Multi-source navigation federated filtering method and system based on exponential attenuation adaptive information distribution

Abstract

The application discloses a kind of multi-source navigation federal filter method and system based on exponential attenuation adaptive information distribution. Specifically: the data of SINS, GNSS and ODOM are obtained;SINS data is taken as the reference system in federal filter, and SINS data is solved in real time, the results of SINS solution are input into the main filter, the first and second sub-filters respectively, Kalman filtering is carried out on GNSS data and ODOM data in the first and second sub-filters respectively, and the filtering results are output to the main filter for fusion;Then, an information distribution strategy based on the exponential attenuation formula of the Frobenius norm of P matrix is constructed, and an information distribution factor is obtained;The main filter multiplies the common state quantity after fusion and the state estimation mean square error matrix of common state by the information distribution factor, and feeds back to each sub-filter for feedback reset.The application has the advantages of high precision, strong stability, strong anti-interference ability and strong adaptability.

Description

Technical Field

[0001] This invention relates to the field of multi-source navigation technology, and in particular to a multi-source navigation federated filtering method and system based on exponential decay adaptive information allocation. Background Technology

[0002] In multi-sensor, multi-source navigation systems, a common filtering method is centralized filtering, which involves designing a high-dimensional synthetic filter that includes all state variables for Kalman filtering. However, this approach has many drawbacks, such as high state dimensionality, high computational cost, and the inability to guarantee the real-time performance of the navigation system. To reduce computational cost, federated filtering can be used for distributed cost reduction.

[0003] In federated filtering with a feedback-reset structure, the information allocation factor is usually set as a constant, with the classic allocation method being average allocation. However, during the federated filtering process, the filtering quality of each sub-filter can change. For example, when the filtering quality of the SINS / GNSS sub-filter deteriorates due to environmental factors in a multi-source navigation system, it is necessary to dynamically adjust the information allocation coefficients to reduce the fusion weight of the SINS / GNSS sub-filters. Therefore, a fixed information allocation method cannot fully leverage the advantages of multi-sensor fusion, necessitating dynamic information allocation.

[0004] From the perspective of system filtering accuracy, the higher the accuracy of the subsystem, the larger the information matrix and the larger the information allocation factor; the lower the accuracy of the subsystem, the smaller the information allocation factor. According to the above rules, the existing information allocation methods include the following two types: (1) The information allocation factor is directly calculated from the weights of the F norm of the state estimation mean square error matrix; (2) First, the information coefficients are constructed using the F norm of the state estimation mean square error matrix, and then the information allocation coefficients are calculated.

[0005] Both of the above information allocation methods are adaptive information allocation methods based on inverse proportional functions. Although they conform to the principle of information allocation, they have hidden dangers. They do not consider the situation where the information allocation factor is close to 0 in extreme cases. This may lead to the common state estimation mean square error matrix of a certain sub-filter being close to 0 after information allocation. This may cause the matrix to be close to singular, which may lead to numerical instability during subsequent information fusion. Numerical calculations will encounter accuracy problems, thus affecting the navigation and positioning accuracy. Summary of the Invention

[0006] The purpose of this invention is to provide a multi-source navigation federated filtering method and system based on exponential decay adaptive information allocation, which enables the information allocation factor to be adaptively adjusted according to the allocation principle, thereby achieving high filtering accuracy, high numerical calculation stability, strong adaptability, and strong fault tolerance.

[0007] The technical solution for implementing this invention is: a multi-source navigation federated filtering method based on exponential decay adaptive information allocation, comprising the following steps:

[0008] Step 1: Acquire data from the Strapdown Inertial Navigation System (SINS), Global Navigation Satellite System (GNSS), and Odometer ODOM;

[0009] Step 2: Use SINS as the reference system in the federated filter, solve the SINS data in real time, and input the SINS solution into the main filter, the first sub-filter and the second sub-filter respectively.

[0010] Step 3: The first sub-filter uses GNSS data as observations for Kalman filtering; the second sub-filter first uses ODOM data and SINS data to perform dead reckoning, and then uses the reckoning results as observations for Kalman filtering.

[0011] Step 4: The filtering results of the first and second sub-filters are output to the main filter. The common state variables output by the two sub-filters, namely velocity, position and attitude, are fused in the main filter.

