CN110888126A - Unmanned ship information perception system data comprehensive processing method based on multi-source sensor - Google Patents

Abstract

The invention discloses a comprehensive processing method of unmanned ship information sensing system data based on a multi-source sensor, which is used for solving the technical problem of poor practicability of the existing data processing method of the unmanned ship information sensing system. The technical scheme includes that data of a multi-source sensor are collected, coordinate transformation and time alignment are carried out according to data formats of the sensors, space-time unification of the data is achieved, then radar, AIS and ESM data are respectively associated with sonar information in a track mode, repeated tracking targets in a monitoring area are eliminated, unidentified non-cooperative underwater targets are marked in a reset point mode, and sea condition grades are reversely deduced through inertial navigation data and wind speed information, so that real scenes of targets and environments in a monitoring sea area are obtained. Because the data such as radar, AIS and ESM are respectively associated with sonar information in a track way, repeated tracking targets in the monitoring area can be eliminated from different detection ranges, accuracy, different target attribute descriptions and the like in a multi-level and multi-dimension way, and the practicability is good.

Description

Unmanned ship information perception system data comprehensive processing method based on multi-source sensor

Technical Field

The invention relates to a data processing method of an unmanned ship information sensing system, in particular to a comprehensive data processing method of the unmanned ship information sensing system based on a multi-source sensor.

Background

China is a developing large and medium-sized country with abundant marine resources and frequent marine disputes, and the research of unmanned ship technology has important military and economic significance for protecting the marine rights and interests of China.

The unmanned ship information perception integrated system is core electronic equipment of an unmanned ship, is an information basis for the unmanned ship to autonomously complete tasks such as environment perception, target detection and the like, and is a basic guarantee for completing autonomous navigation and control. According to different mission tasks, the information perception integrated system can adopt different solutions and carry various sensors, namely a radar, a sonar, an ESM system, an infrared detection system, an AIS, an inertial navigation system and the like. The single sensor can not provide all key information required by a command and control system to complete the detection and sensing functions, and each sensor has respective advantages and defects, so that the fusion of multiple sensors is required.

Document 1 "ship-borne multiple sensor data integrated analysis [ J ] space electronic countermeasure, 1998(1): 46-50" proposes a variety of sensors that can be carried on a ship, such as: ESM system, radar, sonar, infrared detection system, identification of friend or foe (IFF), optical tracking system, etc. the discussion is the necessity of ship-based multi-sensor data synthesis.

Document 2 "association method of high-frequency ground wave radar and AIS target track based on fuzzy double thresholds" proposes an association method of ground wave radar and ship identity Automatic Identification System (AIS) target track based on fuzzy double thresholds, which mainly uses fuzzy membership to describe the association degree between two tracks, and determines the associated track pair through double threshold detection. It can be seen that: through data fusion and track association, on one hand, the problem of repeated tracking in a monitoring area can be solved, and the tracking efficiency is improved; another aspect may focus on unidentified non-cooperative targets by excluding associated cooperative targets. However, neither of these documents suggests how to detect and calibrate non-cooperative underwater targets, i.e. how to strip the underwater targets from sonar information of the above-water targets, which is the key to performing anti-submarine tasks and is one of the most important and frequently performed tasks of unmanned ships.

Document 3 "multiple sonar radar data fusion system implementation [ J ] network new media technology, 2014,3(4): 57-59" introduces a set of multiple sonar radar data fusion system, aims to realize sensing and detection of underwater targets on water, and performs classified management and display on the targets. However, the paper only gives a simple system configuration and a data time and registration method, does not give descriptions of specific sensor forms, fusion algorithms and other key technologies, and does not consider the influence of the environmental changes of the ocean on the sensors and the method for acquiring the actual sea state information.

Disclosure of Invention

In order to overcome the defect that the data processing method of the existing unmanned ship information sensing system is poor in practicability, the invention provides a comprehensive data processing method of the unmanned ship information sensing system based on a multi-source sensor. The method comprises the steps of firstly collecting data of a multi-source sensor, carrying out coordinate transformation and time alignment according to data formats of the sensors to realize the unification of the data in time and space, then respectively carrying out track association on radar, AIS (automatic identification system) and ESM (electronic service management) data and sonar information to eliminate repeated tracking targets in a monitoring area, and repeatedly marking unidentified non-cooperative underwater targets, and reversely deducing sea state grades through inertial navigation data and wind speed information to obtain real scenes for monitoring targets and environments in a sea area. Because the data such as radar, AIS, ESM and the like are respectively associated with sonar information in a track mode, repeated tracking targets in a monitoring area can be eliminated from multi-level and multi-dimensionality of different detection ranges, accuracy, different target attribute description and the like, non-identity non-cooperative underwater targets can be subjected to key calibration, and field sea condition information is considered at the same time to obtain real scenes of the targets and the environment in the monitoring sea area, so that information support is provided for realizing autonomous environment perception and underwater target detection of unmanned ships, completing tasks such as anti-submergence and the like, and the practicability is good.

