CN-122001446-A - Artificial intelligence-based motor emergency weather protection method

CN122001446ACN 122001446 ACN122001446 ACN 122001446ACN-122001446-A

Abstract

The invention provides an artificial intelligence-based maneuvering emergency weather guaranteeing method, which comprises the steps of calculating theoretical azimuth/pitching based on positioning and orientation data, an electronic compass gesture and target satellite parameters, determining a searchable angle range by combining antenna mechanical limit and dividing a sector search area, selecting a coarse scanning scheme by taking on-site satellite information, a candidate coarse scanning scheme and an algorithm state as input and adopting a contextual multi-arm decision algorithm, executing the sector coarse scanning according to the selected scheme, recording signal quality and angle, entering a fine scanning mode when an index meets a locking threshold value, acquiring a locking direction by linkage polarization angle, updating the algorithm state, generating a locking control instruction, enabling a vehicle-mounted antenna to be aligned with a target satellite and receive broadcast data, realizing self-adaptive and continuous satellite searching in a mechanical reachable range, matching a receiver sampling period, shortening satellite time, improving locking success rate and stability, and being suitable for quick access of vehicle-mounted emergency weather service.

Inventors

LUO CHAO
YAN PENG
CHEN JIANGYUN
SHEN YANG

Assignees

西北核技术研究所

Dates

Publication Date: 20260508
Application Date: 20260204

Claims (9)