[0012] Step 5: Construct an information allocation strategy based on the exponential decay formula of the P matrix Frobenius norm, use an adaptive information allocation method based on exponential decay to complete the information allocation, obtain the expression of the information allocation factor, and calculate the information allocation factor.

[0013] Step 6: The main filter multiplies the fused common state variables and the mean square error matrix of the common state estimates by the information allocation factor, and feeds them back to each sub-filter to complete the feedback reset task.

[0014] A multi-source navigation federated filtering system based on exponentially decaying adaptive information allocation is disclosed. This system implements the aforementioned multi-source navigation federated filtering method based on exponentially decaying adaptive information allocation. The system comprises a first module to a fourth module, the functions of which are as follows:

[0015] The first module acquires data from the strapdown inertial navigation system (SINS), the global navigation satellite system (GNSS), and the wheeled odometer (ODOM).

[0016] The second module uses SINS as the reference system in the federated filter, performs real-time calculations on the SINS data, and inputs the results of the SINS calculations into the main filter, the first sub-filter, and the second sub-filter, respectively.

[0017] The first sub-filter uses GNSS data as observations for Kalman filtering; the second sub-filter first uses ODOM and SINS data to perform dead reckoning, and then uses the reckoning results as observations for Kalman filtering; the filtering results of the first and second sub-filters are output to the main filter, where the common state variables output by the two sub-filters, namely velocity, position, and attitude, are fused.

[0018] The third module constructs an information allocation strategy based on the exponential decay formula of the P matrix Frobenius norm, uses an adaptive information allocation method based on exponential decay to complete the information allocation, obtains the expression of the information allocation factor, and calculates the information allocation factor.

[0019] In the fourth module, the main filter feeds back the fused common state variables and the mean square error matrix of the common state estimates by the information allocation factor to each sub-filter, thus completing the task of feedback reset.

[0020] A mobile terminal includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the multi-source navigation federated filtering method based on exponential decay adaptive information allocation.

[0021] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the multi-source navigation federated filtering method based on exponentially decaying adaptive information allocation.

[0022] Compared with the prior art, the significant advantages of this invention are: (1) the information allocation factor can be adaptively adjusted according to the filtering quality of each sub-filter, giving full play to the advantages of multiple sensors and improving the filtering accuracy; (2) the norm is attenuated by using an exponential function, which improves the stability of the filtering process after information allocation. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating a multi-source navigation federated filtering method based on exponential decay adaptive information allocation.

[0024] Figure 2 This is a comparison diagram of the planar trajectory and the baseline trajectory of the original data output by federated filtering using three different information allocation methods in an embodiment of the present invention.

[0025] Figure 3 This is a comparison chart of the three-axis position errors of the original data using federated filtering with three different information allocation methods in an embodiment of the present invention.

[0026] Figure 4This is a comparison chart of the overall spatial location error of the original data using three different information allocation methods in this embodiment of the invention.

[0027] Figure 5 A comparison of the three-axis position errors of federated filtering for three information allocation methods after adding noise to GNSS data in this embodiment of the invention.

[0028] Figure 6 A comparison of the overall spatial location error of federated filtering for three information allocation methods after adding noise to GNSS data in this embodiment of the invention.

[0029] Figure 7 A comparison of the three-axis velocity errors of the federated filtering for three information allocation methods after adding noise to GNSS data in this embodiment of the invention.

[0030] Figure 8 A comparison of heading angle errors of federated filtering for three information allocation methods after adding noise to GNSS data in this embodiment of the invention.

[0031] Figure 9 This refers to the range of values ​​for the information allocation factor β in the information allocation method based on the inverse proportional function in this embodiment of the invention.

[0032] Figure 10 This refers to the range of values ​​for the information allocation factor β in the information allocation method based on exponential decay in this embodiment of the invention. Detailed Implementation

[0033] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0034] like Figure 1 As shown, the present invention provides a multi-source navigation federated filtering method based on exponential decay adaptive information allocation, comprising the following steps:

[0035] Step 1: Acquire data from the Strapdown Inertial Navigation System (SINS), Global Navigation Satellite System (GNSS), and Odometer ODOM;

[0036] Step 2: Use SINS as the reference system in the federated filter, solve the SINS data in real time, and input the SINS solution into the main filter, the first sub-filter and the second sub-filter respectively.

