CN-121995328-A - False AIS signal detection method based on multi-source data

CN121995328ACN 121995328 ACN121995328 ACN 121995328ACN-121995328-A

Abstract

The application relates to the technical field of ship traffic management and offshore monitoring, and discloses a false AIS signal detection method based on multi-source data, which comprises the steps of firstly, extracting a historical track to construct a coverage grid mapping table containing radar detection attributes; the method comprises the steps of receiving real-time data, dividing the real-time data into an AIS to-be-detected group and a radar truth group, screening non-radar blind area targets according to a mapping table, searching related targets in the radar truth group to execute existence reverse verification, adopting a dynamic time warping algorithm to align time sequences and calculate characteristic differences, executing fine comparison verification, constructing a neighborhood flow field model, executing flow field consistency verification according to motion vector deviation, and outputting a false AIS signal list by combining a multi-stage verification result. According to the application, through physical coverage modeling and environmental dynamics constraint, false signals with high simulation degree in radar blind areas can be effectively identified, so that the false alarm rate is reduced.

Inventors

CEN ZHENJIN
ZHU DEGAO
LI XIAOJUN
ZHANG WEIJIA
CHEN YIKANG
ZHAO DEPING
ZHAO RULEI
WANG YOUHE
CHEN HANGWEI
HUANG XIAOWEI
WANG JIANYING
GUO QIANG
ZHANG SHIGUAN
Qin Jinzhi
WU CONG
PANG TAI

Assignees

中国南方电网有限责任公司超高压输电公司广州局海口分局

Dates

Publication Date: 20260508
Application Date: 20260109

Claims (10)