1. An artificial intelligence-based motorized emergency weather protection method is characterized by comprising the following steps: s1, acquiring positioning and orientation data, electronic compass attitude data, target satellite parameters and algorithm state parameters, and calculating a theoretical azimuth angle and a theoretical pitch angle to obtain theoretical pointing information; S2, calculating a searchable angle range according to the theoretical pointing information and the mechanical limit parameters of the antenna, dividing a plurality of sector search areas in the searchable angle range, sequentially generating a rough scanning action scheme aiming at the sector search areas, wherein the rough scanning action scheme comprises a starting scanning angle, a scanning direction, a rotating step length and a stay time in a closed mode, meets the constraint of the mechanical limit parameters of the antenna, and forms site satellite alignment information by the theoretical pointing information, the searchable angle range and the sector search areas; S3, inputting field satellite alignment information, a rough scanning action scheme and algorithm state parameters into a context multi-arm decision algorithm, selecting to obtain the rough scanning action scheme, extracting a starting scanning angle, a scanning direction, a rotation step length and a stay time length from the rough scanning action scheme, generating a rough scanning control instruction, and sending the rough scanning control instruction to a rough scanning stage of a two-stage sector scanning and polarization angle linkage satellite alignment framework; s4, controlling the rough scanning stage to execute a rough scanning control instruction, collecting signal quality indexes output by a receiver, recording corresponding scanning angles, and obtaining a scanning result record; S5, when the signal quality index meets the locking threshold value based on the record of the scanning result, controlling the two-stage sector scanning and polarization angle linkage star framework to enter a fine scanning stage, carrying out joint adjustment on the pointing angle and the polarization angle to obtain locking pointing information, and updating algorithm state parameters based on the record of the scanning result; And S6, generating a locking control instruction according to the locking pointing information, and controlling the vehicle-mounted satellite receiving antenna to aim at the target satellite and receiving satellite broadcast data.
2. The artificial intelligence based motorized emergency weather protection method according to claim 1, wherein S1 is specifically: Acquiring positioning and orientation data, wherein the positioning and orientation data at least comprises current position information and heading reference information of a vehicle-mounted platform, and taking the positioning and orientation data as geographic reference input for calculating subsequent theoretical pointing information; acquiring electronic compass attitude data, wherein the electronic compass attitude data at least comprises a course angle, a pitch angle and a roll angle, and taking the electronic compass attitude data as an attitude input calculated by follow-up theoretical direction information; Acquiring target satellite parameters, wherein the target satellite parameters at least comprise orbit position description information of a target satellite, and calculating a theoretical azimuth angle and a theoretical pitch angle based on the positioning and orientation data and the target satellite parameters; And acquiring algorithm state parameters which at least comprise the selection times of each coarse scanning action scheme, the latest signal quality index and the latest locking judgment result, and packaging the theoretical azimuth angle, the theoretical pitch angle and the electronic compass attitude data together to form theoretical pointing information.
3. The artificial intelligence-based motorized emergency weather protection method according to claim 1, wherein S2 is specifically: According to the theoretical pointing information and the mechanical limiting parameters of the antenna, calculating the azimuth angle lower limit, the azimuth angle upper limit, the pitch angle lower limit and the pitch angle upper limit which can be actually rotated and covered by the vehicle-mounted satellite receiving antenna in the current parking leveling state, so as to form a searchable angle range; Mapping a theoretical azimuth angle and a theoretical pitch angle in the theoretical direction information into a searchable angle range to obtain a reference direction angle positioned in the boundary of the searchable angle range, wherein the reference direction angle is used for limiting the dividing direction of a subsequent sector search area; Judging the azimuth angle continuity according to the searchable angle range, splitting an azimuth angle interval crossing the zero-degree direction into a plurality of angle continuous intervals, and combining each angle continuous interval with a pitch angle interval to form a partitionable search area mother set; dividing a plurality of sector search areas in a search area mother set according to a preset azimuth angle span and a preset pitch angle span, determining a central azimuth angle, a central pitch angle, an azimuth angle span and a pitch angle span for each sector search area, and checking that the boundaries of the sector search areas do not exceed a searchable angle range; Generating a rough scanning action scheme for each sector searching area in sequence, taking the boundary of the sector searching area and the mechanical limit parameter of an antenna as constraint, determining the starting scanning angle of the rough scanning action scheme, determining the scanning direction of enabling a scanning track to cover the sector searching area, determining the rotation step length matched with the sampling period of the signal quality index of a receiver, and determining the stay time for stably sampling the signal quality index; Establishing a corresponding relation between a rough scanning action scheme and fan-shaped search area parameters for each fan-shaped search area, so that the rough scanning action scheme can be directly converted into a rough scanning control instruction parameter source under the condition of not introducing manual secondary setting; The theoretical pointing information, the searchable angle range and the plurality of sector search areas are combined to form site satellite-to-satellite information, and the site satellite-to-satellite information and each rough scanning action scheme together form an optional action space and a constraint space of the context multi-arm decision algorithm.
4. The artificial intelligence based maneuver emergency weather protection method as claimed in claim 3 wherein when dividing the plurality of sector search areas within the search area master set by a predetermined azimuth span, the center azimuth of the sector search area containing the reference pointing angle is calculated based on an azimuth function, wherein the azimuth function is specifically: ; Wherein, the For the center azimuth of the sector search area containing the reference azimuth, For the lower azimuth limit of the preferentially generated azimuth angle continuum, For the mapped reference azimuth angle, For a preset azimuth span, Is a downward rounding operation, and 、、 The same angle unit is adopted, and the vehicle-mounted controller is used for Starting from the continuous interval along the azimuth angle Generating central azimuth angles of other sector search areas for stepping and using the central azimuth angles in pitch angle direction To step at Generating a central pitch angle internally so as to determine a central azimuth angle, a central pitch angle, an azimuth angle span and a pitch angle span for each sector search area, and calculating a sector boundary and a searchable angle range according to the central angle and the span for each sector search area by the vehicle-mounted controller Performing boundary check when sector boundary exceeds At this time, the sector boundary is clipped to fall back to And (3) inner part.
