CN116540287B

CN116540287B - Intelligent navigation method, intelligent navigation device, computer equipment and storage medium

Info

Publication number: CN116540287B
Application number: CN202310385539.2A
Authority: CN
Inventors: 朱祥维; 宋江波; 刘若帆; 李婉清; 何少烽; 伊瑞特
Original assignee: Sun Yat Sen University; Sun Yat Sen University Shenzhen Campus
Current assignee: Sun Yat Sen University; Sun Yat Sen University Shenzhen Campus
Priority date: 2023-04-11
Filing date: 2023-04-11
Publication date: 2024-03-19
Anticipated expiration: 2043-04-11
Also published as: CN116540287A

Abstract

The application belongs to the technical field of intelligent navigation, and discloses an intelligent navigation method, an intelligent navigation device, computer equipment and a storage medium, wherein the method comprises the following steps: constructing a sliding window, defining a sliding window threshold value, and constructing an initial variable node and a priori factor node; according to the current time t _k Constructing a first variable node according to the inertia estimation state value of (2) and the next time t _k+1 Constructing a second variable node by the inertia state predicted value of the first variable node; according to the next time t _k+1 The inertia estimation state value and the inertia state prediction value of the model (1) construct a current factor node; constructing a corresponding additional factor node according to the measurement information of the additional sensor to obtain a factor graph to be optimized; when the number of variable nodes in the sliding window reaches the threshold value of the sliding window, the variable nodes and factor nodes in the sliding window are optimized and then calculated, and an optimal navigation result is obtained. The method and the device can solve the problem of heterogeneous system frequency, and achieve the effect of improving the accuracy of navigation results.

Description

Intelligent navigation method, intelligent navigation device, computer equipment and storage medium

Technical Field

The present disclosure relates to the field of intelligent navigation technologies, and in particular, to an intelligent navigation method, an intelligent navigation device, a computer device, and a storage medium.

Background

With the increase of human ocean exploration activities, the requirements on the navigation positioning capability of autonomous underwater vehicles (Autonomous Underwater Vehicle, AUV) are higher and higher, and the multi-sensor fusion navigation positioning technology becomes a hot spot and trend of research. As the number and variety of sensors installed and on the AUV increases, how to effectively fuse the data of each sensor has been a hotspot and difficulty of research. Conventional fusion methods based on filtering are centralized kalman filtering (Centralized Kalman Filter, CKF), federal kalman filtering (Federal Kalman Filter, FKF). When the number of the fused sensors is increased, the state dimension and the calculated amount of CKF are correspondingly increased, so that the filter is unstable and easy to diverge; compared with CKF, FKF based on distributed structure has the advantages of good fault tolerance, small calculated amount, flexible design and the like, and FKF and related improved algorithm are widely applied in underwater multisource information fusion technology.

However, the information update frequencies of the different sensors are often not synchronized, i.e. the inter-frequency heterogeneous problem. The FKF information fusion mechanism needs to be performed when measurement information of all sensors is received. Therefore, the accuracy and real-time performance of the FKF-based integrated navigation system may be affected. In addition, the working states of different sensors are uncertain, the sensors are unstable or fail in certain scenes, and the FKF method can face the problems of information faults, system reconstruction after information source change and the like.

Therefore, the existing underwater multisensor combination navigation system has the problem of heterogeneous frequencies, and the underwater multisensor data fusion algorithm applied to the navigation system has the problem of low accuracy of calculated navigation results.

Disclosure of Invention

The application provides an intelligent navigation method, an intelligent navigation device, computer equipment and a storage medium, which can solve the problem of different frequency isomerism in the existing navigation system and improve the calculation precision of navigation results.

In a first aspect, an embodiment of the present application provides an intelligent navigation method, where the method includes:

constructing a sliding window, defining a sliding window threshold value, constructing an initial variable node and a priori factor node according to pose information of an aircraft at an initial moment, and connecting the initial variable node and the priori factor node to obtain an initialization factor graph;

according to the acquired current moment t of the aircraft _k (k∈N ^* ) Inertial measurement value of (c) and current time t _k The inertia estimated state value of (2) gets the next time t _k+1 According to the current time t _k Constructing a first variable node according to the inertia estimation state value of (2) and the next time t _k+1 Constructing a second variable node by the inertia state predicted value of the first variable node; if k=1, the first variable node replaces the initial variable node in the initial factor graph and is connected with the prior factor node;

According to the acquired next time t of the aircraft _k+1 Inertial estimated state value of (2) and next time t _k+1 Constructing a current factor node by using the inertia state predicted value of (1); the first variable node and the second variable node are respectively connected with the current factor node;

detecting the current time t _k Whether additional sensor measurement information is received, if at the current time t _k When additional sensor measurement information is received, constructing corresponding additional factor nodes according to the additional sensor measurement information, and independently connecting the additional factor nodes with the first variable nodes to obtain a factor graph to be optimized;

and the sliding window slides on the factor graph to be optimized according to a preset rule until the number of variable nodes in the factor graph to be optimized in the sliding window reaches a sliding window threshold value, and the variable nodes in the sliding window and factor nodes connected with the variable nodes are calculated after being optimized to obtain an optimal navigation result of the aircraft.

Further, according to the acquired current time t of the aircraft _k Inertial measurement of (2)Value and current time t _k The inertia estimated state value of (2) gets the next time t _k+1 An inertial state prediction value of (1), comprising:

according to the acquired current moment t of the aircraft _k Inertial measurement value of (c) and current time t _k The inertia estimation state value of (2) is calculated based on the inertia state transition equation to obtain the next time t _k+1 Is a predicted value of the inertial state of the vehicle.

Further, according to the acquired next time t of the aircraft _k+1 Inertial estimated state value of (2) and next time t _k+1 The inertia state predicted value of (1) constructs a current factor node, comprising: according to the acquired next time t of the aircraft _k+1 Inertial estimated state value of (2) and next time t _k+1 The inertia state predictor of (2) constructs the current factor node based on the residual function.

Further, the additional sensor measurement information includes: long baseline positioning information, doppler velocity measurement information, magnetic compass information, pressure sensing information and navigation satellite positioning information; the additional factor node includes: long baseline factor node, doppler velocity factor node, magnetic compass factor node, pressure sensing factor node, and navigation satellite factor node.

