Closed space unmanned system positioning method under external interference

Abstract

The application discloses a closed space unmanned system positioning method under external interference and relates to the technical field of unmanned system positioning. The application accurately identifies abnormal interference by using the detection mechanism of the extended sign signal design, and accurately estimates the positioning of the unmanned system by using the forward filtering and backward smoothing based on the generalized statistics. The application can realize high-precision positioning of the unmanned system in a complex scene where the error characteristics of wireless measurement signals present non-Gaussian coupling unknown interference, reduces the influence of unknown external disturbance and non-Gaussian noise on the positioning effect, improves the positioning accuracy of the unmanned system in indoor, downhole, basement and other closed and limited spaces, guarantees the normal operation of the unmanned system, and realizes the task target.

Description

Technical Field

[0001] This invention relates to the field of unmanned system positioning technology, and more specifically to a method for locating unmanned systems in enclosed spaces under external interference. Background Technology

[0002] When conducting mining and other operations in enclosed spaces such as underground or indoor environments, the geological environment is extremely complex. Wireless positioning signals are subject to reflection, refraction, and multipath phenomena, causing the sensor error characteristics to no longer conform to a standard Gaussian distribution. Furthermore, wireless communication and wireless measurement are vulnerable to malicious attacks or interference from adversaries in complex adversarial environments. This can lead to unreliable signals received by the sensor, potentially causing intermittent interruptions in the measurement signal and worsening the positioning performance of the unmanned system. Therefore, inventing a positioning method for unmanned systems in enclosed spaces under external interference is particularly important.

[0003] Existing technologies, such as CN113324547, disclose a multi-AUV cooperative localization method based on an iterative extended RTS smoothing filter algorithm. This method uses the RTS optimal smoothing algorithm to correct the forward extended Kalman filter results, is simple and easy to implement, and effectively improves the navigation and positioning accuracy of the following AUVs. However, this method approximates process noise and measurement noise as zero-mean Gaussian white noise and assumes that the communication channel transmission is accurate and anomalies-free. This is not applicable to complex situations in confined downhole spaces where process noise and measurement noise do not follow a Gaussian distribution and are likely to be cross-coupled with network attacks.

[0004] Existing technology, such as CN115615433, discloses a hybrid localization method and system based on extended Kalman and RTS smoothing algorithms. This system uses the RTS smoothing algorithm to smooth the robot's position in a certain direction based on the system's position change error, and averages the smoothing results to obtain the optimal position estimate in that direction, effectively improving the estimation accuracy when the local direction is stationary. However, in complex environments with signal interference, the position measurement error increases, and the best smoothed position fitting result cannot be obtained. In addition, it only considers single Gaussian white noise and does not take into account the non-Gaussian nature of sensor measurement errors in complex and constrained environments.

[0005] Existing technology, such as CN118921622A, discloses an indoor positioning fusion method based on RTS smoothing of UWB and IMU. This method utilizes the complementary characteristics of UWB and IMU and employs extended Kalman filtering to fuse positioning information. Based on this, RTS smoothing is used to more accurately estimate the system state, overcoming the limitations of a single positioning system in enclosed environments and the excessive positioning error. However, in complex, enclosed environments, besides sensor inherent biases and accumulated integration errors leading to decreased positioning accuracy, the influence of non-Gaussian noise and interference from network attacks can also cause significant positional deviations when estimating unmanned systems.

[0006] The existing technology has the following shortcomings, specifically: 1. When facing non-Gaussian noise in the process and measurement, and when it is cross-coupled with unknown external interference, the existing technology struggles to achieve accurate detection and forward filtering positioning. This leads to inaccurate positioning of the unmanned system in complex environments, causing frequent path planning errors and deviations from the target during task execution, resulting in unstable operation and limiting its application in some special scenarios.

[0007] 2. Existing technologies have limitations in positioning principles and algorithms, lacking optimized methods like robust RTS smoothing technology to effectively handle complex data and noise interference. When performing filtered positioning, the processing methods for measurement data are relatively conventional, failing to accurately extract precise location information from raw data containing significant noise and interference. This leads to frequent positioning deviations and repeated calibrations during the process, severely impacting work efficiency and increasing human and material costs. Summary of the Invention

[0008] The purpose of this invention is to provide a positioning method for unmanned systems in enclosed spaces under external interference, which solves the problems existing in the background art.

[0009] To solve the above technical problems, the present invention adopts the following technical solution: The present invention provides a method for locating an unmanned system in a closed space under external interference, including: Step 1, constructing a model, considering multi-source external interference, introducing a flag signal that conforms to a Gaussian distribution, and constructing a state space model for the location problem of an unmanned system in a closed space.

[0010] Step 2: Interference detection, obtaining the predicted system state at time k-1 based on time updates. and prediction error covariance P k|k-1 The novelty of the flag signal is calculated using one-step prediction information, and the presence of interference is detected based on the chi-square statistic of the novelty. Corresponding measures are then taken based on the detection results.

[0011] Step 3: Measurement Update. The proposed generalized statistics are used to measure the similarity between expected and actual quantities, and the system state variables are calculated. With covariance matrix P k|k The state is updated based on the posterior estimation results.