[0037] Step 3: The first sub-filter uses GNSS data as observations for Kalman filtering; the second sub-filter first uses ODOM data and SINS data to perform dead reckoning, and then uses the reckoning results as observations for Kalman filtering.

[0038] Step 4: The filtering results of the first and second sub-filters are output to the main filter. The common state variables output by the two sub-filters, namely velocity, position and attitude, are fused in the main filter.

[0039] Step 5: Construct an information allocation strategy based on the exponential decay formula of the P matrix Frobenius norm, use an adaptive information allocation method based on exponential decay to complete the information allocation, obtain the expression of the information allocation factor, and calculate the information allocation factor.

[0040] Step 6: The main filter multiplies the fused common state variables and the mean square error matrix of the common state estimates by the information allocation factor, and feeds them back to each sub-filter to complete the feedback reset task.

[0041] As a specific example, in step 3, sub-filter 1 uses GNSS data as observations for Kalman filtering; sub-filter 2 first uses ODOM and SINS data to perform dead reckoning, and then uses the reckoning results as observations for Kalman filtering, as detailed below:

[0042] Step 3.1, the state-space model and observation model of the sub-filter are as follows:

[0043]

[0044] in, Let be the state estimate of sub-filter i at time k; Let be the one-step state transition matrix of sub-filter i from time k-1 to time k; Let be the noise matrix of sub-filter i; These are the observation value, observation matrix, and observation noise matrix of sub-filter i at time k, respectively;

[0045] Step 3.2, State variables of sub-filter 1 Selected as:

[0046]

[0047] Where (φ) T 、(δv n ) T (δp) T 、(ε b ) T , These represent attitude error, velocity error, position error, gyroscope bias, and accelerometer bias, respectively.

[0048] Observation vector of sub-filter 1 and observation matrix They are respectively:

[0049]

[0050] in, and These represent the velocity and position information obtained from the strapdown inertial navigation system, respectively. and These represent the velocity and position information obtained from GNSS observations, respectively; 0 3×3 I is a 3×3 zero matrix. 3×3 It is a 3×3 identity matrix;

[0051] Step 3.3, State variables of sub-filter 2 Selected as:

[0052]

[0053] Where (φ) T 、(δv n ) T (δp) T 、(ε b ) T , These represent attitude error, velocity error, position error, gyroscope bias, and accelerometer bias, respectively. (δp) D ) T This indicates the displacement error obtained from dead reckoning using wheeled odometers.

[0054] Observation vector of sub-filter 2 and observation matrix They are respectively:

[0055]

[0056] in, This represents the position information calculated by the strapdown inertial navigation system. This indicates the position information obtained from dead reckoning using wheeled odometers; 0 3×3 I is a 3×3 zero matrix. 3×3 It is a 3×3 identity matrix;

[0057] Step 3.4, the sub-filters, according to Kalman filtering, have:

[0058]

[0059] In the formula, For the prior estimate at time k, This is the posterior estimate at time k-1;

[0060] Step 3.5: Calculate the mean square error matrix for one-step state prediction as follows:

[0061]

[0062] In the formula, P k-1 Let Q be the mean square error matrix for the state estimation at time k-1. k-1 The system process noise matrix;

[0063] Step 3.6: Calculate the filter gain as follows:

[0064]

[0065] In the formula, K k R is the filter gain at time k. k Let be the observation noise matrix at time k;

[0066] Step 3.7, Update the state estimate as follows:

[0067]

[0068] In the formula This is the posterior estimate at time k;

[0069] Step 3.8: Update the mean squared error matrix of the state estimate as follows:

[0070] P k =(IK k H k )P k / k-1 (12)

[0071] In the formula P k Let be the mean square error matrix of the state estimation at time k.