Further, the method further comprises: when the detection time of the sliding window on the number of variable nodes in the sliding window reaches a preset detection time threshold, the variable nodes in the sliding window and factor nodes connected with the variable nodes are optimized and then calculated, and an optimal navigation result of the aircraft is obtained.

Further, the method further comprises: when the number of variable nodes in the factor graph to be optimized in the sliding window reaches a sliding window threshold value, or when the detection time of the sliding window to the number of variable nodes in the sliding window reaches a preset detection time threshold value, the part, which is positioned outside the sliding window, of the factor graph to be optimized is marginalized by adopting a ShuerBu method.

Further, the optimizing and calculating the variable nodes in the sliding window and the factor nodes connected with the variable nodes to obtain the optimal navigation result of the aircraft comprises the following steps: and optimizing variable nodes in the sliding window and factor nodes connected with the variable nodes by adopting a Gaussian-Newton iteration method or a Levenberg-Marquardt method, and then calculating to obtain an optimal navigation result of the aircraft.

Further, the method further comprises:

detecting whether the received long baseline positioning information comes from a preset number or more of long baseline positioning sensors;

when the received long baseline positioning information is detected to come from a preset number of long baseline positioning sensors or more, constructing a long baseline factor node by adopting a long baseline loose coupling method according to the long baseline positioning information;

When the received long baseline positioning information is detected to come from the long baseline positioning sensors with the number below a preset number, a long baseline factor node is constructed by adopting a long baseline tight coupling method according to the long baseline positioning information.

Further, the constructing a long baseline factor node by adopting the long baseline loose coupling method includes: and differencing the long baseline positioning information and the position information calculated by the strapdown inertial navigation system to obtain long baseline measurement information, and constructing a long baseline factor node according to the long baseline measurement information.

Further, the construction of the long baseline factor node by the long baseline tight coupling method includes: and making a difference between the long baseline positioning information and the slope distance difference in the strapdown inertial navigation system to obtain long baseline measurement information, and constructing a long baseline factor node according to the long baseline measurement information.

Further, the method further comprises:

detecting whether the received navigation satellite positioning information comes from a preset number or more of navigation satellites;

when the received navigation satellite positioning information is detected to come from a preset number of navigation satellites or more, constructing navigation satellite factor nodes by adopting a navigation loose coupling method according to the navigation satellite positioning information;

When the received navigation satellite positioning information is detected to come from navigation satellites with the number below a preset value, constructing a navigation satellite factor node by adopting a navigation tight coupling method according to the navigation satellite positioning information.

Further, the method for constructing the navigation satellite factor node by adopting the navigation loose coupling method comprises the following steps: and (3) differencing the navigation satellite positioning information and the position information calculated by the strapdown inertial navigation system to obtain navigation satellite measurement information, and constructing a navigation satellite factor node according to the navigation satellite measurement information.

Further, the method for constructing the navigation satellite factor node by adopting the navigation tight coupling method comprises the following steps: and taking pseudo-range information in the navigation satellite positioning information as navigation satellite measurement information, and constructing a navigation satellite factor node according to the navigation satellite measurement information.

In a second aspect, an embodiment of the present application provides an intelligent apparatus, where the apparatus is applied to an intelligent navigation device, the apparatus includes:

the initialization module is used for constructing a sliding window, defining a sliding window threshold value, constructing an initial variable node and a priori factor node according to pose information of the initial moment of the aircraft, and connecting the initial variable node and the priori factor node to obtain an initialization factor graph;

The variable node construction module is used for constructing a variable node according to the acquired current moment t of the aircraft _k (k∈N ^* ) Inertial measurement value of (c) and current time t _k The inertia estimated state value of (2) gets the next time t _k+1 According to the current time t _k Constructing a first variable node according to the inertia estimation state value of (2) and the next time t _k+1 Constructing a second variable node by the inertia state predicted value of the first variable node; the variable node construction module is further used for replacing the initial variable node in the initial factor graph with the first variable node when k=1, and connecting the first variable node with the prior factor node;

a factor node construction module for constructing a factor node according to the acquired next time t of the aircraft _k+1 Inertial estimated state value of (2) and next time t _k+1 Constructing a current factor node by using the inertia state predicted value of (1); the first variable node and the second variable node are respectively connected with the current factor node;

an additional node construction module for detecting the current time t _k Whether additional sensor measurement information is received, if at the current time t _k When additional sensor measurement information is received, constructing corresponding additional factor nodes according to the additional sensor measurement information, and independently connecting the additional factor nodes with the first variable nodes to obtain a factor graph to be optimized;

The navigation result calculation module is used for enabling the sliding window to slide on the factor graph to be optimized according to a preset rule until the number of variable nodes in the factor graph to be optimized in the sliding window reaches a sliding window threshold value, and calculating after optimizing the variable nodes in the sliding window and factor nodes connected with the variable nodes to obtain an optimal navigation result of the aircraft.

In a third aspect, embodiments of the present application provide a computer device including a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor executing the steps of the intelligent navigation method of any of the embodiments described above when the computer program is executed.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the intelligent navigation method of any of the embodiments described above.

In summary, compared with the prior art, the technical scheme provided by the embodiment of the application has the beneficial effects that at least:

according to the intelligent navigation method, the measurement information obtained by various sensors at various moments is converted into the factor nodes and the variable nodes which are placed in the constructed factor graph, so that different frequency fusion among different sensors is realized, and the problem of different frequency isomerism in the navigation system in the prior art is solved; meanwhile, the sliding window method is adopted when the factor graph is optimized to calculate and solve the navigation result, so that the calculated amount is reduced to a certain extent when the navigation result is calculated, the real-time performance of calculated data is ensured by calculating the sensor information received at the latest moment, and the calculation precision of the navigation result is improved.

Drawings

Fig. 1 is a flowchart of an intelligent navigation method according to an exemplary embodiment of the present application.

Fig. 2 is a schematic diagram of a factor graph architecture provided in an exemplary embodiment of the present application.

Fig. 3 is a flowchart of a long baseline positioning information adaptive coupling step provided in an exemplary embodiment of the present application.

Fig. 4 is a flowchart of a navigation satellite positioning information adaptive coupling step according to an exemplary embodiment of the present application.

Fig. 5 is a block diagram of an intelligent navigation device according to an exemplary embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.