5. The artificial intelligence based motorized emergency weather protection method according to claim 1, wherein S3 is specifically: the method comprises the steps of reading theoretical pointing information, a searchable angle range and fan-shaped search area parameters in site opposite star information, and reading candidate rough scanning action scheme parameters and algorithm state parameters to form an input set of a context multi-arm decision algorithm; Constructing a satellite alignment feature vector according to each rough scanning action scheme, and splicing a theoretical azimuth angle and a theoretical pitch angle, an azimuth angle lower limit pitch angle upper limit pitch angle of a searchable angle range, a course angle pitch angle roll angle in electronic compass attitude data, a center azimuth angle center pitch angle span of a fan-shaped search area, a rotational step length stay time of a scanning angle of the rough scanning action scheme, a last signal quality index in the last time of selection frequency in algorithm state parameters according to a fixed element sequence to form a twenty-dimensional satellite alignment feature vector; Inputting star feature vectors into a first branch of a selection value prediction neural network, forming ten three-dimensional first branch input by using sector search area parameters of electronic compass gesture data in a theoretical azimuth theoretical pitch angle searchable angle range, and sequentially passing through a first full-connection layer and a second full-connection layer to obtain sixteen-dimensional first branch feature vectors, wherein the first full-connection layer comprises thirty-two neurons, and the second full-connection layer comprises sixteen neurons; Inputting a star feature vector into a second branch of a selection value prediction neural network, inputting a coarse scanning action scheme parameter and an algorithm state parameter into a seven-dimensional second branch, and sequentially passing through a third full-connection layer and a fourth full-connection layer to obtain a sixteen-dimensional second branch feature vector, wherein the third full-connection layer comprises thirty-two neurons, and the fourth full-connection layer comprises sixteen neurons; Splicing the first branch feature vector and the second branch feature vector to form a thirty-two-dimensional fusion feature vector, and sequentially outputting one-dimensional selection values through a fifth full-connection layer and an output full-connection layer, wherein the fifth full-connection layer comprises sixteen neurons, and the output full-connection layer comprises one neuron; Repeatedly executing star feature vector construction and selection value output on all rough scanning action schemes to form a selection value sequence corresponding to the rough scanning action scheme, and determining the rough scanning action scheme according to the selection value sequence by adopting a selection value maximum principle; and extracting the rotation step length stay time of the scanning direction of the starting scanning angle from the rough scanning action scheme, generating a rough scanning control instruction and sending the rough scanning control instruction to a rough scanning stage of the two-stage sector scanning and polarization angle linkage star alignment framework.
6. The method for ensuring motor-driven emergency weather based on artificial intelligence according to claim 5, wherein when the rough scanning action scheme is determined by adopting a maximum selection value principle according to a selection value sequence, the selection times of the selection values corresponding to the rough scanning action schemes in the state parameters of the predictive selection value and the algorithm are calculated through a rough selection function, wherein the rough selection function is specifically ; Wherein, the To select the selection value for selecting the current coarse scanning operation scheme, Predicting the predicted selection value output by the neural network for the selection value, For the preset weight coefficient, the weight coefficient is set, Scheme for algorithm state parameter medium and rough sweeping action Corresponding number of selections, subscript For coarse scanning action scheme index, the vehicle-mounted controller is used for selecting a value sequence according to the selection The coarse sweeping action scheme is determined by adopting the principle of maximum selection value, specifically, the selection is carried out to ensure that Maximum coarse sweeping motion scheme As the scheme of the rough sweeping action, the subscript is as follows Indexing the selected coarse scanning action scheme.
7. The artificial intelligence based motorized emergency weather protection method of claim 1, wherein S4 is specifically Receiving a rough scanning control instruction, analyzing to obtain a starting scanning angle, a scanning direction, a rotating step length and a residence time length, and converting the parameters into an angle rotating sequence in a rough scanning stage; Controlling the rough scanning stage to drive the vehicle-mounted satellite receiving antenna to rotate from a starting scanning angle according to the angle rotation sequence, and keeping the stay time after each rotation step length is in place; Acquiring signal quality indexes output by a receiver in each stay time period, and synchronously acquiring a scanning angle corresponding to the signal quality indexes, wherein the scanning angle comprises a current azimuth angle and a current pitch angle; And combining the signal quality index and the corresponding scanning angle according to the acquisition sequence to form a scanning result record, and taking the scanning result record as input for locking threshold judgment and algorithm state parameter updating.
8. The artificial intelligence based motorized emergency weather protection method according to claim 1, wherein S5 is specifically: Correlating the signal quality index in the scanning result record with the corresponding scanning angle, selecting the maximum signal quality index and the corresponding scanning angle, and judging the maximum signal quality index and the locking threshold value; When the locking threshold is judged to be met, controlling the two-stage sector scanning and polarization angle linkage star alignment framework to enter a fine scanning stage, and setting the corresponding scanning angle as a fine scanning initial pointing angle; The fine scanning stage is controlled to adjust the pointing angle point by point around the fine scanning initial pointing angle according to the fine scanning step length, the polarization angle is synchronously adjusted at each pointing angle adjusting point to acquire signal quality indexes, and locking pointing information that the signal quality indexes meet locking threshold values is determined; And updating an algorithm state parameter based on the scanning result record and the judging result, adding one to the selection times of the coarse scanning action scheme, setting the latest signal quality index as the maximum signal quality index, and setting the latest locking judging result as the judging result.
9. The artificial intelligence based motorized emergency weather protection method according to claim 1, wherein step S6 is specifically: Reading locking pointing information, extracting a locking azimuth angle, a locking pitch angle and a locking polarization angle, converting the locking azimuth angle and the locking pitch angle into control angles under the current zero reference of the vehicle-mounted satellite receiving antenna, and simultaneously verifying that the control angles meet the mechanical limit parameters of the antenna and are positioned in a searchable angle range; Inputting the control angle and the locking polarization angle as target angles, and generating a locking control instruction, wherein the locking control instruction comprises a azimuth axis target angle, a pitching axis target angle, a polarization axis target angle, an in-place judging threshold value and an in-place maintaining mode in a closed manner; Controlling a vehicle-mounted satellite receiving antenna to execute a locking control instruction, sequentially driving a azimuth axis, a pitching axis and a polarization axis to rotate to corresponding target angles, judging in place based on an antenna feedback angle and an in-place judging threshold value, and entering a holding state corresponding to an in-place holding mode after in place; and controlling the receiver to enter a satellite broadcast data receiving mode in a holding state, and periodically collecting signal quality indexes for outputting satellite broadcast data when the locking threshold is confirmed to be continuously met.