The technical scheme adopted by the invention for solving the technical problems is as follows: a comprehensive data processing method of an unmanned ship information perception system based on a multi-source sensor is characterized by comprising the following steps:

reading the data rate according to the data type of the multi-source sensor, analyzing the read data, and then transmitting the data into a data preprocessing part.

And step two, the preprocessing part analyzes and normalizes the read data, performs coordinate conversion on the data in different coordinate systems, and performs time synchronization on the data transmitted at different times.

Each sensor outputs a string of data sequences with time synchronization, equal interval and numerical value normalization, generates standardized preprocessed data and sends the preprocessed data to the fusion processing and information comprehensive processing unit.

And step three, the processing of multi-source sensor data fusion comprises two parts of track initiation and track association. And (3) establishing an initial wave gate at the beginning of the flight path by adopting a speed method, if three point paths fall into the related wave gates, successfully initiating the flight path, performing filtering initialization and establishing the point paths and the related wave gates of the flight path. And the track correlation is to correlate the sonar information with radar, AIS and ESM data, and if the sonar information is correlated with any information, the target is considered to be an overwater target, and if the sonar information is not correlated with any information, the target is considered to be an underwater target.

Step four, the comprehensive content of the multi-source sensor data comprises the following three parts:

unmanned ship early warning of overturning: according to the maximum roll angle and the maximum pitch angle which can be borne by the unmanned ship, the maximum roll and pitch change rate is divided into nine grades of 1-9, wherein the grade 1 is basically free of overturning danger, and the grade 9 is the potential possibility of overturning danger. And judging the current corresponding overturning grade of the ship through the transverse and longitudinal rocking angles of the unmanned ship returned by the inertial navigation system, and judging whether the ship has overturning danger.

Sea state grade estimation: the method comprises the steps of collecting wind speed information of a weather instrument of the unmanned ship and roll angle information of inertial navigation respectively, estimating a roll motion response function of the unmanned ship on line through the roll angle information, then performing a wave spectrum in a reverse mode, extracting wave height information from the wave spectrum, and estimating the sea condition grade of the current sea area jointly according to the relation between the wave grade and the wave height and the wind speed.

Target hazard level and data confidence assignment: after the tracks are associated, according to the attributes of the targets, namely whether the targets are water targets or underwater targets, whether the targets have electronic countermeasure conditions or not and whether the targets transmit AIS information or not, the danger levels are given to the targets. And according to the data source, the target type, the correlation quality and the environmental factors of the target, comprehensively judging the confidence degree of the output target information according to different weight distribution.

And fifthly, generating a comprehensive report.

And sequentially outputting the UTC time, the target batch number, the target type, the target direction, the direction unit, the target distance, the target strength, the target speed, the target threat level, the target information confidence coefficient, the sea condition information and the ship information obtained by the four steps.

The invention has the beneficial effects that: the method comprises the steps of firstly collecting data of a multi-source sensor, carrying out coordinate transformation and time alignment according to data formats of the sensors to realize the unification of the data in time and space, then respectively carrying out track association on radar, AIS (automatic identification system) and ESM (electronic service management) data and sonar information to eliminate repeated tracking targets in a monitoring area, and repeatedly marking unidentified non-cooperative underwater targets, and reversely deducing sea state grades through inertial navigation data and wind speed information to obtain real scenes for monitoring targets and environments in a sea area. Because the data such as radar, AIS, ESM and the like are respectively associated with sonar information in a track mode, repeated tracking targets in a monitoring area can be eliminated from multi-level and multi-dimensionality of different detection ranges, accuracy, different target attribute description and the like, non-identity non-cooperative underwater targets can be subjected to key calibration, and field sea condition information is considered at the same time to obtain real scenes of the targets and the environment in the monitoring sea area, so that information support is provided for realizing autonomous environment perception and underwater target detection of unmanned ships, completing tasks such as anti-submergence and the like, and the practicability is good.