[0072] As a specific example, in step 4, the filtering results of sub-filter 1 and sub-filter 2 are output to the main filter. In the main filter, the common state variables output by the two sub-filters, namely velocity, position, and attitude, are fused, as follows:

[0073] First, take the common state of the sub-filters and let... Represents the common state of sub-filter i, taking That is, attitude error, velocity error, position error, gyroscope bias, and accelerometer bias are taken as common states; let Let represent the unique state of sub-filter i; in the main filter, only common states shared by all sub-filters can be fused and reset, and non-common states cannot be fused; let the state estimate and mean square error matrix of the i-th sub-filter at time k be:

[0074]

[0075] Each sub-filter undergoes information fusion:

[0076]

[0077] After information fusion, the globally optimal state estimate is obtained. and the mean square error matrix of state estimation

[0078] As a specific example, step 5 constructs an information allocation strategy based on the exponential decay formula of the P matrix Frobenius norm, uses an adaptive information allocation method based on exponential decay to complete the information allocation, obtains the expression for the information allocation factor, and calculates the information allocation factor β, as follows:

[0079] Step 5.1: This invention employs a federated reset structure in a federated filter, multiplying the mean square error matrix of the common state estimation by the information allocation factor β and feeding it back to each sub-filter:

[0080]

[0081] Where, β i Assign factors to information, satisfying the following relationship:

[0082]

[0083] Where, β m The information allocation factor of the main filter, β i The information allocation factor for each sub-filter varies depending on the structure of the federated filter and the selection strategy for information allocation. As a result, the federated filter has different values.

[0084] Step 5.2: To fully utilize the information from each sub-filter in the main filter, improve filtering accuracy, and maintain a certain degree of fault tolerance, an adaptive information allocation method is adopted; in the sub-filters, This describes the filtering quality of sub-filter i on the common state at time k; As an information matrix, when The smaller the value, the better the quality of the system's state estimation, and the larger the information matrix will be.

[0085] From equation (16), it can be seen that in information allocation, the mean square error matrix of the estimated common state after information allocation of the sub-filters is inversely proportional to the information allocation factor β. The larger β is, the smaller the error covariance matrix after information allocation. From equation (15), it can be seen that the fusion of state variables is essentially based on the inverse matrix of the error covariance matrix of each sub-filter as the weight, i.e. The smaller the value, the greater the fusion weight; based on the above analysis, the following conclusions can be drawn:

[0086] The information allocation factor β affects the system's weight in utilizing navigation information from sub-filters; a smaller β indicates a lower utilization weight. To ensure the fusion result is more significantly influenced by the higher-precision sub-filters, the information allocation factor of the higher-precision sub-filters should be larger. Based on these conclusions, the following information allocation strategy is derived:

[0087]

[0088] Step 6: The main filter multiplies the fused common state variables and the mean square error matrix of the common state estimates by the information allocation factor β, and feeds them back to each sub-filter to complete the feedback reset task.

[0089] This invention also provides a multi-source navigation federated filtering system based on exponentially decaying adaptive information allocation. This system is used to implement the aforementioned multi-source navigation federated filtering method based on exponentially decaying adaptive information allocation. The system includes a first module to a fourth module, wherein the functions of each module are as follows:

[0090] The first module acquires data from the strapdown inertial navigation system (SINS), the global navigation satellite system (GNSS), and the wheeled odometer (ODOM).

[0091] The second module uses SINS as the reference system in the federated filter, performs real-time calculations on the SINS data, and inputs the results of the SINS calculations into the main filter, the first sub-filter, and the second sub-filter, respectively.

[0092] The first sub-filter uses GNSS data as observations for Kalman filtering; the second sub-filter first uses ODOM and SINS data to perform dead reckoning, and then uses the reckoning results as observations for Kalman filtering; the filtering results of the first and second sub-filters are output to the main filter, where the common state variables output by the two sub-filters, namely velocity, position, and attitude, are fused.

[0093] The third module constructs an information allocation strategy based on the exponential decay formula of the P matrix Frobenius norm, uses an adaptive information allocation method based on exponential decay to complete the information allocation, obtains the expression of the information allocation factor, and calculates the information allocation factor.

[0094] In the fourth module, the main filter feeds back the fused common state variables and the mean square error matrix of the common state estimates by the information allocation factor to each sub-filter, thus completing the task of feedback reset.

[0095] The present invention also provides a mobile terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the multi-source navigation federated filtering method based on exponential decay adaptive information allocation.

[0096] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the multi-source navigation federated filtering method based on exponential decay adaptive information allocation.

