CN-122001952-A - Unmanned system distributed data transmission method based on fuzzy confidence and hysteresis trigger

CN122001952ACN 122001952 ACN122001952 ACN 122001952ACN-122001952-A

Abstract

The application relates to a distributed data transmission method of an unmanned system based on fuzzy confidence and hysteresis trigger, which is characterized in that uncertainty feature vectors and task value feature vectors corresponding to the unmanned system are extracted in parallel through target observation information acquired by any unmanned system at the current moment; the method comprises the steps of carrying out double-layer fuzzy reasoning on an uncertainty feature vector and a task value feature vector, outputting comprehensive confidence coefficient corresponding to an unmanned system at the current moment, calculating a high trigger threshold and a low silence threshold based on the comprehensive confidence coefficient of the unmanned system at the past moment, updating a data transmission state of the unmanned system at the current moment according to a hysteresis comparison result of the comprehensive confidence coefficient corresponding to the unmanned system at the current moment and the high trigger threshold or the low silence threshold, and executing a differential data transmission strategy according to the corresponding comprehensive confidence coefficient when the updated data transmission state is active by the unmanned system, so that reliable and real-time transmission of high-value low-quality data under a low-bandwidth dynamic environment is realized.

Inventors

XU XUESONG
LI YIGAN
PENG HAN
LIU SIQI
WANG WANTAO
MA AO
XIA JING

Assignees

湖南工商大学

Dates

Publication Date: 20260508
Application Date: 20260408

Claims (10)

1. The unmanned system distributed data transmission method based on fuzzy confidence and hysteresis trigger is characterized by comprising the following steps: s1, for target observation information acquired by any unmanned system at the current moment, extracting an uncertainty feature vector and a task value feature vector corresponding to the unmanned system in parallel; s2, calculating a high trigger threshold and a low silence threshold based on the comprehensive confidence coefficient of the unmanned system at the past moment, and updating the data transmission state of the unmanned system at the current moment according to the hysteresis comparison result of the comprehensive confidence coefficient corresponding to the unmanned system at the current moment and the high trigger threshold or the low silence threshold; and S3, when the updated data transmission state is active, the unmanned system executes a differential data transmission strategy according to the corresponding comprehensive confidence.
2. The unmanned system distributed data transmission method of claim 1, wherein the target observation information comprises an image, a point cloud, or a radio frequency signal.
3. The unmanned system distributed data transmission method of claim 1, wherein extracting the uncertainty feature vector of the unmanned system comprises: Calculating the structural similarity index of the target observation information to obtain a quality index; Calculating the shielding ratio of the target observation information through a segmentation algorithm; estimating the motion ambiguity corresponding to the unmanned system based on the inertial measurement unit data; And calculating the signal-to-noise ratio of the target observation information.
4. A method of unmanned system distributed data transmission according to claim 3, wherein extracting the task value feature vector of the unmanned system comprises: the unmanned system acquires a task and calculates a normalized distance between the unmanned system and a target point in the task; inquiring the task urgency of the acquired task; a time-dependent decay factor based on the discovery period is calculated.
5. The unmanned system distributed data transmission method of claim 3, wherein the subjecting the uncertainty feature vector and the task value feature vector to double-layer fuzzy reasoning comprises: blurring each element in the task value feature vector into a corresponding language value through a membership function, and mapping each language value corresponding to the task value feature vector into a rule modulation factor according to a preset mapping table; And mapping the uncertainty characteristic vector into the comprehensive confidence corresponding to the unmanned system at the current moment based on the rule modulation factor and a preset fuzzy mapping rule.
6. The unmanned system distributed data transmission method of claim 5, wherein the preset fuzzy mapping rule is: And if the quality index is a first preset value, the shielding ratio is a second preset value, the motion blur degree is a third preset value and the signal to noise ratio is a fourth preset value, taking the value calculated by the rule modulation factor as the comprehensive confidence degree.
7. The unmanned system distributed data transmission method of claim 1, wherein calculating the high trigger threshold and the low silence threshold based on the integrated confidence of the unmanned system at the past time comprises: setting a confidence coefficient sliding window with the length of W, wherein the confidence coefficient sliding window carries out sliding sampling on the comprehensive confidence coefficient corresponding to the past continuous moment of the unmanned system according to a preset step length; Calculating the mean value and standard deviation of each comprehensive confidence coefficient in the confidence coefficient sliding window corresponding to the current moment; And calculating a high trigger threshold and a low silence threshold based on the mean value and the standard deviation, wherein a calculation formula comprises: ; ; Wherein, the Indicating a high trigger threshold value and, Indicating a low silence threshold value, The mean value of the integrated confidence level is represented, The standard deviation of the integrated confidence level is represented, A first positive coefficient is represented and a second positive coefficient is represented, A second positive coefficient is represented and is used to represent, A third positive coefficient is represented and is used to represent, Representing the load factor of the network at the current time.
8. The method for distributed data transmission in an unmanned system according to claim 7, wherein the load factor of the network at the present time comprises a channel occupancy rate obtained by the unmanned system at the present time based on channel interception.
9. The unmanned system distributed data transmission method of claim 7, wherein updating the data transmission state of the unmanned system at the current time comprises: Acquiring the data transmission state of the unmanned system at the current moment, and triggering data transmission and updating the data transmission state of the unmanned system at the current moment to be active if the comprehensive confidence coefficient corresponding to the unmanned system at the current moment is larger than the high trigger threshold and the data transmission state of the unmanned system at the current moment is silent; If the comprehensive confidence coefficient corresponding to the unmanned system at the current moment is smaller than the low silence threshold, and the data transmission state of the unmanned system at the current moment is active, stopping data transmission and updating the data transmission state of the unmanned system at the current moment into silence; and the rest conditions maintain the data transmission state of the unmanned system at the current moment.
10. The unmanned system distributed data transmission method of claim 4, wherein S3 comprises: When the updated data transmission state is active, if the comprehensive confidence coefficient corresponding to the unmanned system is greater than a first interval threshold value, compressing target observation information acquired by the unmanned system into a first data packet by adopting a lightweight video encoder; If the comprehensive confidence coefficient corresponding to the unmanned system is larger than the second interval threshold value and smaller than or equal to the first interval threshold value, a super-resolution network is called to enhance the low-quality area in the target observation information in real time according to the shielding ratio and the motion ambiguity, the enhanced area and other high-quality areas are spliced and compressed to obtain compressed data, a binary mask for identifying the enhanced area is generated, and the compressed data and the binary mask are packed into a second data packet; If the comprehensive confidence coefficient corresponding to the unmanned aerial vehicle system is smaller than or equal to a second interval threshold value and the task emergency degree is larger than the emergency degree threshold value, extracting a target category and a boundary box in target observation information through a lightweight target detection network, calculating a perception hash value of a target area, and packaging a timestamp, the target category, the boundary box, the perception hash value and a re-observation request mark into a third data packet; And broadcasting the first data packet, the second data packet or the third data packet corresponding to the unmanned system to the neighbor unmanned system through a distributed communication protocol.