Referring to fig. 1, an embodiment of the present application provides an intelligent navigation method, which is applied to an intelligent navigation device, and is described by taking an execution subject as an example of the intelligent navigation device, and the method specifically includes the following steps:

Step S1, constructing a sliding window, defining a sliding window threshold, constructing an initial variable node and a priori factor node according to pose information of an aircraft at an initial moment, and connecting the initial variable node and the priori factor node to obtain an initialization factor graph.

Wherein, the factor graph is expressed as G= (F, X, E) as a bipartite graph model, and describes the joint probability distribution of random variables. Wherein the node class is divided into factor nodes f _i E F and variable node x _j Two classes, E and X, wherein the former is a local function in factorization, and the latter is a variable in a global multi-element function; if and only if the factor node f _i And variable node x _j There is a connecting edge when related, a watchShown as e _ij E, edges typically act as constraints for variable nodes.

Thus, the definition of factor graph G can be described as factoring f (X): f (X) = pi _i f _i (x _i ) X represents a group comprising a group X _i A subset of the total set of variable nodes involved. Factor node f _i (x _i ) Can use a residual function f _i (x _i )＝d(err _i (x _i ,z _i ) Described, where d (·) is the corresponding residual function (also called cost function), z _i Is the actual measurement value.

Because the information fusion problem of an AUV multi-source navigation system has many similarities to a factor graph, the factor graph can be introduced therein. AUV at t _i The navigation state of the moment is described as a variable node x _i And is defined at t _k Time of day set containing state variable information of all past time of dayDefinition t _k The set of all measurement information in time +.>Wherein z is _i Representing t _i Actual measurement information obtained by different navigation sensors at the moment. The joint probability density (Probability Distribution Function, PDF) and factoring thereof, which is defined to yield all state variables and measurements, can be expressed as:

wherein P (x) ₀ ) The information in the node is the prior factor of the initial moment;representing a measurement model->A set of variable nodes therein; />Representing a process model, wherein->To distinguish inertial measurements from inertial sensor assemblies IMUs (Inertial Measurement Unit, inertial measurement units).

In the joint probability density function P (X _k |Z _k ) Each factor represents an independent term, i.e

For Gaussian noise distribution, we can factor f each representative of the metrology model _i Expressed as a residual function:

each factor f representing a process model _i Expressed as a residual function:

wherein operator is represented by square mahalanobis distance, Σ is a measurement noise covariance matrix, Λ is a process noise covariance matrix, h _i (. Cndot.) is a nonlinear measurement function of the system, f _i (. Cndot.) is a nonlinear state transfer function of the system.

The Maximum posterior probability density (Maximuma Posteriori, MAP) estimation converts the AUV multisensor information problem to an equivalent nonlinear optimization problem. At this time, the maximum a posteriori probability estimated density X of all the state variables can be obtained for the minimum value that can take the negative logarithm ^* ：

In order to simplify the operation, the information P (x ₀ ). The above formula represents a least squares problem which can be solved iteratively according to nonlinear optimization theory by adjusting the variable X _k Minimizing it, then this optimal estimateThe overall residual function f (X) can be minimized. The above formula is a general representation method, and the residual function can represent positioning means such as inertial navigation, LBL (Long BaseLine), GNSS and the like.

Specifically, initial parameters, namely initial variable nodes and prior factor nodes, are set according to the initial moment pose information of an AUV (Autonomous Underwater Vehicle ), a sliding window is constructed, a sliding window threshold is defined, and an initial variable node X is obtained ₀ = { } and a priori factor node f _prior The initialization is completed by adding = { } to the initialization factor table.

Step S2, according to the acquired current moment t of the aircraft _k (k∈N ^* ) Inertial measurement value of (c) and current time t _k The inertia estimated state value of (2) gets the next time t _k+1 According to the current time t _k Constructing a first variable node according to the inertia estimation state value of (2) and the next time t _k+1 Constructing a second variable node by the inertia state predicted value of the first variable node; if k=1, the first variable node replaces the initial variable node in the initial factor graph and is connected with the prior factor node.

Wherein the inertial measurements of the aircraft include specific forces and angular rates of inertial sensors on the aircraft.

Specifically, if at t _k The IMU measurements (specific force and angular rate) are obtained at the moment, at which point the factor node f _IMU Connection t _k ,t _k+1 Time variable node X _k ,X _k+1 T can be obtained from IMU state transfer equation _k+1 State prediction value of timeIf no other sensor information is present at this time, +.>Can be used as the state variable +.>Added to variable node X _k+1 = { }. It is generally believed that IMU measurements can be obtained at all times during the navigation process.

In the concrete implementation process, k is E N ^* ，N ^* Is a positive integer, i.e. from the construction of the initialization factor graph, the IMU factor node and the variable node at corresponding moments are added to the factor graph at each moment in the following, so that the node information in the factor graph is gradually increased along with the time, when k=1, i.e. IMU measurement information is received for the first time after the construction of the initialization factor graph, and the variable node X constructed according to the measurement information ₁ Then replaces the original empty initial variable node X ₀ And is connected with the prior factor node.

Step S3, according to the acquired next time t of the aircraft _k+1 Inertial estimated state value of (2) and next time t _k+1 Constructing a current factor node by using the inertia state predicted value of (1); the first variable node and the second variable node are respectively connected with the current factor node.

Specifically, at the current time t _k Receiving measurements from an IMUAfter that, the current factor node f is defined _IMU . The current factor node of the IMU is connected with the current time t _k And next time t _k+1 Is a variable node X of (2) _k And X _k+1 . The discretized form of the time update of the navigation state can be expressed as +.>At the current time t _k Inertial measurement of (2)And an inertia estimated state value x _k Is used for predicting the next time t _k+1 Inertia state prediction value +.>This prediction is associated with the next time t _k+1 Inertial estimated state value x received by aircraft _k+1 The difference between the estimates is the residual function represented by the current factor node:

step S4, detecting the current time t _k Whether additional sensor measurement information is received, if at the current time t _k And when receiving the measurement information of the additional sensor, constructing a corresponding additional factor node according to the measurement information of the additional sensor, and independently connecting the additional factor node with the first variable node to obtain a factor graph to be optimized.

Wherein the additional sensor measurement information includes: long Baseline (LBL) positioning information, doppler (Doppler Velocity Log, DVL) velocity information, magnetic compass (Magnetic Compass Pilot, MCP) heading angle information, pressure Sensor (PS) depth information, GNSS global navigation satellite positioning information, and the like. Correspondingly, the additional factor node comprises: LBL factor node, DVL factor node, MCP factor node, PS factor node, and GNSS factor node.

