set; is the number of signal transmission error types; For the The samples for The true labels of each type; The model predicts The samples belong to The probability of a type; The transmission error diagnosis model is trained using the training set, and the model parameters are updated through the back propagation algorithm to minimize the loss function. The performance of the transmission error diagnosis model is evaluated by calculating the accuracy index using the validation set. The SGD optimization algorithm is selected as the optimizer, and the model is tuned according to the performance feedback of the validation set. The model parameters are adjusted until the performance is no longer significantly improved or the preset stopping condition is reached. The test set is used to evaluate the performance of the model in the prediction task, and the trained transmission error diagnosis model is used to predict the current signal transmission error data to obtain the signal transmission error type.

9. The chip adaptive signal transmission path optimization system according to claim 8, characterized in that: The chip signal transmission monitoring terminal issues an error processing instruction according to the signal transmission error type, and the method for generating a corresponding error optimization strategy includes: The error severity analysis of the signal transmission error type is performed using an error severity evaluation model, and the error severity evaluation model is: ;in, is the severity of the error; is the signal packet loss rate during chip transmission; is the signal retransmission rate; is the bit error rate during signal transmission; is the signal loss rate; is the packet arrival sequence error rate; Preset error severity threshold, further divide the preset error severity threshold into low error severity threshold and high error severity threshold ; The severity of the error The lower error severity threshold and high error severity threshold For comparison, if , then the signal transmission error type is determined to be a low error type; if , then the signal transmission error type is determined to be a moderate error type; if , then the signal transmission error type is determined to be a height error type; For low error types, the chip signal transmission monitoring terminal issues a third-level error processing instruction to generate a mild error optimization strategy; for moderate error types, the chip signal transmission monitoring terminal issues a second-level error processing instruction to generate a moderate error optimization strategy; for high error types, the chip signal transmission monitoring terminal issues a first-level error processing instruction to generate a high error optimization strategy.

10. The chip adaptive signal transmission path optimization system according to claim 9, characterized in that: The mild error optimization strategy includes adjusting the signal buffer settings, optimizing the signal encoding settings and monitoring and adjusting the bus load; the moderate error optimization strategy includes enhancing the signal retransmission mechanism, adjusting the signal frequency and implementing error correction codes; The height error optimization strategy includes increasing the number of signal data packet retransmissions, increasing the signal buffer capacity, and switching the signal transmission path.

Images