[0097] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0098] Example

[0099] This embodiment uses an unmanned vehicle equipped with a ROS system to collect data, and performs a hardware-in-the-loop simulation experiment on the collected data in MATLAB to verify the effectiveness and accuracy of the multi-source navigation federated filtering method based on exponential decay adaptive information allocation proposed in this invention.

[0100] To compare the multi-source navigation federated Kalman filtering method based on exponential decay adaptive information allocation with the federated Kalman filtering method based on traditional information allocation, we first used an unmanned vehicle to complete a sports car experiment to collect data. Then, we compared the federated Kalman filtering results of several information allocation methods using the collected actual sports car data to verify the advantages of the present invention.

[0101] The specific working process of this embodiment is as follows: First, an unmanned vehicle experimental platform is built, equipped with a strapdown inertial navigation system (SINS), a global navigation satellite system (GNSS), an odometry system (ODOM), an integrated navigation system (IPMV), and an onboard navigation computer. Then, the vehicle runs laps in an open square and records data from various sensors. The high-precision integrated navigation system serves as the benchmark, and its output speed, position, and attitude are taken as the true values. The raw sensor data is recorded using a rosbag package. Finally, the data is extracted from the rosbag package and saved as a txt file. MATLAB is used to read the txt file, import the sensor data, and process it.

[0102] In the hardware-in-the-loop simulation experiment, the collected data were filtered and estimated using both the traditional federated Kalman filter method for information allocation and the multi-source navigation federated filtering method based on exponential decay adaptive information allocation of this invention. The difference between the filtered result and the IPMV data was then used to obtain the estimation error. The root mean square error (RMES) of the experimental results was calculated to measure the magnitude of the error, and the method of this invention was compared with the multi-source navigation federated filtering method based on the traditional information allocation method. To demonstrate the effectiveness of the method of this invention, zero-mean noise was added to the collected GNSS data in MATLAB, and this noisy GNSS data was then used for further filtering and comparison.

[0103] Table 1 Comparison of root mean square error of simulation results based on original data.

[0104] method Spatial position error (m) Spatial velocity error (m / s) Heading angle error (°) average 0.8570 0.0940 0.7581 Dync1 0.9001 0.0956 0.8137 Dync2 0.8563 0.0945 0.7594

[0105] Table 1 shows a comparison of simulation results using raw data collected by an unmanned vehicle. In the table, `average` represents the traditional onboard federated filtering method with average information allocation; `Dync1` represents the onboard federated filtering method based on an inverse proportional function-based adaptive information allocation; and `Dync2` represents the onboard federated filtering method of this invention based on an exponentially decaying adaptive information allocation. Spatial position error in the table is the total position error after the vector summation of the three-axis position errors, and spatial velocity error is the total velocity error after the vector summation of the three-axis velocity errors. The table shows that, without human interference, the output of this invention's method is almost identical to that of the average information allocation federated filtering method, while the result of the onboard federated filtering method based on the inverse proportional function-based information allocation has a larger error.

[0106] Figure 2 This is a comparison chart of the planar trajectories obtained by the three methods: average, Dync1, and Dync2, and the baseline trajectory. Figures 3-4 The three-axis position error and spatial position error are output by the three methods in this simulation experiment.

[0107] Table 2 Comparison of root mean square error of simulation results for GNSS data location after adding noise.

[0108] method Spatial position error (m) Spatial velocity error (m / s) Heading angle error (°) average 4.3857 0.5040 3.5548 Dync1 4.1507 0.5405 2.2931 Dync2 2.1795 0.2642 1.8221

[0109] Table 2 compares the root mean square errors (RMSEs) of the federated filtering results for the three information allocation methods after adding noise to the GNSS position and velocity. It can be seen that the method of this invention yields the smallest error. Compared to the vehicle-mounted federated filtering method with average information allocation, the RMSE of spatial position is reduced by approximately 50%, the overall RMSE of spatial velocity is reduced by 47%, and the RMSE of heading angle is reduced by approximately 44%. Compared to the vehicle-mounted federated filtering method based on inverse proportional function information allocation, the RMSE of spatial position is reduced by approximately 47%, the overall RMSE of spatial velocity is reduced by 51%, and the RMSE of heading angle is reduced by approximately 20.53%.