Specifically, referring to fig. 2, if measurement information measured by sensors such as LBL position information, DVL speed information, MCP heading angle information, PS depth information, etc. is acquired at certain times, corresponding factor nodes are constructed and added to the factor graph; the factor node is a square f in the graph, the circle x is a variable node, and the priority is a priori factor node; the correspondence of variable nodes to factor nodes depends on which sensor measurements the current aircraft state is related to, if at the current time t _k Without the addition of additional sensor measurement information, the current time t _k Is constructed by only IMU measurement information, if additional sensor measurement information is added, the current time t _k Is constructed by IMU inertial measurement information and other sensor measurement information. At the current time t _k Measured value of IMU and current time t _k And next time t _k+1 The variable nodes at two moments are related, and the measurement information of other additional sensors is only related to the current moment t _k Is related to the variable node of (c).

And S5, sliding the sliding window on the factor graph to be optimized according to a preset rule until the number of variable nodes in the factor graph to be optimized in the sliding window reaches a sliding window threshold value, and calculating after optimizing the variable nodes in the sliding window and factor nodes connected with the variable nodes to obtain an optimal navigation result of the aircraft.

The preset rule may be to slide on the factor graph at a certain preset speed, for example, a distance of M variable nodes sliding forward every minute, and the specific value of M or the preset speed may be set manually.

Specifically, when the system is in long running time, variable nodes in a factor graph to be optimized can be increased sharply along with time, and in order to balance the calculated amount of the system and ensure the real-time performance of the system, the application adopts the idea of sliding window. The number of variable nodes in the factor graph to be optimized is limited by a fixed size, and when a window slides, new variable nodes and factor nodes are added continuously and old variable nodes and old factor nodes are removed.

In order to utilize past state and measurement information as much as possible, the factor nodes and variable nodes removed from the window are subjected to edge processing, and the past information is reserved in the calculation of the optimal navigation result through a Shuerbu method.

The IMU state of the factor graph in the sliding window is x= [ X ₁ ,…, _n ]Wherein n is the maximum number of states in the sliding window, _k is the kth IMU state in the sliding window. And obtaining the initial position of the aircraft according to the GNSS, constructing a priori factor node according to the initial position information at the initial moment, updating the variable node and the factor node through the received IMU measurement information, and simultaneously detecting whether measurement information from other sensors exists.

The number of variable nodes in the sliding window is used for navigation result calculation if the sliding window threshold is reached. And if the size of the sliding window threshold is N, calculating the optimal navigation result of factor graph optimization when the number of variable nodes in the sliding window reaches the threshold N. Within the sliding window, assume at some point the IMU state x _i The corresponding factor node is connected to the related variable node and IMU factor node f when other available sensor information, namely LBL position information, DVL speed information, MCP course angle information, PS depth information and other measured values measured by the sensors are sequentially detected _IMU As a binary edge, is connected to two adjacent variable nodes associated therewith. Variable x for all moments in a sliding window _i And the set of (2) is the variable to be optimized.

As the running time of the system in practice is increased, the amount to be optimized is increased, the updating can not be performed every moment, and the updating and optimizing of all variable nodes at one time are difficult. Thus, a sliding window method must be used to control the scale of the computation.

The factor graph method is applied to the coding field at the earliest, and is gradually paid attention to other fields such as artificial intelligence, signal processing and the like as a probability graph model. The factor graph-based optimization method is also very widely mature in its application to SLAM problems. The map optimization also provides a new idea for information fusion of the navigation system: at some point, when a sensor is available, its measurement information can be represented as a corresponding factor node, and the state quantity to be estimated is represented as a variable node. The factor nodes provide state constraints for the associated variable nodes. The whole factor graph is recursively and updated according to a nonlinear optimization theory, so that different frequencies of different sensors can be fused, and the plug-and-play of different types of underwater navigation sensors can be met. In order to reduce the complexity of system calculation, the method based on the sliding window idea is adopted in the optimization process of the factor graph, so that the system calculation amount and the accuracy are balanced to a certain extent.

According to the intelligent navigation method provided by the embodiment, the sensor measurement information received by the aircraft at a plurality of moments is converted into the factor nodes and the variable nodes which are placed in the constructed factor graph, so that the different frequency fusion among different sensors is realized, and the different frequency heterogeneous problem of the navigation system in the prior art is solved; meanwhile, the sliding window method is adopted when the factor graph is optimized to calculate and solve the navigation result, so that the calculation amount of calculating the navigation result is reduced to a certain extent, the real-time performance of calculated data is ensured by calculating the sensor information received at the latest moment, and the calculation precision of the navigation result is improved.

In some embodiments, the step S2 is based on the acquired current time t of the aircraft _k Inertial measurement value of (c) and current time t _k The inertia estimated state value of (2) gets the next time t _k+1 An inertial state prediction value of (1), comprising:

Wherein the inertial state transition equation is x _k For the current time t _k Inertia estimated state value of ∈j->For the current time t _k Inertia measurement of>For the next time t _k+1 Is a predicted value of the inertial state of the vehicle.

The above embodiment can be quickly based on the current time t _k The inertia measurement information of (a) gets the next time t _k+1 To establish its current time t _k And the next time t _k+1 Is a second variable node of (a).

In some embodiments, the following time t of the acquired aircraft in step S3 _k+1 Inertial estimated state value of (2) and next time t _k+1 The inertial state prediction value of (1) constructs the current factor node, and may further include:

according to the acquired next time t of the aircraft _k+1 Inertial estimated state value of (2) and next time t _k+1 The inertia state predictor of (2) constructs the current factor node based on the residual function.

The residual function is specifically:

wherein x is _k+1 For the next time t _k+1 Is set to the inertia estimated state value of (a),for the next time t _k+1 Is x _k For the current time t _k Inertia estimated state value of ∈j->For the current time t _k Inertial measurement of (a).

In the above embodiment, the current time is represented by the residual functionEngraving t _k For the next time t _k+1 Predicted value of inertial state of (c) and t at next time _k+1 The difference between the inertial estimated state values actually received by the aircraft, so as to be able to determine the current instant t _k Variable node of (2) and the next time t constructed with inertia state prediction value _k+1 Is connected to the variable nodes of (a).

