CN-121984614-A - Triggering type online calibration trusted fusion composition method and system

CN121984614ACN 121984614 ACN121984614 ACN 121984614ACN-121984614-A

Abstract

The invention discloses a triggering type online calibration trusted fusion composition method and system, and relates to the technical field of spectrum detection. The method comprises the steps of obtaining multi-band detection data output by a plurality of detection nodes, establishing a calibration parameter set and a trust state set, determining a trigger index according to a calibration drift amount and a consistency deviation amount between the nodes, correcting a trigger threshold according to the trust state, triggering on-line calibration when conditions are met, updating calibration parameters based on reference measurement data and completing multi-band detection data correction, combining data quality weights and the trust state to perform fusion processing, obtaining a fusion result and node confidence coefficient, and constructing a spectrum environment diagram. The method and the device can improve the pertinence of on-line calibration triggering, improve the consistency of the multi-detection node detection data after calibration, enhance the stability and the reliability of the fusion result, and realize the structural expression of the detection result and the reliability degree thereof.

Inventors

GUO QIANG
LIANG ZELIN
MAO NING
GU RONGZHI
LU MAOLIN
HAO YANAN
ZHAO XIAOYA
Ning Yedong

Assignees

中国铁塔股份有限公司天津市分公司

Dates

Publication Date: 20260505
Application Date: 20260409

Claims (10)