Description

Artificial intelligence-based motor emergency weather protection method Technical Field The invention relates to the technical field of vehicle satellite communication and emergency weather guarantee, in particular to a maneuvering emergency weather guarantee method based on artificial intelligence. Background Vehicle-mounted emergency weather protection relies on rapid access of satellite broadcast data. After the vehicle is parked in the field, the target satellite is aligned according to the position, the heading and the gesture of the platform, and searching and locking are completed within the mechanical limit constraint. The field environment is changeable, the platform is difficult to be completely horizontal, the azimuth angle can cross zero degree, and the signal quality of the receiver can enter data receiving when meeting the threshold value, so the continuity and the executable of the star flow are critical. In the prior art, the theoretical azimuth and the pitching are usually calculated by using a positioning and electronic compass, the search boundary is determined by combining the mechanical limit of an antenna, the rough scanning is performed according to the preset step length and the stay time length, and the azimuth/pitching range is covered sequentially, which is usually the sector or the raster scanning with the theoretical direction as the center. When the signal quality reaches the threshold value, fine tuning is carried out, the direction is optimized by small step length, locking is completed, and the polarization angle is set after the partial scheme is locked. The fixed rules are adopted for selecting the scanning angle and the scanning direction, and the state transition is realized by threshold judgment. The method has the following defects that continuity of a zero-degree azimuth interval and sector cutting processing are imperfect, scanning track jump is easy to generate, coarse scanning step length and stay time are not matched with a sampling period of a receiver, signal judgment stability is affected, a starting scanning angle and scanning direction selection lack of context and historical state basis, fine scanning and polarization are not linked, and locking pointing is difficult to determine at one time. Thus, a motorized emergency weather protection method that addresses the above-described deficiencies of the prior art is a problem that one skilled in the art would need to address. Disclosure of Invention The application aims to provide an artificial intelligence-based maneuvering emergency weather guarantee method, which aims to solve the core technical problems that under a vehicle-mounted maneuvering emergency scene, the positioning orientation, the gesture and the mechanical limit are unified into a site satellite-to-satellite context, a searchable range and a sector search area are automatically divided, and a feasible rough and fine scanning flow is intelligently selected and executed under zero-crossing orientation and polarization linkage constraint, so that the satellite-to-satellite process can be directly executed on site and stably locked. According to the embodiment of the invention, the manual intelligent-based maneuvering emergency weather protection method comprises the following steps of: s1, acquiring positioning and orientation data, electronic compass attitude data, target satellite parameters and algorithm state parameters, and calculating a theoretical azimuth angle and a theoretical pitch angle to obtain theoretical pointing information; S2, calculating a searchable angle range according to the theoretical pointing information and the mechanical limit parameters of the antenna, dividing a plurality of sector search areas in the searchable angle range, sequentially generating a rough scanning action scheme aiming at the sector search areas, wherein the rough scanning action scheme comprises a starting scanning angle, a scanning direction, a rotating step length and a stay time in a closed mode, meets the constraint of the mechanical limit parameters of the antenna, and forms site satellite alignment information by the theoretical pointing information, the searchable angle range and the sector search areas; S3, inputting field satellite alignment information, a rough scanning action scheme and algorithm state parameters into a context multi-arm decision algorithm, selecting to obtain the rough scanning action scheme, extracting a starting scanning angle, a scanning direction, a rotation step length and a stay time length from the rough scanning action scheme, generating a rough scanning control instruction, and sending the rough scanning control instruction to a rough scanning stage of a two-stage sector scanning and polarization angle linkage satellite alignment framework; s4, controlling the rough scanning stage to execute a rough scanning control instruction, collecting signal quality indexes output by a receiver, recording corresponding scanni