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

Drawings

FIG. 1 is a flow chart of a comprehensive data processing method of an unmanned ship information perception system based on a multi-source sensor.

FIG. 2 is a diagram of data transmission between the unmanned ship integrated information processing modules in the method of the present invention.

Detailed Description

Reference is made to fig. 1-2. The comprehensive processing method of the unmanned ship information perception system data based on the multi-source sensor comprises the following specific steps:

the method comprises the following steps: multi-source sensor data is read.

Referring to table 1, the system has six heterogeneous sensors such as passive sonar and radar, and the data content and data rate of each sensor are shown in table 1.

TABLE 1 data content and data Rate for each sensor

Data source	Data content	Data rate
			Passive sonar data	Target azimuth, target number 4 batches	500～5000ms
Radar data	Target track information (azimuth and distance), target number 10 batches	2s
			AIS data	Location (latitude and longitude), MMSI number, target number 10 batches	500ms
Inertial navigation data	Attitude and azimuth information of ship body	5s
			Meteorological instrument data	Wind speed information	500ms
ESM data	Target azimuth, target number 10 batches	500ms

Each sensor has an independent asynchronous working mode, acquires original data according to different data rates and transmits the original data to a system input interface.

When the data acquisition software module is activated, the following working steps are sequentially completed:

① starting sensor collection software to read data sent from the input interface at regular time;

② comparing the data with the last data, if there is any update, moving to the buffer area, and marking the mark of "new data";

③ parse the new data and make a processing request to the pre-processing module.

Step two: multi-source sensor data preprocessing

The work flow of the pretreatment is carried out in the following order:

① different sensor measurements describing the same physical quantity are resolved and normalized to have the same units and dimensions.

②, because different sensors obtain different data rates and are not synchronous, the time synchronization is needed, the processing method is a) determining the standard interval of data, selecting the common multiple (or approximate value) of the two sensor data intervals as the time interval of the new sequence to normalize the data transmission rate, b) synchronizing the processing, preprocessing the original data in the same time interval by fitting compression or interpolation method to form the normalized sequence with the standard interval.

③, a carrier coordinate system is established by taking the unmanned ship as the origin and the water surface as the coordinate plane, and the data of the position information of each sensor in the carrier coordinate system is calculated by the coordinate conversion formulas of translation, rotation and the like of the data of each sensor.

Finally, each sensor outputs a string of data sequences with high reliability of time synchronization, equal interval and numerical value normalization, generates standardized preprocessed data and sends the preprocessed data to a fusion processing unit and an information comprehensive processing unit.

Step three: and fusing multi-source sensor data.

① track start.

In the initial stage of the track, the track head is not processed, and the initial wave gate is established by using a speed method by taking the track head as the center. And if the second point track falls into the initial wave gate, receiving the data of the second point track, establishing a second related wave gate, otherwise, canceling the track, and reselecting the track head to start. If the third point track falls into the relevant wave gate, the track is started successfully, the track mark character is set to be 1, and at the moment, the filtering initialization is carried out and the wave gate relevant to the point track and the track is established. And if the third point track does not fall into the second relevant wave gate, canceling the track and restarting.

② track association.

Adopting an angle tracking algorithm based on sliding window type double-threshold detection, firstly calculating the difference of the mean values of azimuth data of two nodes with the same track in a sliding window, then comparing the sum of squares of the mean values of all nodes in the sliding window with a set first threshold, and if the sum is greater than the first threshold, adding 1 to the value of an association frequency counter; and then, carrying out the same processing on the track data in the next sliding window, and finally, after processing a plurality of sliding windows, comparing the count value of the associated number counter with a set second threshold value. If the number of the flight paths is larger than the second threshold, the flight paths are judged to be related, namely the two flight paths are the same target, otherwise the flight paths are judged to be not related, namely the two flight paths respectively represent different targets.

The method comprises the steps of firstly associating 4 batches of sonar track information (azimuth angles) after track formation with 10 batches of radar track information (azimuth angles) at the same time, if the sonar track information is not associated with the radar track information (azimuth angles), sequentially associating the 4 batches of sonar track information (azimuth angles) with 10 batches of AIS data and 10 batches of ESM data, and if the sonar track information is associated with any one of radar, AIS and ESM data, considering that a target is an overwater target, and considering that the target is an underwater target if the sonar track information is not associated with any data.

Step four: multi-source sensor data synthesis.