In some embodiments, the additional sensor measurement information includes: long baseline positioning information, doppler velocity measurement information, magnetic compass information, pressure sensing information and navigation satellite positioning information; the additional factor node includes: long baseline factor node, doppler velocity factor node, magnetic compass factor node, pressure sensing factor node, and navigation satellite factor node.

The long baseline positioning information is positioning information sent by an underwater long baseline positioning system, the Doppler velocity measurement information is velocity information measured by a Doppler log, the magnetic compass information can be specifically magnetic compass heading angle information, the pressure sensing information can be specifically pressure sensor depth information, and the navigation satellite positioning information can be specifically GNSS global navigation satellite positioning information.

In a specific implementation process, a method for constructing a doppler velocimetry node, i.e., a DVL factor node, in the additional factor nodes is as follows, and firstly, a DVL measurement equation may be expressed as follows:

Wherein n is ^DVL Is to measure noise, h ^DVL Is a measurement function.

At t _m After receiving DVL measurement information at any time, adding a new DVL factor node f _DVL . The DVL factor node is only connected with the variable node x _m I.e. t _m The navigational state of the moment of time is relevant. Thus, the factor node f can be _DVL Defined as a unitary edge factor:

similarly, the magnetic compass factor node in the additional factor nodes, namely the MCP factor node, is constructed by the following steps:

at t _s And when the speed measurement information of the MCP is received at the moment, constructing MCP factor nodes, and adding the MCP factor nodes into a factor graph frame. The MCP measurement equation can be expressed as:

wherein n is ^MCP Measurement noise, h, which is a measurement equation of MCP ^MCP ( _s ) Is a measurement function.

MCP factor node is only connected with t _s Time variable node x _s The correlation, i.e. only one edge is connected between them, the MCP measurement information acts as a constraint. f (f) _MCP Is a unitary edge, namely:

the construction method of the pressure sensing factor node, namely the PS factor node comprises the following steps: at t _l And when elevation measurement information of PS is received at any time, constructing PS factor nodes, and adding the PS factor nodes into a factor graph frame. The PS metrology equation can be expressed as:

wherein n is ^PS Measurement noise, h, which is the measurement equation for PS ^PS ( _l ) Is a measurement function. PS factor node is only connected with t _l Time variable node x _l The PS measurement information is used as a constraint in relation, i.e. there is only one edge between them. f (f) _PS Is a unitary edge, namely:

over time, the available measurement information, such as IMU, LBL, DVL, MCP, PS, GNSS, is received at other times, corresponding additional factor nodes can be expanded to factor graphs, and variable nodes can be recursively and updated according to measurement equations of different sensors and by utilizing corresponding error function optimization solutions. When the availability of the sensor is changed, seamless switching of the working mode of the AUV integrated navigation system can be realized by deleting or adding the factor nodes in the factor graph model.

According to the embodiment, whether other sensor information is received or not is detected at each moment, and the detected other sensor information is added into the factor graph algorithm, so that different frequency fusion among different sensors is realized, the problem of different frequency isomerism in a navigation system in the prior art is solved, and meanwhile, the addition of multiple sensor information enables more reference information and more accurate calculation of an optimal navigation result when the factor graph in a sliding window is optimized.

In some embodiments, the above method may further comprise: when the detection time of the sliding window on the number of variable nodes in the sliding window reaches a preset detection time threshold, the variable nodes in the sliding window and factor nodes connected with the variable nodes are optimized and then calculated, and an optimal navigation result of the aircraft is obtained.

Specifically, when the detection time reaches the time threshold T, even if the number of variable nodes in the sliding window at this time is smaller than the sliding window threshold, the node information within the sliding window at this time is used to calculate the optimal navigation result. Because the intelligent navigation device needs new data at this time to correct the history navigation information, even if the number of nodes is not enough, the optimization calculation is performed.

According to the embodiment, the problem that navigation data cannot be updated when the number of nodes does not meet the sliding window threshold is solved by increasing the detection time threshold, the navigation data can be updated when a certain preset time is reached, the navigation data is updated in time, and the situation that intelligent navigation cannot be achieved due to the fact that the navigation data cannot be received for a sufficient number of sensor measurement information is avoided.

In some embodiments, the method may further comprise: when the number of variable nodes in the factor graph to be optimized in the sliding window reaches a sliding window threshold value, or when the detection time of the sliding window to the number of variable nodes in the sliding window reaches a preset detection time threshold value, the part, which is positioned outside the sliding window, of the factor graph to be optimized is marginalized by adopting a ShuerBu method.

The method of the Shu's complement is a process of changing a matrix into a diagonal matrix, and in marginalization, the Shu's complement method converts node information outside a sliding window as constraint information into prior distribution of variables to be optimized, and is actually a problem of obtaining probability distribution of a variable subset from joint distribution.

The embodiment fully utilizes the state and the measurement information of the sensor at the past moment, optimizes the node information at the latest moment in the sliding window, and simultaneously considers the sensor information at the past moment as much as possible, thereby further improving the precision of the optimal navigation result.

In some embodiments, in step S5, the optimizing the variable node in the sliding window and the factor node connected to the variable node to calculate to obtain the optimal navigation result of the aircraft may include: and optimizing variable nodes in the sliding window and factor nodes connected with the variable nodes by adopting a Gaussian-Newton iteration method or a Levenberg-Marquardt method, and then calculating to obtain an optimal navigation result of the aircraft.

The Gaussian-Newton iteration method (Gauss-Newton iteration method, GNA) is an iteration method for solving regression parameters in a nonlinear regression model to carry out least square, the method uses a Taylor series expansion type to approximately replace the nonlinear regression model, then the regression coefficient is corrected for multiple times through multiple iterations, so that the regression coefficient is continuously approximate to the optimal regression coefficient of the nonlinear regression model, and finally the sum of squares of residual errors of an original model is minimized.

The Levenberg-Marquardt method (Levenberg-Marquardt algorithm, LMA or LM for short) is also known in mathematics and computing as the damped least squares method (DLS) for solving the nonlinear least squares problem. These minimization problems occur especially in least squares curve fitting. LMA interpolates between gaussian-newton iteration (GNA) and gradient descent. LMA is more robust than GNA, which means that in many cases it can find a solution even if it starts far from the final minimum. LMAs tend to be slower than GNAs for well performing functions and reasonable startup parameters. LMA can also be considered as a gaussian-newton iteration method using a trusted domain method.

LMAs are used in many software applications to solve the general curve fitting problem. By using the gaussian-newton algorithm, it generally converges faster than the first order method.