1. The triggering type online calibration trusted fusion composition method is characterized by comprising the following steps of: The method comprises the steps of S1, acquiring multi-band detection data output by a plurality of detection nodes, establishing a calibration parameter set and a trust state set, S2, determining a calibration drift amount and a consistency deviation amount between nodes based on the multi-band detection data in a preset sliding time window, aggregating the calibration drift amount and the consistency deviation amount between the nodes to obtain a trigger index, correcting a reference trigger threshold according to the trust state set to obtain a trigger threshold, triggering an online calibration event when the trigger index reaches the trigger threshold, S3, after triggering the online calibration event, estimating the calibration amount based on reference measurement data and updating the calibration parameter set, performing time reference alignment, frequency reference alignment, phase reference alignment and frequency band response correction on the multi-band detection data according to the updated calibration parameter set to obtain calibrated detection data, S4, determining a data quality weight set based on the calibrated detection data and updating the trust state set, determining a fusion weight based on the data quality weight set and the trust state set, performing fusion processing according to frequency bands to obtain a fusion result and node confidence level, S5, constructing an environment map based on the fusion result and the node confidence level, wherein the environment map comprises a corresponding spatial frequency band unit, a node side map, a corresponding to the confidence level map and a node side map, and a side map corresponding to the two-to a side map of the node confidence level.
2. The method for triggering on-line calibration trusted fusion composition according to claim 1 is characterized in that multi-frequency band detection data are associated to form a data unit according to detection nodes, frequency bands and acquisition time sequences, the calibration parameter set comprises time deviation parameters, frequency deviation parameters, phase deviation parameters, frequency band response error parameters and corresponding drift baseline parameters according to detection nodes and frequency band indexes, and the trust state set comprises node-level trust states and frequency band-level trust states, wherein the node-level trust states and the frequency band-level trust states are determined according to reference measurement consistency in an initialization stage and the difference degree of detection values of all detection nodes to the same frequency band.
3. The method for triggering on-line calibration trusted fusion composition according to claim 1, wherein the calibration drift amount is determined according to a difference between a current calibration parameter of each detection node and a corresponding drift baseline parameter, the difference comprises a time deviation difference, a frequency deviation difference, a phase deviation difference and a frequency band response error difference, the inter-node consistency deviation amount is determined according to a deviation of a detection value of each detection node in the same frequency band relative to a frequency band reference detection value, the frequency band reference detection value is determined according to a node level trust state weighting of each detection node participating in frequency band detection, the triggering index is determined according to the calibration drift amount of each detection node and the inter-node consistency deviation amount of each frequency band in a combined manner, the triggering threshold is determined according to a reference triggering threshold and node level trust state correction, and when the node level trust state is reduced, the triggering threshold is reduced.
4. A method for triggering an online calibration trusted fusion patterning as defined in claim 3, wherein the triggering criteria is And the trigger threshold Each of which is determined as follows: ; When (when) Triggering an online calibration event; In the formula, Represent the first Calibration drift amounts of the individual detection nodes; represent the first The amount of deviation in the node consistency of each frequency band; represent the first Node level trust status of individual probing nodes; represent the first A frequency band level trust status for each frequency band; representing the number of the detection nodes; representing the number of frequency bands; Representing the aggregation factor; representing a baseline trigger threshold; Representing the threshold correction coefficient.
5. The method for triggering on-line calibration trusted fusion composition according to claim 1, wherein the estimating calibration quantity based on reference measurement data and updating calibration parameter sets comprise obtaining reference measurement data corresponding to each detection node and each frequency band, wherein the reference measurement data is any one of standard signal injection measurement data, air interface reference path measurement data and bidirectional round trip time measurement data, respectively determining time deviation calibration quantity, frequency deviation calibration quantity, phase deviation calibration quantity and frequency band response error calibration quantity of each detection node on a corresponding frequency band according to deviation between the reference measurement data and a preset reference value, updating the time deviation calibration quantity, the frequency deviation calibration quantity, the phase deviation calibration quantity and the frequency band response error calibration quantity to corresponding parameters in the calibration parameter sets to obtain updated calibration parameter sets, and sequentially executing time reference alignment, frequency reference alignment, phase reference alignment and frequency band response correction on the multi-band detection data according to the updated calibration parameter sets to obtain calibrated detection data.
6. The triggered online calibration trusted fusion composition method of claim 1 is characterized by comprising the steps of determining fusion weights based on a data quality weight set and a trust status set, executing fusion processing according to frequency bands, respectively determining corresponding data quality weights for calibrated detection data of all detection nodes on the corresponding frequency bands, determining the data quality weights according to two optional items of reconstruction errors, prediction errors, cross-window stability differences and reference consistency differences of the calibrated detection data, updating the trust status set according to historical consistency changes, abnormal correction results and reference similarity changes of all the detection nodes, combining the node level trust status and the frequency band level trust status to obtain comprehensive trust status of the corresponding detection nodes on the corresponding frequency bands according to each frequency band, determining fusion weights of all the detection nodes on the corresponding frequency bands based on the data quality weights and the comprehensive trust status of the corresponding detection nodes, carrying out normalization constraint on the fusion weights, carrying out weighted fusion on the calibrated detection data of all the detection nodes on the corresponding frequency bands according to the fusion weights after normalization constraint, obtaining fusion results of the corresponding frequency bands, and determining the degree of the weighted fusion results of the corresponding detection nodes with a large degree of deviation of the confidence level from the calibrated detection nodes on the corresponding frequency bands.
7. The triggered online calibration trusted fusion composition method of claim 6, wherein the fusion weights are determined as follows: ; In the formula, And Respectively representing the indexes of the detection nodes; Representing a frequency band index; And Respectively represent the first The detection node and the first The detection node is at the first Data quality weights of the individual frequency bands; And Respectively represent the first The detection node and the first The detection node is at the first Comprehensive trust status of individual frequency bands; And Respectively represent the first The detection node and the first The detection node is at the first Initial fusion weights of the individual frequency bands; represent the first The detection node is at the first Normalized fusion weights of the individual frequency bands; And Respectively representing a data quality weight index parameter and a comprehensive trust state index parameter; representing a fusion weight upper bound parameter; and the normalized fusion weight is used for the weighted fusion processing of the corresponding frequency band.
8. The triggered online calibration trusted fusion composition method according to claim 1 is characterized in that the construction of the spectrum environment graph based on the fusion result and the node confidence coefficient comprises the steps of taking the fusion result and the node confidence coefficient of each space unit on each frequency band as graph node attribute values of the combination of the corresponding space unit and the corresponding frequency band, determining a preset adjacent relation according to the adjacent relation of the space grid, the overlapping relation of the coverage area and the geographic adjacent relation, and establishing corresponding graph edges in the same frequency band for the adjacent space units corresponding to the preset adjacent relation.
9. The triggered online calibration trusted fusion composition method of claim 1, wherein the determining of the edge confidence comprises calculating a result difference degree for fusion results of graph nodes at two ends of a graph edge in the same frequency band and using the result difference degree as a fusion result consistency metric value, determining the edge confidence degree of a corresponding graph edge according to a preset edge confidence degree calculation rule according to the fusion result consistency metric value and the node confidence degree of the graph nodes at two ends of the graph edge, reducing the edge confidence degree when the fusion result consistency metric value is increased or the node confidence degree of any graph node is reduced, and recording a corresponding space unit, a corresponding frequency band, a corresponding fusion result and a corresponding node confidence degree by each graph node in a spectrum environment graph, wherein each graph edge records a corresponding edge confidence degree.
10. A triggered online calibration trusted fusion composition system for use in a triggered online calibration trusted fusion composition method as claimed in any one of claims 1 to 9, comprising: The system comprises a state modeling module, a joint triggering module, a trusted state aggregation module, an environment module, a context module and a graph-side graph, wherein the state modeling module is used for acquiring multi-band detection data output by a plurality of detection nodes and establishing a calibration parameter set and a trusted state set, the joint triggering module is used for determining a calibration drift amount and an inter-node consistency deviation amount based on the multi-band detection data in a preset sliding time window, aggregating the calibration drift amount and the inter-node consistency deviation amount to obtain a triggering index, correcting a reference triggering threshold according to the trusted state set to obtain a triggering threshold, triggering an on-line calibration event when the triggering index reaches the triggering threshold, the on-line calibration module is used for estimating the calibration amount based on reference measurement data and updating the calibration parameter set after the on-line calibration event, executing time reference alignment, frequency reference alignment, phase reference alignment and frequency response correction according to the updated calibration parameter set to the multi-band detection data, determining a data quality set and updating the trusted state set based on the calibrated detection data, determining a fusion weight based on the data quality weight set and the trusted state set, executing fusion processing according to the frequency band, obtaining a fusion result and node confidence, and constructing an environment graph based on the fusion result and the node confidence, wherein the environment graph comprises a corresponding graph and the node confidence edge graph and the graph-side graph.

