An unmanned system information elasticity adaptation method based on multi-granularity characterization

Abstract

The application discloses a kind of unmanned system information flexible adaptation transmission methods based on multi-granularity characterization, belong to communication transmission field.The method is first to evaluate communication quality, then constructs multi-stage compression code rate information content granularity division level, the value range of communication quality is evenly divided into four segments, each interval corresponds to a content granularity level;Finally, according to content granularity level, the transmission mode is flexibly adapted, and the corresponding transmission mode is selected to transmit unmanned system information.The application constructs two communication quality evaluation modes, and the mean of the communication quality evaluation results obtained by the two modes is used as the communication quality evaluation result at the next time.The content granularity level is directly determined based on the communication quality evaluation level, and a communication content flexible transmission mechanism is constructed, which automatically selects the optimal transmission data according to the state of the unmanned aerial vehicle link.

Description

Technical Field

[0001] This invention relates to the field of communication transmission, and in particular to a method for flexible adaptation and transmission of information in unmanned systems based on multi-granularity characterization, which solves the problem of efficient information transmission in complex environments. Background Technology

[0002] Information elasticity adaptation of unmanned systems based on multi-granularity representation can improve the adaptability of unmanned systems to complex environments, optimize resource allocation, enhance decision-making accuracy, and strengthen system robustness. These advantages make multi-granularity representation an important research direction in the field of unmanned system information processing.

[0003] From a demand perspective:

[0004] (1) Unmanned systems often need to operate in complex and ever-changing environments, which place high demands on their information processing capabilities. Multi-granularity representation enables unmanned systems to adjust the fineness and depth of information processing according to different environments and task requirements, thereby better adapting to complex environments.

[0005] (2) Unmanned systems have limited resources (such as computing resources and communication resources). How to achieve efficient information processing with limited resources is a key issue. Multi-granularity representation can flexibly adjust the granularity of information processing according to task requirements, thereby achieving efficient resource utilization while ensuring the quality of task completion.

[0006] (3) In certain emergency situations, unmanned systems need to respond quickly and make decisions. Multi-granularity representation enables unmanned systems to process information at different time scales, thereby achieving rapid response to emergency situations.

[0007] From the perspective of advantages:

[0008] (1) Multi-granularity characterization allows unmanned systems to flexibly adjust the granularity of information processing according to mission requirements and environmental changes. This flexibility enables unmanned systems to better adapt to complex and ever-changing environments and mission requirements;

[0009] (2) Through multi-granularity characterization, unmanned systems can rationally allocate resources based on the importance and urgency of tasks. For important and urgent tasks, more resources can be invested in fine-grained processing; while for general tasks, coarser granularity can be used to save resources.

[0010] (3) Multi-granularity characterization enables unmanned systems to perform decision analysis at different time scales and information granularities. This multi-faceted analysis helps to discover potential problems and risks, thereby improving the accuracy and reliability of decision-making;

[0011] (4) In complex and ever-changing environments, unmanned systems may face various unpredictable challenges. Multi-granularity representation can enable unmanned systems to perceive and respond to these challenges at different information granularities, thereby enhancing the robustness and adaptability of the system.

[0012] Currently, there is a lack of flexible adaptation and transmission schemes for unmanned system information based on multi-granularity characterization in existing technologies. Summary of the Invention

[0013] In view of this, the present invention proposes an information elastic adaptation and transmission method for unmanned systems based on multi-granularity characterization to solve the problem of efficient information transmission in complex environments.

[0014] The technical solution adopted in this invention is as follows:

[0015] A method for elastically adapting and transmitting information in unmanned systems based on multi-granularity representation includes the following steps:

[0016] Step 1: Assess the communication quality;

[0017] Step 2: Construct a multi-level content granularity hierarchy for compressed bitrate information;

[0018] Step 3: Divide the range of communication quality values ​​into four segments, with each segment corresponding to a content granularity level.

[0019] Step 4: Flexibly adapt the transmission method according to the content granularity level, and select the appropriate transmission method to transmit unmanned system information.

[0020] Furthermore, the specific method of step 1 is as follows:

[0021] Step 101: Construct a communication quality level assessment model; the communication quality level assessment model is an SVMML model, using a Gaussian kernel function;

[0022] Step 102: Obtain the communication quality parameters of the data link for the first n time steps. The communication quality parameters include signal strength, bit error rate, and signal-to-noise ratio.