In the embodiment, the node information is optimized and calculated by adopting the Gaussian-Newton iteration method or the Levenberg-Marquardt method, and compared with the Newton method for solving the least square problem, the two methods have smaller calculation amount, and the calculation amount and the power consumption of the optimal navigation result in the application are reduced to a certain extent.

In some embodiments, referring to fig. 3, the method may further comprise the steps of:

step S61, detecting whether the received long baseline positioning information is from a preset number of long baseline positioning sensors or more.

Step S62, when the received long baseline positioning information is detected to come from a preset number of long baseline positioning sensors and above, a long baseline factor node is constructed by adopting a long baseline loose coupling method according to the long baseline positioning information.

Step S63, when the received long baseline positioning information is detected to come from the long baseline positioning sensors with the number smaller than the preset number, a long baseline factor node is constructed by adopting a long baseline tight coupling method according to the long baseline positioning information.

Wherein the preset number is preferably 4.

Specifically, assuming that LBL long baseline positioning information is received at time k, setting n as the number of received long baseline positioning sensors, presetting the number as 4, and constructing f _LBL Factor node:

the embodiment solves the problems that the received long baseline positioning information has instability and uncertainty due to the fact that the number of the long baseline positioning sensors is large, and the method for adaptively coupling the long baseline positioning information not only improves the robustness of the intelligent navigation system, but also reduces the calculated amount and the power consumption of the optimal navigation result.

In some embodiments, step S62 may specifically include: and differencing the long baseline positioning information and the position information calculated by the strapdown inertial navigation system to obtain long baseline measurement information, and constructing a long baseline factor node according to the long baseline measurement information.

In some embodiments, step S63 may specifically include: and making a difference between the long baseline positioning information and the slope distance difference in the strapdown inertial navigation system to obtain long baseline measurement information, and constructing a long baseline factor node according to the long baseline measurement information.

In the specific implementation process, at t _n When the measurement information of the LBL is received at the moment, an LBL factor node is constructed and added into a factor graph frame. The LBL participation fusion mode is divided into long baseline loose coupling and long baseline tight coupling, wherein the long baseline loose coupling means that the measurement information input by the LBL is obtained by differencing long baseline positioning information and position information calculated by a strapdown inertial navigation system; the long baseline tight coupling is obtained by taking the difference between the measurement information input by the LBL and the slope distance difference in the strapdown inertial navigation system. The strapdown inertial navigation system (Strap-down Inertial Navigation System, SINS) is an inertial navigation system which directly connects an inertial instrument to a carrier and completes the function of a navigation platform by a computer, so that position information and skew difference are directly calculated by the system, and the application does not describe the position information and the skew difference in detail.

The number of the received long baseline positioning sensors is detected through the self-adaptive coupling method, whether the long baseline loose coupling or the long baseline tight coupling is used is judged, and corresponding LBL factor nodes are constructed according to different input long baseline measurement information.

The LBL measurement equation can be expressed as:

wherein n is ^LBL Measurement noise, h, which is the measurement equation of LBL ^LBL ( _n ) Is a measurement function, and note that the long baseline measurement information in the measurement functions corresponding to different coupling modes is different. t is t _n The LBL factor node is only equal to t _n Time variable node x _n The correlation, i.e. there is an edge connection between them, the LBL measurement information acts as a constraint. f (f) _LBL The factor is a unitary edge, namely:

the above embodiment enables the construction of the LBL factor node to be switched in a dynamic adaptive manner. When the long baseline positioning information is good in performance, constructing an LBL factor node by adopting long baseline loose coupling; when the number of the long baseline positioning sensors is smaller than the preset number, namely the number is insufficient to calculate the information such as the position, the LBL factor node is constructed by adopting a long baseline tight coupling mode, and the dynamic fusion mode is adopted, so that the robustness of the system is improved, and the calculated amount and the power consumption of the system are reduced.

In some embodiments, referring to fig. 4, the method may further comprise the steps of:

in step S71, it is detected whether the received positioning information of the navigation satellites is from a predetermined number or more of navigation satellites.

Step S72, when the received navigation satellite positioning information is detected to come from a preset number of navigation satellites and above, constructing a navigation satellite factor node by adopting a navigation loose coupling method according to the navigation satellite positioning information.

Step S73, when the received navigation satellite positioning information is detected to come from the navigation satellites below the preset number, constructing a navigation satellite factor node by adopting a navigation tight coupling method according to the navigation satellite positioning information.

Wherein the preset number is preferably 4.

Setting the positioning information of the received navigation satellites at the moment k, n is the number of the received navigation satellites, and constructing f _GNSS Factor node:

specifically, if the number of navigation satellites received by the AUV when the AUV floats on the water surface in a certain water area is greater than or equal to 4, performing navigation loose coupling by taking the difference between the position information calculated by the GNSS and the SINS as measurement information; and if the number of the received navigation satellites is smaller than 4, performing navigation tight coupling by taking GNSS pseudo-range information as measurement information.

The method solves the problems that the received navigation satellite positioning information has instability and uncertainty due to the fact that the number of the navigation satellites is large in general, and the method improves the robustness of the intelligent navigation system and reduces the calculated amount and the power consumption of the optimal navigation result due to the self-adaptive coupling mode of the navigation satellite positioning information.

In some embodiments, step S72 may specifically include: and (3) differencing the navigation satellite positioning information and the position information calculated by the strapdown inertial navigation system to obtain navigation satellite measurement information, and constructing a navigation satellite factor node according to the navigation satellite measurement information.

In some embodiments, step S73 may specifically include: and taking pseudo-range information in the navigation satellite positioning information as navigation satellite measurement information, and constructing a navigation satellite factor node according to the navigation satellite measurement information.

In the specific implementation process, t _p And when global measurement information of the GNSS is received at any moment, constructing GNSS factor nodes, and adding the GNSS factor nodes into a factor graph frame. The GNSS participating fusion mode is divided into navigation loose coupling and navigation tight coupling, wherein the navigation loose coupling means that GNSS measurement information is obtained by making difference between navigation satellite positioning information and position information calculated by a strapdown inertial navigation system To (a) to (b); the navigation tight coupling refers to navigation pseudo-range information, such as pseudo-range, doppler shift and the like, of which the GNSS measurement information is not resolved. Firstly, detecting the number of navigation satellites of received GNSS signals through a proposed self-adaptive coupling method, judging whether to use navigation loose coupling or navigation tight coupling, and constructing corresponding GNSS factor nodes according to different input measurement information.