Description

Triggering type online calibration trusted fusion composition method and system Technical Field The invention relates to the technical field of spectrum detection, in particular to a triggering type online calibration trusted fusion composition method and system. Background In the multi-detection node and multi-band spectrum detection scene, the prior art generally adopts a plurality of detection nodes to respectively collect spectrum information of different frequency bands, and then performs calibration, fusion and output on the collection result. In the existing scheme, the calibration process adopts a static calibration mode after fixed period calibration or initialization, the fusion process adopts fixed weight, experience rule or single evaluation factor for weighting, and the result output is expressed as discrete detection result, spectrum distribution result or simple graph display result. In the actual operation process, each detection node is easily influenced by factors such as clock drift, frequency drift, device state change, environment interference, operation condition change and the like, so that the measurement results among different detection nodes gradually generate deviation. For the fixed period calibration mode, when the running state of the system changes less, the repeated calibration can bring unnecessary resource consumption, and when the running state of the system changes more quickly, the system cannot respond in time, so that the subsequent detection result is distorted. For the static calibration mode, since the calibration parameters remain unchanged for a long time, it is difficult to reflect the dynamic changes of the states and the measurement conditions of the detection nodes in the running process in time. On the other hand, in multi-detection node, multi-band cooperative processing, the output quality, stability and consistency of different detection nodes on different frequency bands are not generally the same. If the existing fusion mode only depends on fixed weights or single evaluation factors, the actual reliability of different detection nodes on different frequency bands is difficult to comprehensively reflect. When the local nodes are abnormal, the partial frequency bands are mismatched or the fluctuation of the detection result is increased, the influence of the low-reliability data on the whole fusion result is difficult to be effectively restrained by the existing scheme. In the prior art, only the fusion result is output as a result, and a unified organization mode for spatial association relation, frequency band association relation and result reliability degree is lacking, so that the detection result and the association relation thereof are difficult to express simultaneously in a complex electromagnetic environment, and the analysis and the utilization of the regional spectrum state are not facilitated. Therefore, a new spectrum detection processing scheme is needed to solve the problems of inaccurate calibration triggering, insufficient anti-abnormal capability of fusion processing and lack of trusted structural expression of detection results in multi-detection nodes and multi-band scenes. Disclosure of Invention Aiming at the problems of inaccurate calibration triggering, insufficient reliability of fusion results and lack of trusted structure expression in the existing multi-detection node and multi-band frequency spectrum detection scheme, the invention provides a trigger type online calibration trusted fusion composition method and system, so as to improve the pertinence of calibration triggering, the stability of fusion processing and the structuring degree of detection result expression. In order to achieve the above object, the present invention is realized by the following technical scheme: in one aspect, the invention provides a triggered online calibration trusted fusion composition method, which comprises the following steps: S1, acquiring multi-frequency band detection data output by a plurality of detection nodes, and establishing a calibration parameter set and a trust state set; S2, determining a calibration drift amount and a consistency deviation amount between nodes based on multi-band detection data in a preset sliding time window, and aggregating the calibration drift amount and the consistency deviation amount between nodes to obtain a trigger index; S3, after triggering an online calibration event, estimating a calibration quantity based on reference measurement data, updating a calibration parameter set, and performing time reference alignment, frequency reference alignment, phase reference alignment and frequency band response correction on the multi-band detection data according to the updated calibration parameter set to obtain calibrated detection data; S4, determining a data quality weight set based on the calibrated detection data, updating a trust state set, determining a fusion weight based on the data quality weight se