[0110] Figures 5-8 The figures show the triaxial position error, overall spatial position error, triaxial velocity error, and heading angle error of the federated filtering outputs for three information allocation methods after adding GNSS noise. As can be seen from the figures, the method of this invention has significant advantages over the other two methods, with a more pronounced reduction in error.

[0111] It is evident that the multi-source navigation federated filtering method based on exponential decay adaptive information allocation proposed in this invention, compared with the federated filtering method based on traditional information allocation, can not only adaptively adjust the information allocation factor, fully leverage the advantages of multiple sensors, and improve filtering accuracy, but also exhibit better adaptive fault tolerance performance when the observation sensors have significant noise interference, reducing errors caused by additional noise. Furthermore, the use of an exponential function to decay the norm in the information allocation formula limits the information allocation factor to a reasonable range, avoiding extreme cases and making the filtering process after information allocation more stable.

[0112] Figures 9-10 The figure shows the range of values ​​for β based on exponential decay information allocation and inverse proportional function information allocation. As can be seen from the figure, in the exponential decay information allocation method, the value of the information allocation factor β is smoother and is limited to a reasonable range, avoiding extreme values ​​that could lead to instability in numerical calculations.

[0113] In summary, the multi-source navigation federated filtering method based on exponential decay adaptive information allocation proposed in this invention improves the calculation formula of the information allocation factor, enhances the adaptive performance under large noise interference, improves the filtering accuracy and numerical calculation stability, and improves the fault tolerance performance.

[0114] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A multi-source navigation federated filtering method based on exponential decay adaptive information distribution, characterized in that, Comprising the following steps: Step 1, obtaining the data of a strapdown inertial navigation system SINS, a global navigation satellite system GNSS and a wheeled odometry ODOM; Step 2, taking the SINS as a reference system in a federated filter, and performing real-time calculation on the SINS data, and inputting the calculation results of the SINS into a main filter, a first sub-filter and a second sub-filter respectively; Step 3, the first sub-filter performs Kalman filtering with GNSS data as observation; the second sub-filter first performs dead reckoning with the data of the ODOM and the SINS, and then performs Kalman filtering with the calculation results as observation; Step 4, the filtering results of the first sub-filter and the second sub-filter are output to the main filter, and the common state quantities, i.e. velocity, position and attitude, output by the two sub-filters are fused in the main filter; Step 5, constructing an information distribution strategy based on an exponential decay formula of the Frobenius norm of a P matrix, and completing information distribution by using an adaptive information distribution method based on an exponential decay, to obtain an expression of an information distribution factor, and calculating the information distribution factor, which is specifically as follows: Step 5.1, using a federated filter with a feedback reset structure, multiplying the state estimation mean square error matrix of the common state by the information distribution factor and feeding back to each sub-filter: （16） wherein, state estimation that is globally optimal, is the state estimation mean square error matrix; The information is assigned a factor that satisfies the following relationship: （17） wherein, is an information allocation factor of the main filter, is an information allocation factor of each sub-filter, the information allocation factor , has different value modes, so that the federal filter has different structures; Step 5.2, the information distribution factor affects the utilization weight of the navigation information of the sub-filter by the system, and the smaller the information distribution factor is, the lower the utilization weight is, in order to make the fusion result be more affected by the sub-filter with high precision, the information distribution factor of the sub-filter with higher precision should be larger, therefore, the information distribution strategy based on the exponential decay formula of the Frobenius norm of the P matrix is constructed as: （18） Step 6, the main filter multiplies the common state quantity after fusion and the state estimation mean square error matrix of the common state by the information distribution factor, and feeds back to each sub-filter, to complete the task of feedback reset.