The system obtains GNSS measurement information at a time that generally provides global positioning of the system with initial values and error correction before and when the aircraft is in water. The GNSS metrology equation can be expressed as:

wherein n is ^GNSS Measurement noise, h, which is the measurement equation of GNSS ^GNSS ( _p ) The measurement function is used for paying attention to the difference of navigation satellite measurement information in the measurement functions of different coupling modes. The GNSS factor node is only related to the variable node x at the moment _p In relation, only one side is connected between the two sides, and GNSS measurement information is used as constraint. f (f) _GNsS Is a unitary edge.

The above embodiment enables the construction of the GNSS factor node to be switched in a dynamic adaptive manner. When the navigation satellite positioning information is well represented, a GNSS factor node is constructed by adopting navigation loose coupling; when the number of the received navigation satellites is smaller than the preset number, namely insufficient to calculate the information such as the position, the GNSS factor node is constructed by adopting a navigation tight coupling mode. By adopting the dynamic fusion mode, the robustness of the system is improved, and the calculated amount and the power consumption of the system are reduced.

Referring to fig. 5, another embodiment of the present application provides an intelligent navigation apparatus, which is applied to an intelligent navigation device, and includes:

the initialization module 101 is configured to construct a sliding window, define a sliding window threshold, construct an initial variable node and a priori factor node according to pose information of an initial moment of the aircraft, and connect the initial variable node and the priori factor node to obtain an initialization factor graph.

A variable node construction module 102, configured to, according to the acquired current time t of the aircraft _k (k∈N ^* ) Inertial measurement value of (c) and current time t _k The inertia estimated state value of (2) gets the next time t _k+1 According to the current time t _k Constructing a first variable node according to the inertia estimation state value of (2) and the next time t _k+1 And constructing a second variable node from the inertia state predicted value of (a).

The variable node construction module 102 is further configured to replace the initial variable node in the initial factor graph with the first variable node when k=1, and connect the first variable node with the a priori factor node.

A factor node construction module 103 for constructing a factor node according to the acquired next time t of the aircraft _k+1 Inertial estimated state value of (2) and next time t _k+1 Constructing a current factor node by using the inertia state predicted value of (1); the first variable node and the second variable node are respectively connected with the current factor node.

An additional node construction module 104 for detecting a current time t _k Whether additional sensor measurement information is received, if at the current time t _k And when receiving the measurement information of the additional sensor, constructing a corresponding additional factor node according to the measurement information of the additional sensor, and independently connecting the additional factor node with the first variable node to obtain a factor graph to be optimized.

The navigation result calculation module 105 is configured to enable the sliding window to slide on the factor graph to be optimized according to a preset rule until the number of variable nodes in the factor graph to be optimized in the sliding window reaches a sliding window threshold value, and calculate after optimizing the variable nodes in the sliding window and the factor nodes connected with the variable nodes, so as to obtain an optimal navigation result of the aircraft.

According to the intelligent navigation device provided by the embodiment, sensor measurement information received by an aircraft at a plurality of moments is converted into factor nodes and variable nodes which are placed in a constructed factor graph through the initialization module 101, the variable node construction module 102, the factor node construction module 103 and the additional node construction module 104, so that different frequency fusion among different sensors is realized, and the problem of different frequency isomerism in a navigation system in the prior art is solved; meanwhile, the navigation result calculation module 105 adopts a sliding window method when the factor graph is optimized to calculate and solve the navigation result, so that the calculation amount of calculating the navigation result is reduced to a certain extent, the real-time performance of calculated data is ensured by calculating by using the sensor information received at the latest moment, and the calculation precision of the navigation result is improved.

The specific limitation of the intelligent navigation device provided in this embodiment may be referred to the above embodiments of an intelligent navigation method, which is not described herein. Each module in the intelligent navigation apparatus may be implemented in whole or in part by software, hardware, or a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

Embodiments of the present application provide a computer device that may include a processor, memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program, when executed by a processor, causes the processor to perform the steps of a smart navigation method as in any of the embodiments described above.

The working process, working details and technical effects of the computer device provided in this embodiment may be referred to the above embodiments of an intelligent navigation method, which are not described herein.

The present embodiments provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of a smart navigation method as in any of the embodiments described above. The computer readable storage medium refers to a carrier for storing data, and may include, but is not limited to, a floppy disk, an optical disk, a hard disk, a flash Memory, and/or a Memory Stick (Memory Stick), etc., where the computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable devices.

The working process, working details and technical effects of the computer readable storage medium provided in this embodiment can be referred to the above embodiments of an intelligent navigation method, and are not described herein.

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the various embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The above examples merely represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the invention. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application is to be determined by the claims appended hereto.

Claims

1. An intelligent navigation method, characterized in that the method comprises:

according to the acquired current moment t of the aircraft _k (k∈N ^* ) Inertial measurement value of (c) and current time t _k The inertia estimated state value of (2) gets the next time t _k+1 According to the current time t _k Constructing a first variable node according to the inertia estimation state value of (a) and (b) at the next moment t _k+1 Constructing a second variable node by the inertia state predicted value of the first variable node; if k=1, replacing the initial variable node in the initial factor graph by the first variable node, and connecting the first variable node with the prior factor node;

according to the acquired next time t of the aircraft _k+1 Inertial estimated state value of (c) and said next time t _k+1 Constructing a current factor node by using the inertia state predicted value of (1); the first variable node and the second variable node are respectively connected with the current factor node;

detecting the current time t _k Whether additional sensor measurement information is received, if at the current time t _k Receiving the additional sensorThe measurement information is used for constructing a corresponding additional factor node according to the measurement information of the additional sensor, and the additional factor node is independently connected with the first variable node to obtain a factor graph to be optimized; wherein the additional sensor measurement information includes long baseline positioning information; the additional factor nodes include long baseline factor nodes;

Specifically, detecting whether the received long baseline positioning information comes from a preset number of long baseline positioning sensors or more;

when the received long baseline positioning information is detected to come from a preset number of long baseline positioning sensors or more, constructing the long baseline factor node by adopting a long baseline loose coupling method according to the long baseline positioning information;

when the received long baseline positioning information is detected to come from the long baseline positioning sensors with the quantity below a preset value, constructing the long baseline factor node by adopting a long baseline tight coupling method according to the long baseline positioning information;

and the sliding window slides on the factor graph to be optimized according to a preset rule until the number of variable nodes in the factor graph to be optimized in the sliding window reaches the sliding window threshold, and the variable nodes in the sliding window and factor nodes connected with the variable nodes are calculated after being optimized to obtain an optimal navigation result of the aircraft.