① unmanned ship overturn warning.

According to the maximum roll angle and the maximum pitch angle which can be borne by the unmanned ship, the maximum roll and pitch change rate is divided into nine grades of 1-9, wherein the grade 1 is basically free of overturning danger, and the grade 9 is highly possible to have overturning danger. The current corresponding overturning grade of the ship can be judged through the transverse and longitudinal rocking angles of the unmanned ship returned by the inertial navigation system, and whether the ship has the overturning danger or not is judged.

② sea state rating estimation.

The sea condition grade estimation method based on unmanned ship inertial navigation and meteorological instrument information is adopted, wind speed information of the unmanned ship meteorological instrument and roll angle information of inertial navigation are respectively collected, a roll motion response function of the unmanned ship is estimated on line through the roll angle information, then a sea wave spectrum is inverted, sea wave height information is extracted from the sea wave spectrum, and finally the sea condition grade of the current sea area is estimated in a combined mode according to the relation between the sea wave grade and the wave height and the wind speed.

③ target hazard level and data confidence value.

Judging the threat level of the target according to the track association result, generally dividing the threat level into four levels, which are 1-4: a sonar associated with an AIS indicates a relatively safe level 4 if it is not associated with other sensors, indicating only underwater targets, representing the most dangerous level 1. Meanwhile, the confidence of the output data is comprehensively judged in a weighting processing mode according to the data source, the target type, the correlation quality and the environmental factors.

Step five: and forming an unmanned ship information perception comprehensive report.

After the repeated targets are eliminated through the track association, an unmanned ship information perception comprehensive report is formed and output, and the report contains the following contents: UTC time, target batch number, target type, target direction, target distance, target speed, target threat level, target information confidence, sea state information and ship information.

Claims (1)

1. A comprehensive data processing method of an unmanned ship information perception system based on a multi-source sensor is characterized by comprising the following steps:

reading the data rate according to the data type of the multi-source sensor, analyzing the read data, and transmitting the data into a data preprocessing part;

the preprocessing part analyzes and normalizes the read data, performs coordinate conversion on the data in different coordinate systems, and performs time synchronization on the data transmitted at different times;

each sensor outputs a string of data sequences with time synchronization, equal interval and numerical value normalization, generates standardized preprocessed data and sends the preprocessed data to a fusion processing and information comprehensive processing unit;

step three, the processing of multi-source sensor data fusion comprises two parts of track initiation and track association; the initial wave gate of the flight path is established by adopting a speed method, if three point paths fall into the related wave gates, the initial wave gate of the flight path is successful, the filtering initialization is carried out, and the point paths and the related wave gates of the flight path are established; the track association is to associate sonar information with radar, AIS and ESM data, if the sonar information is associated with any information, the target is considered as an overwater target, and if the sonar information is not associated with any information, the target is considered as an underwater target;

step four, the comprehensive content of the multi-source sensor data comprises the following three parts:

unmanned ship early warning of overturning: dividing the maximum rolling and pitching change rate into nine grades of 1-9 according to the maximum rolling angle and the maximum pitching angle which can be borne by the unmanned ship, wherein the grade 1 is basically free of overturning danger, and the grade 9 is the condition that the overturning danger is possibly existed greatly; judging the current corresponding overturning grade of the ship through the transverse and longitudinal rocking angles of the unmanned ship transmitted back by the inertial navigation system, and judging whether the ship has overturning danger;

sea state grade estimation: respectively collecting wind speed information of a weather instrument of the unmanned ship and roll angle information of inertial navigation, estimating a roll motion response function of the unmanned ship on line through the roll angle information, then performing a wave spectrum in a reverse mode, extracting wave height information from the wave spectrum, and jointly estimating the sea condition grade of the current sea area according to the relation between the wave grade and the wave height and the wind speed;

target hazard level and data confidence assignment: after the tracks are associated, according to the attributes of the targets, namely whether the targets are water targets or underwater targets, whether the targets have electronic countermeasure conditions or not and whether the targets transmit AIS information or not, the risk level of the targets is endowed; according to the data source, the target type, the correlation quality and the environmental factors of the target, comprehensively judging the confidence coefficient of the output target information according to different weight distribution;

step five, generating a comprehensive report;

and sequentially outputting the UTC time, the target batch number, the target type, the target direction, the direction unit, the target distance, the target strength, the target speed, the target threat level, the target information confidence coefficient, the sea condition information and the ship information obtained by the four steps.

Images