Description

Unmanned system distributed data transmission method based on fuzzy confidence and hysteresis trigger Technical Field The application relates to the technical field of unmanned system communication, in particular to an unmanned system distributed data transmission method based on fuzzy confidence and hysteresis trigger. Background The existing unmanned system reduces transmission by setting a threshold value based on an event-triggered distributed data transmission method, but the following outstanding defects still exist: 1. the data confidence evaluation and the task value evaluation are mutually split and are rigidified, the reliability and the task value of the data are generally calculated independently in the prior art, and then simple linear weighting or sequential judgment is carried out, so that in a typical conflict scene of 'high task value and low observation confidence', the system delays or discards key information due to the rigidified quality judgment standard, and the transmission decision cannot be dynamically adjusted according to the essential requirement of the task. 2. The trigger mechanism has poor immunity and is easy to cause network oscillation, the current method mostly adopts a trigger mode of a single static threshold value, when the data confidence level or the network state slightly fluctuates, frequent trigger-silence state switching, namely 'ping-pong effect', is easy to generate a large number of invalid control instructions and aggravate network instability. 3. The data transmission strategy is extensive, the bandwidth utilization rate is low, namely, a unified compression coding mode is generally adopted after triggering, and the resource allocation is not refined according to the final value and quality evaluation result of the data, so that the link congestion or high-value information loss can be caused under the condition of extreme bandwidth limitation. Disclosure of Invention In order to solve the above-mentioned problems, it is necessary to provide an unmanned system distributed data transmission method based on fuzzy confidence and hysteresis trigger, which comprises the following steps: s1, for target observation information acquired by any unmanned system at the current moment, extracting an uncertainty feature vector and a task value feature vector corresponding to the unmanned system in parallel; s2, calculating a high trigger threshold and a low silence threshold based on the comprehensive confidence coefficient of the unmanned system at the past moment, and updating the data transmission state of the unmanned system at the current moment according to the hysteresis comparison result of the comprehensive confidence coefficient corresponding to the unmanned system at the current moment and the high trigger threshold or the low silence threshold; and S3, when the updated data transmission state is active, the unmanned system executes a differential data transmission strategy according to the corresponding comprehensive confidence. Preferably, the target observation information includes an image, a point cloud or a radio frequency signal. Preferably, extracting an uncertainty feature vector of the unmanned system includes: Calculating the structural similarity index of the target observation information to obtain a quality index; Calculating the shielding ratio of the target observation information through a segmentation algorithm; estimating the motion ambiguity corresponding to the unmanned system based on the inertial measurement unit data; And calculating the signal-to-noise ratio of the target observation information. Preferably, extracting a task value feature vector of the unmanned system includes: the unmanned system acquires a task and calculates a normalized distance between the unmanned system and a target point in the task; inquiring the task urgency of the acquired task; a time-dependent decay factor based on the discovery period is calculated. Preferably, the uncertainty feature vector and the task value feature vector are subjected to double-layer fuzzy reasoning, which comprises the following steps: blurring each element in the task value feature vector into a corresponding language value through a membership function, and mapping each language value corresponding to the task value feature vector into a rule modulation factor according to a preset mapping table; And mapping the uncertainty characteristic vector into the comprehensive confidence corresponding to the unmanned system at the current moment based on the rule modulation factor and a preset fuzzy mapping rule. Preferably, the preset fuzzy mapping rule is: And if the quality index is a first preset value, the shielding ratio is a second preset value, the motion blur degree is a third preset value and the signal to noise ratio is a fourth preset value, taking the value calculated by the rule modulation factor as the comprehensive confidence degree. Preferably, calculating the high trigger