2. The method according to claim 1, wherein said determining is based on the acquired current time t of the aircraft _k Inertial measurement value of (c) and current time t _k The inertia estimated state value of (2) gets the next time t _k+1 An inertial state prediction value of (1), comprising:

according to the acquired current moment t of the aircraft _k Inertial measurement value of (c) and said current time t _k The inertia estimated state value of (2) is calculated based on an inertia state transition equation to obtain the next time t _k+1 Is a predicted value of the inertial state of the vehicle.

3. According to the weightsThe method according to claim 1, characterized in that the following time t of the aircraft is acquired _k+1 Inertial estimated state value of (c) and said next time t _k+1 The inertia state predicted value of (1) constructs a current factor node, comprising:

according to the acquired next time t of the aircraft _k+1 Inertial estimated state value of (c) and said next time t _k+1 And constructing the current factor node based on a residual function.

4. The method of claim 1, wherein the additional sensor measurement information comprises: long baseline positioning information, doppler velocity measurement information, magnetic compass information, pressure sensing information and navigation satellite positioning information; the additional factor node includes: long baseline factor node, doppler velocity factor node, magnetic compass factor node, pressure sensing factor node, and navigation satellite factor node.

5. The method according to claim 1, wherein the method further comprises:

when the detection time of the sliding window to the number of variable nodes in the sliding window reaches a preset detection time threshold, the variable nodes in the sliding window and factor nodes connected with the variable nodes are optimized and then calculated, and the optimal navigation result of the aircraft is obtained.

6. The method of claim 5, wherein the method further comprises:

and when the number of variable nodes in the factor graph to be optimized in the sliding window reaches the threshold of the sliding window or when the detection time of the sliding window to the number of variable nodes in the sliding window reaches a preset detection time threshold, carrying out marginalization on the part, which is positioned outside the sliding window, of the factor graph to be optimized by adopting a ShuerBu method.

7. The method according to claim 1, wherein the optimizing the variable nodes and the factor nodes connected to the variable nodes in the sliding window to calculate an optimal navigation result of the aircraft includes:

and optimizing the variable nodes in the sliding window and factor nodes connected with the variable nodes by adopting a Gaussian-Newton iteration method or a Levenberg-Marquardt method, and then calculating to obtain the optimal navigation result of the aircraft.

8. The method of claim 7, wherein constructing the long baseline factor node using a long baseline loosely coupled method comprises: and differencing the long baseline positioning information and the position information solved by the strapdown inertial navigation system to obtain long baseline measurement information, and constructing the long baseline factor node according to the long baseline measurement information.

9. The method of claim 7, wherein constructing the long baseline factor node using a long baseline tight coupling method comprises: and differencing the long baseline positioning information and the skew difference in the strapdown inertial navigation system to obtain long baseline measurement information, and constructing the long baseline factor node according to the long baseline measurement information.

10. The method according to claim 4, wherein the method further comprises:

when the received navigation satellite positioning information is detected to come from a preset number of navigation satellites or more, constructing the navigation satellite factor node by adopting a navigation loose coupling method according to the navigation satellite positioning information;

When the received navigation satellite positioning information is detected to come from the navigation satellites with the number below a preset value, constructing the navigation satellite factor node by adopting a navigation tight coupling method according to the navigation satellite positioning information.

11. The method of claim 10, wherein constructing the navigation satellite factor node using a navigation loose coupling method comprises:

and differencing the navigation satellite positioning information and the position information calculated by the strapdown inertial navigation system to obtain navigation satellite measurement information, and constructing the navigation satellite factor node according to the navigation satellite measurement information.

12. The method of claim 10, wherein said constructing said navigational satellite factor node using a navigational tight coupling method comprises:

and taking the pseudo-range information in the navigation satellite positioning information as navigation satellite measurement information, and constructing the navigation satellite factor node according to the navigation satellite measurement information.

13. An intelligent navigation apparatus, characterized by being applied to an intelligent navigation device, the apparatus comprising:

A variable node construction module for constructing a variable node according to the acquired current moment t of the aircraft _k (k∈N ^* ) Inertial measurement value of (c) and current time t _k The inertia estimated state value of (2) gets the next time t _k+1 According to the current time t _k Constructing a first variable node according to the inertia estimation state value of (a) and (b) at the next moment t _k+1 Constructing a second variable node by the inertia state predicted value of the first variable node; the variable node construction module is further configured to replace the initial variable node in the initialization factor graph with the first variable node when k=1, and connect the first variable node with the prior factor node;

factor node structureA building module for building up a module according to the acquired next time t of the aircraft _k+1 Inertial estimated state value of (c) and said next time t _k+1 Constructing a current factor node by using the inertia state predicted value of (1); the first variable node and the second variable node are respectively connected with the current factor node;

an additional node construction module for detecting the current time t _k Whether additional sensor measurement information is received, if at the current time t _k If the additional sensor measurement information is received, constructing a corresponding additional factor node according to the additional sensor measurement information, and independently connecting the additional factor node with the first variable node to obtain a factor graph to be optimized; wherein the additional sensor measurement information includes long baseline positioning information; the additional factor nodes include long baseline factor nodes;

Specifically, the additional node construction module is used for detecting whether the received long baseline positioning information is from a preset number of long baseline positioning sensors or more; when the received long baseline positioning information is detected to come from a preset number of long baseline positioning sensors or more, constructing the long baseline factor node by adopting a long baseline loose coupling method according to the long baseline positioning information; when the received long baseline positioning information is detected to come from the long baseline positioning sensors with the quantity below a preset value, constructing the long baseline factor node by adopting a long baseline tight coupling method according to the long baseline positioning information;

the navigation result calculation module is used for enabling the sliding window to slide on the factor graph to be optimized according to a preset rule until the number of variable nodes in the factor graph to be optimized in the sliding window reaches the sliding window threshold value, and calculating the variable nodes in the sliding window and factor nodes connected with the variable nodes after optimizing to obtain an optimal navigation result of the aircraft.

14. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any one of claims 1 to 12 when the computer program is executed.

15. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 12.