2. The multi-source navigation federated filtering method based on exponential decay adaptive information distribution according to claim 1, characterized in that, In step 3, the first sub-filter performs Kalman filtering with GNSS data as observation; the second sub-filter first performs dead reckoning with the data of the ODOM and the SINS, and then performs Kalman filtering with the calculation results as observation, which is specifically as follows: Step 3.1, the state space model and the observation model of the sub-filter are as follows: （1） wherein, is a sub-filter at a time instant; is a sub-filter from a time instant to a time instant; is a sub-filter of the noise matrix of the sub-filter , , are respectively a sub-filter at a time instant; Step 3.2 State variable of the first sub-filter is chosen as: （2） wherein , , , , respectively represent attitude error, velocity error, position error, gyroscope bias and accelerometer bias; the observation vector of the first sub-filter and the observation matrix are respectively: （3） （4） wherein, and respectively represent velocity and position information solved by a strapdown inertial navigation system, and respectively represent velocity and position information observed by a GNSS; is a zero matrix of 3x3, is a unit matrix of 3x3; Step 3.3 State variable of the second sub-filter is chosen as: （5） wherein , , , , denote the attitude error, velocity error, position error, gyroscope bias and accelerometer bias, respectively, denotes the displacement error of the wheeled odometry for dead reckoning. the observation vector of the second sub-filter and the observation matrix are respectively: （6） （7） wherein, represents position information calculated by a strapdown inertial navigation system, represents position information calculated by a wheeled odometry; is a 3x3 zero matrix, is a 3x3 identity matrix; Step 3.4, according to the Kalman filtering, the sub-filter has: （8） wherein is a prior estimate of the time instant, is a posterior estimate of the time instant; Step 3.5, the state one-step prediction mean square error matrix is calculated as: （9） wherein is the state estimation mean square error matrix at time is the system process noise matrix; Step 3.6, the filtering gain is calculated as: （10） wherein is filtering gain at time is observation noise matrix at time Step 3.7, the state estimation is updated as: （11） In the formula is the posterior estimate of the time instant; Step 3.8, the state estimation mean square error matrix is updated as: （12） In the formula is state estimation mean square error matrix at the time 3. The multi-source navigation federated filtering method based on exponential decay adaptive information allocation according to claim 1, characterized in that, In step 4, the filtering results of the first sub-filter and the second sub-filter are output to the main filter, and the common state quantities, i.e. velocity, position and attitude, output by the two sub-filters are fused in the main filter, which is specifically as follows: First, take the common state of the sub-filter, let represent the common state of the sub-filter , take , that is, the attitude error, the speed error, the position error, the gyroscope zero offset, the accelerometer zero offset as the common state; let represent the specific state of the sub-filter i; in the main filter, only the common state shared by all sub-filters can be fused and reset, and the non-common state cannot be fused; set the state estimation and its mean square error matrix of the i-th sub-filter at time k as: ， （13） After information fusion of each sub-filter: （14） （15） After information fusion, the globally optimal state estimation is obtained and the state estimation mean square error matrix .

4. A multi-source navigation federated filtering system based on exponential decay adaptive information distribution, characterized in that, The system is used to realize the multi-source navigation federated filtering method based on adaptive information distribution based on exponential decay according to any one of claims 1-3, and the system comprises a first module to a fourth module, and the functions of each module are specifically as follows: The first module acquires data of a strapdown inertial navigation system (SINS), a global navigation satellite system (GNSS) and a wheeled odometer (ODOM); The second module takes the SINS as a reference system in a federal filter, and solves SINS data in real time, and inputs a result of the SINS solution into a main filter, a first sub-filter and a second sub-filter respectively; The first sub-filter performs Kalman filtering with GNSS data as observation, and the second sub-filter performs Kalman filtering with a result of dead reckoning with ODOM data and SINS data as observation; the filtering results of the first sub-filter and the second sub-filter are output to the main filter, and common state quantities, i.e. velocity, position and attitude, output by the two sub-filters are fused in the main filter; The third module constructs an information distribution strategy based on an exponential decay formula of a P-matrix Frobenius norm, adopts an adaptive information distribution method based on exponential decay to complete information distribution, obtains an expression of an information distribution factor, and calculates the information distribution factor; The fourth module feeds back the common state quantities after fusion and the state estimation mean square error matrix of the common state to each sub-filter by multiplying the information distribution factor, and completes the task of feedback reset.

5. A mobile terminal comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the multi-source navigation federal filtering method based on adaptive information distribution based on exponential decay according to any one of claims 1-3.

6. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the steps in the multi-source navigation federal filtering method based on adaptive information distribution based on exponential decay according to any one of claims 1-3.

Images