1. The false AIS signal detection method based on the multi-source data is characterized by comprising the following steps of: extracting ship track data with radar signals from historical multi-source data, and constructing a radar coverage grid mapping table containing radar detection capability attributes by adopting GeoHash codes; receiving real-time ship multi-source track data, constructing a track sequence according to a ship identification code, and dividing the track sequence into an AIS to-be-detected group and a radar truth value group according to the data source attribute of data points in the track sequence; Matching tracks in the AIS to-be-detected group according to the radar coverage grid mapping table, and screening ship targets to be detected in non-radar blind areas; Searching the related radar targets around the ship targets to be detected in the radar truth value group, and judging the existence of radar echoes of the ship targets to be detected according to the search result to be used as existence reverse verification; aiming at the condition that the associated radar target is searched, carrying out space-time alignment and feature difference calculation on the ship target to be detected and each searched associated radar target by adopting a dynamic time warping algorithm, and judging the physical identity of the ship target to be detected and the associated radar target according to a feature difference calculation result to be used as comparison verification; selecting radar truth value group targets in the neighborhood range of the ship targets to be detected to construct a flow field model, and executing flow field consistency verification according to motion vector deviation; And outputting a false AIS signal list by combining the judgment result of the existence reverse verification, the judgment result of the comparison verification and the judgment result of the flow field consistency verification.
2. The method for detecting false AIS signals based on multi-source data according to claim 1, wherein said constructing a radar coverage grid mapping table including radar detection capability attributes using GeoHash codes comprises: Performing GeoHash spatial discretization coding on longitude and latitude of historical track data with radar signal marks extracted from the historical multi-source data; taking GeoHash coded values as aggregation keys, and counting radar point position density and minimum detection ship length in each grid; The minimum detected ship length is obtained by one mode selected from the minimum value in a ship length set corresponding to all track points falling into the grid and the numerical value in a set percentile in the ship length set; and taking the GeoHash coded values as keys, taking the radar point density and the minimum detection ship length as values, and storing the values into the radar coverage grid mapping table.
3. The method for detecting false AIS signals based on multi-source data according to claim 1, wherein dividing the track sequence into AIS to-be-detected groups and radar truth groups according to data source attributes of data points in the track sequence comprises: Maintaining a sequence of trajectories within a predetermined time window for each vessel; Under the condition that one track sequence only contains data from an AIS base station, classifying the track sequence into the AIS to-be-detected group; In the case that one track sequence contains data from a radar station, and in the case that one track sequence contains fused track data generated by fusion of AIS data with radar data, the track sequence is classified into the radar truth group.
4. The method for detecting false AIS signals based on multi-source data according to claim 2, wherein said matching the tracks in the AIS to-be-detected group according to the radar coverage grid mapping table, screening the ship targets to be detected in the non-radar blind area, comprises: Acquiring a track sequence of a target to be screened in the AIS to be tested, and a captain key field contained in the track sequence; calculating GeoHash codes for each track point in the track sequence, and inquiring in the radar coverage grid mapping table; Determining that the track point meets radar detection conditions under the condition that GeoHash codes corresponding to the track point exist in the radar coverage grid mapping table and the numerical value of the key field of the captain exceeds the minimum detection captain recorded by the grid; And calculating the coverage rate of the track points meeting radar detection conditions in the whole track sequence, and determining that the target to be screened belongs to the ship target to be detected of the non-radar blind area under the condition that the coverage rate exceeds a coverage rate set threshold value.
5. A false AIS signal detection method based on multi-source data as claimed in claim 4, wherein said presence reverse verification comprises: Searching targets in the radar truth value group in a set searching radius by taking a position corresponding to the latest timestamp in the ship target track sequence to be detected as a center; under the condition that the search result is an empty set, determining that the ship target to be detected belongs to a false AIS signal; and under the condition that the search result is not null, taking the set formed by all the searched targets as an associated radar target candidate set, and entering the comparison verification step.
6. The method of claim 5, wherein said employing dynamic time warping algorithm to time-space align said ship target to be detected with each of said associated radar targets retrieved comprises: Constructing a cost matrix aiming at the track sequence of the ship target to be detected and the track sequence of any target in the associated radar target candidate set, wherein elements of the cost matrix represent absolute values of time stamp difference values of data points in the two sequences; and searching a path which passes through the cost matrix and has the minimum total time difference cost by adopting a dynamic programming method, and outputting a group of time-aligned track point pair sequences.
7. The method for detecting false AIS signals based on multi-source data according to claim 6, wherein said determining physical identity of the ship target to be detected and the associated radar target according to the feature difference calculation result includes: based on the time-aligned track point pair sequences, respectively calculating a position distance difference average value, a captain attribute difference average value, a navigational speed difference average value and a heading difference average value; determining a set similarity judgment threshold, and determining that any one of the position distance difference average value, the captain attribute difference average value, the navigational speed difference average value and the heading difference average value is not a physical entity corresponding to the ship target to be detected under the condition that any one of the position distance difference average value, the captain attribute difference average value, the navigational speed difference average value and the heading difference average value exceeds the corresponding set similarity judgment threshold; and under the condition that all targets in the associated radar target candidate set cannot meet the similarity condition, identifying that the ship target to be detected belongs to a false AIS signal.
8. The method for detecting false AIS signals based on multi-source data according to claim 1, wherein the selecting radar truth group targets in the ship target neighborhood range to be detected to construct a flow field model comprises: Taking a grid at a position corresponding to the latest timestamp in the ship target track sequence to be detected as a reference, and selecting grids adjacent to the periphery of the reference as a flow field sampling area; Searching all targets belonging to the radar truth value group and AIS targets verified to be real in the past detection period in the flow field sampling area to form a flow field reference data set; And skipping the flow field consistency check under the condition that the target number in the flow field reference data set does not reach a preset statistics lower limit threshold value.
9. The method for detecting false AIS signals based on multi-source data according to claim 8, wherein said calculating a neighborhood flow field average vector in said constructed flow field model comprises: obtaining the ground speed and the ground heading of each reference target in the flow field reference data set; decomposing the motion vector of each reference target into a forward direction component and a north direction component by adopting a trigonometric function relation; Respectively calculating the arithmetic average value of the forward direction components and the arithmetic average value of the forward direction components of all the reference targets in the region; and synthesizing a local flow field average vector based on the arithmetic average value of the forward direction component and the arithmetic average value of the forward direction component, wherein the local flow field average vector comprises flow velocity and flow direction.
10. The method of claim 9, wherein performing a flow field consistency check based on motion vector bias comprises: Acquiring the navigational speed and the navigational course of the ship target to be detected so as to form a motion vector; calculating a consistency deviation index, wherein the consistency deviation index is obtained by weighting and summing a navigational speed difference item and a heading direction difference item; The navigational speed difference item belongs to the ratio relation between the absolute value of the difference between the target navigational speed of the ship to be detected and the average vector flow speed of the local flow field and the flow speed; the course difference term is calculated based on cosine similarity between the target course of the ship to be detected and the average vector flow direction of the local flow field; and under the condition that the consistency deviation index exceeds a preset flow field consistency threshold value, determining that the ship target to be detected belongs to a false AIS signal.