[0023] Step 103: Input the signal strength, bit error rate, and signal-to-noise ratio of the previous n time steps into the corresponding LSTM model, and output the signal strength, bit error rate, and signal-to-noise ratio of the next time step to obtain the communication quality parameters of the data link at the next time step.

[0024] Step 104: Input the communication quality parameters of the data link at the next time moment into the communication quality level evaluation model to obtain the communication quality evaluation result Eq1 at the next time moment;

[0025] Step 105: Input the communication quality parameters of each of the first n time steps into the communication quality level evaluation model to obtain the communication quality evaluation results of each of the first n time steps. Input the communication quality evaluation results of each of the first n time steps into the corresponding LSTM model to obtain the communication quality evaluation result Eq2 of the next time step.

[0026] Step 106: If the difference between Eq1 and Eq2 is less than the threshold, then Otherwise, at the next time step, steps 102-106 are repeated until the difference between Eq1 and Eq2 is less than the threshold, and the final communication quality assessment result Eq is obtained.

[0027] Furthermore, the specific method for step 2 is as follows:

[0028] The transmitted information is divided into four levels of granularity of compressed bitrate information content: original video stream, compressed video stream, stream containing target image, and message information.

[0029] Furthermore, step 4 is specifically implemented as follows:

[0030] If the content granularity level is the original video stream, then the unmanned system information is transmitted directly as a video stream.

[0031] If the content granularity level is compressed video stream, the video stream collected by the unmanned system is first compressed in H265 encoding format before transmission.

[0032] If the content granularity level is to include target images, then target recognition is performed on the video stream acquired by the unmanned system, and then the image frames of the identified targets are transmitted.

[0033] If the content granularity level is message information, then target identification is performed on the video stream collected by the unmanned system, and then the category and location of the identified target are transmitted in message form.

[0034] Compared with the prior art, the present invention has the following advantages:

[0035] 1. This invention constructs two communication quality assessment modes: The first mode is based on a regression prediction model to predict the communication quality parameters at the next time step, and then obtains the communication quality assessment result at the next time step based on the communication quality assessment model; the second mode is based on the communication quality assessment model and the communication quality parameters at the previous n time steps to obtain the communication quality assessment results at the previous n time steps, and then inputs them into the regression prediction model to obtain the communication quality assessment result at the next time step. If the communication quality assessment results of the two modes are close, the average of the communication quality assessment results obtained by the two modes is taken as the communication quality assessment result at the next time step; otherwise, the calculation continues until the communication quality assessment results of the two modes are close or the same.

[0036] 2. This invention constructs a method for determining communication quality assessment levels, which directly determines the content granularity level.

[0037] 3. This invention constructs a flexible transmission mechanism for communication content: when the communication quality is good, video streams are transmitted; when the communication quality is poor, images containing the target are transmitted; and when the communication quality is worst, communication messages are transmitted, thus meeting the minimum communication quality requirements. Detailed Implementation

[0038] The following will clearly and completely describe the concept, technical advantages, and resulting technical effects of the present invention with reference to embodiments, so as to fully understand the purpose, features, and effects of the present invention. It should be noted that the specific embodiments described herein are only used to explain the present invention and do not limit the present invention.

[0039] A method for elastically adapting and transmitting information in unmanned systems based on multi-granularity representation, the method comprising:

[0040] S1: Evaluate the communication quality using the following method:

[0041] Obtain the data link communication quality parameters for the first n time steps, including: signal strength, bit error rate, signal-to-noise ratio, etc.

[0042] Based on the LSTM model, the communication quality parameters for the next time step are obtained, including signal strength, bit error rate, and signal-to-noise ratio. The specific operation method is as follows:

[0043] ① Input the signal strength of the previous n time steps into the LSTM model, and output the signal strength of the next time step;

[0044] ②The bit error rate and signal-to-noise ratio, etc., can also be obtained by analogy with the above method to obtain the bit error rate, signal-to-noise ratio, etc. at the next time step;

[0045] To construct a communication quality level assessment model, the specific method is as follows: input the communication quality parameters at the same time into the SVMML model to obtain the communication quality assessment results;

[0046] Based on the communication quality level assessment model, the communication quality assessment result Eq is obtained in the following way:

[0047] ① Input the communication quality parameters of the first n time steps into the LSTM model to obtain the communication quality parameters of the next time step, and input them into the SVMML model to obtain the communication quality evaluation result Eq1 of the next time step;

[0048] ② Input the communication quality parameters of each of the first n time steps into the SVMML model to obtain the communication quality evaluation result of each time step. Input the communication quality evaluation results of the first n time steps into the LSTM model to obtain the communication quality evaluation result Eq2 of the next time step.

[0049] ③If the difference between Eq1 and Eq2 is less than a given threshold, then Otherwise, continue executing the first two steps at the next time step until the difference between Eq1 and Eq2 is less than a given threshold.

[0050] S2: Constructing multi-level compression bitrate information content granularity classification levels:

[0051] The transmitted information is divided into content granularity levels such as raw video stream, compressed video stream, stream containing target image, and message information.

[0052] S3: Divide the maximum and minimum values ​​of the theoretical values ​​of the communication quality assessment results into 4 levels, each corresponding to a content granularity level;

[0053] Step 4: Flexibly adapt the transmission method according to the content granularity level, and select the appropriate transmission method to transmit unmanned system information:

[0054] If the content granularity level is the original video stream, then the unmanned system information is transmitted directly as a video stream.

[0055] If the content granularity level is compressed video stream, the video stream collected by the unmanned system is first compressed in H265 encoding format before transmission.

[0056] If the content granularity level is to include target images, then target recognition is performed on the video stream acquired by the unmanned system, and then the image frames of the identified targets are transmitted.

[0057] If the content granularity level is message information, then target identification is performed on the video stream collected by the unmanned system, and then the category and location of the identified target are transmitted in message form.

[0058] This invention constructs two communication quality assessment modes, takes the average of the communication quality assessment results obtained from the two modes as the communication quality assessment result at the next moment, directly determines the content granularity level based on the communication quality assessment level, and constructs a flexible transmission mechanism for communication content, which automatically selects the best data transmission based on the UAV link status.

Claims

1. A method for information resilient transmission of unmanned systems based on multi-granularity characterization, characterized in that, The method comprises the following steps: Step 1: evaluating the communication quality; the specific method is: Step 101: constructing a communication quality level evaluation model; the communication quality level evaluation model is an SVMML model, and a Gaussian kernel function is adopted; Step 102, obtain the previous The communication quality parameters of the data link at each moment, including signal strength, bit error rate, and signal-to-noise ratio; Step 103, the front The signal strength, bit error rate, and signal-to-noise ratio at each time step are input into the corresponding LSTM model, and the signal strength, bit error rate, and signal-to-noise ratio at the next time step are output, thereby obtaining the communication quality parameters of the data link at the next time step. Step 104, input the communication quality parameter of the next time data link into the communication quality level evaluation model, and obtain the communication quality evaluation result of the next time ; Step 105, input the communication quality parameter of each of the previous time instants into the communication quality level evaluation model respectively to obtain the communication quality evaluation result of each of the previous time instants, input the communication quality evaluation result of each of the previous time instants into the corresponding LSTM model to obtain the communication quality evaluation result of the next time instant ; Step 106, if the difference between is less than a threshold, then , otherwise, after the next time, repeat the execution of steps 102-106 until the difference between is less than a threshold, obtaining the final communication quality evaluation result ; Step 2: constructing a multi-level compressed code rate information content granularity level; the specific method is: The transmission information is divided into four compressed code rate information content granularity levels: original video code stream, compressed video code stream, target image containing, and message information; Step 3: evenly dividing the value range of the communication quality into four segments, and each segment corresponds to a content granularity level; Step 4: elastically adapting the transmission mode according to the content granularity level, and selecting a corresponding transmission mode to transmit the unmanned system information; the specific method is: If the content granularity level is original video code stream, the unmanned system information is directly transmitted in the form of video code stream; If the content granularity level is compressed video code stream, the video code stream collected by the unmanned system is compressed in the H265 coding format, and then transmitted; If the content granularity level is target image containing, the video code stream collected by the unmanned system is subjected to target recognition, and then the image frame of the recognized target is transmitted; If the content granularity level is message information, the video code stream collected by the unmanned system is subjected to target recognition, and then the category and position of the recognized target are transmitted in the form of a message.

Images