Description

False AIS signal detection method based on multi-source data Technical Field The invention relates to the technical field of ship traffic management and offshore monitoring, in particular to a false AIS signal detection method based on multi-source data. Background The ship Automatic Identification System (AIS) is used as a ship-borne navigation and communication device, and can support the perception of traffic situation of peripheral ships and shore-based facilities by broadcasting dynamic information such as identity, position, speed, heading and the like, so that the application is wide in the aspects of ensuring navigation safety and assisting traffic management. However, since the AIS protocol is open and lacks a strict identity authentication mechanism, the system is at risk of fraud attacks. The non-cooperative targets can generate non-existent false targets or falsify real tracks through forging or falsifying AIS signals to interfere with shipping order, and even be used for covering illegal activities such as smuggling, illegal fishing and the like. Existing identification techniques for false AIS signals rely mainly on feature analysis of the AIS data itself. For example, the judgment is performed by analyzing the kinematic rationality or the signal receiving range of the ship dynamic data, but the method is difficult to cope with the track spoofing with high simulation degree, and an attacker can simulate the false track conforming to the motion rule to avoid detection. Another common method is time slot conflict analysis based on SOTDMA protocol, but in the sparse sea area of the ship, time slot resources are abundant, an attacker can easily select idle time slots to send false messages without generating conflicts, and in the dense sea area of the ship, normal time slot conflicts frequently occur, so that the attack behavior and normal network congestion are difficult to effectively distinguish, and the misjudgment rate is higher. Relying on a single AIS data source for verification makes it difficult to accurately identify falsified signals that are falsified without an external physical truth reference. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a false AIS signal detection method based on multi-source data, which solves the technical problems that radar blind areas and false signals are difficult to distinguish and false tracks are difficult to identify based on multidimensional features in ship supervision in the prior art. The false AIS signal detection method based on the multi-source data comprises the steps of establishing a radar coverage grid mapping table through mining historical data, wherein the mapping table essentially forms a physical view diagram of a radar system and is used for eliminating blind area interference caused by objective factors such as terrain shielding or distance limitation before real-time detection and ensuring that only targets which should be detected theoretically are verified. In a real-time detection flow, the method establishes verification logic with radar observation data as a true value reference. Aiming at AIS targets in non-blind areas, the system adopts a three-level verification mechanism from thick to thin and from individual to environment: the first stage is existence reverse verification, and whether physical radar echoes exist at the periphery of the AIS target is rapidly judged by utilizing the spatial proximity so as to identify completely imaginary ghost ship signals; The second stage is space-time feature comparison verification, a dynamic time warping algorithm is introduced to solve the problem of asynchronous sampling frequency of heterogeneous sensors, and on the basis of time sequence alignment, the fine differences of the targets in position, captain, navigational speed and navigational direction are accurately calculated to identify the track counterfeiting or attribute tampering of the transfer grafting wood; And thirdly, checking consistency of the flow field, constructing a dynamic flow field model reflecting the hydrologic characteristics of the current sea area by utilizing a real radar target in the adjacent area, and comparing the motion vector of the target to be tested with the environmental flow field to identify complex deceptive signals which are smooth in track but violate the sea current motion rule. Finally, the system integrates the three-dimensional verification results and outputs a false AIS signal list with high confidence. The invention provides a false AIS signal detection method based on multi-source data. The beneficial effects are as follows: 1. According to the invention, the radar detection boundary is quantized by constructing the radar coverage grid mapping table containing the minimum detection captain attribute. The targets in radar dead zones or with the size smaller than radar detection precision are removed before detection, so that false judgment of r