[0012] Step 4: Backward smoothing. Using the estimation results from forward filtering, and within the proposed framework based on generalized statistics, the recursive estimation results of the system state are calculated. and P k|N This yields the positioning results corrected by the unmanned system.

[0013] Preferably, step 1, model construction, is specifically implemented as follows: Step 1-1: To model the positioning problem of an unmanned system in an actual underground enclosed space under unknown interference, a binary variable following a Bernoulli distribution is introduced. and δ k This indicates whether a denial-of-service attack or spoofed data interference has occurred, from which the standard state-space model is derived:

[0014] x k =F k x k-1 +w k ;

[0015]

[0016] Where, x k z represents the system state at time k. k F represents the sensor measurement at time k. k and H k Represents the state transition matrix and observation matrix, and the process noise w. k Measurement noise v k They are independent and their corresponding nominal covariance matrices are Q k and R k ω k This represents the probability of injected spoofed data interference. and the probability of denial-of-service disturbance All are unknowns. To avoid loss of generality, the interference from false data is modeled as a covariance matrix η with zero mean. k Non-Gaussian noise.

[0017] Step 1-2: Introduce a flag signal u independent of measurement noise. k As an extension term of the measurement model, its mean is the zero variance matrix B. k The pseudo-measurement model considering multi-source disturbances is expressed as follows:

[0018]

[0019] Where the pseudo-measurement matrix H j,k With H k Same, ω j,k This indicates unknown external interference.

[0020] Preferably, step 2, interference detection, is specifically implemented as follows: Step 2-1: Based on the positioning results of the unmanned system at the previous moment... and P k-1|k-1 Update the time:

[0021]

[0022] P k|k-1 =F k P k-1|k-1 F k T +Q k ;

[0023] in, and P k|k-1 These are the one-step state prediction of the system at time k and its prediction error covariance matrix, respectively.

[0024] Step 2-2: Calculate the chi-square standard value based on the innovation of the extended term according to the prediction results, and perform outlier detection by comparing it with the threshold.

[0025]

[0026] Where, γ k It is the residual based on the extended term. It is the expected covariance information of the sensor when it is not subject to external interference.

[0027] Steps 2-3: Standardize the innovation based on the chi-square distribution, introduce a chi-square threshold with confidence level ζ, and determine whether interference exists by comparing it with the threshold:

[0028]

[0029] Where θ is a threshold obtained from the chi-square distribution table based on the degrees of freedom of the measurement dimension and the confidence level.

[0030] Preferably, step 3, measurement update, is specifically implemented as follows: Step 3-1: Define generalized statistics:

[0031]

[0032] Where β1 and β2 are user-defined kernel function mixing coefficients, and σ and ε are kernel bandwidth and degrees of freedom parameters, respectively.

[0033] Step 3-2: Define the cost function based on generalized statistics, maximize the lower bound of the cost function using gradient descent, and obtain the optimal posterior estimate of the forward filtering using a fixed-point iterative algorithm.

[0034]

[0035] in, and P k|k It is the posterior state estimate and its estimation error covariance matrix, while K k This is the forward filter gain, and its expression is as follows:

[0036]

[0037] Among them, B p,k|k-1 and B v,k It is the Cholesky decomposition of the prediction covariance and the measurement noise covariance, I nx It is an identity matrix with the same dimension as the state vector, Π k It is an adjustment matrix.

[0038] Preferably, step 4, backward smoothing, is specifically implemented as follows: Step 4-1: Input the initial value of the forward filter. P k|k .

[0039] Step 4-2: Define the cost function based on the generalized statistics, and solve it using fixed-point iteration to obtain and output the optimal smoothing estimation result:

[0040]

[0041] in, and P k|N It is the smoothed estimate and its covariance information at the current moment, while G k It is the smoothing gain, and its expression is as follows:

[0042]

[0043] Among them, B w,k|k+1 and B p,k|k Q k+1 and P k|k Cholesky decomposition yields Θ k+1 It is the adjustment weight matrix for smoothing estimation, and β is the mixing coefficient of the generalized statistic.

[0044] The beneficial effects of this invention are as follows: 1. The closed-space unmanned system positioning method based on RTS smoothing under external interference proposed in this invention can still achieve accurate detection and forward filtering positioning even in complex and harsh environments where process noise and measurement noise are non-Gaussian noise and are cross-coupled with unknown external interference. Furthermore, robust RTS smoothing technology is used to further improve the accuracy of filtering positioning, which can meet the application requirements of extremely high positioning accuracy, such as high-precision positioning of equipment in underground mines. This enables unmanned systems to stably and accurately determine their position even in challenging environments, greatly expanding their application scenarios. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a flowchart of the method steps proposed in this invention.

[0047] Figure 2 This is a comparison chart of the positional root mean square error between the method proposed in this invention and existing maximum entropy filtering and smoothing methods and classical Kalman filtering and smoothing methods.

[0048] Figure 3 This is a comparison chart of the positional root mean square error between the method proposed in this invention and existing maximum entropy filtering and smoothing methods and classical Kalman filtering and smoothing methods.

[0049] Figure 4 This is a diagram illustrating a practical application scenario of the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Reference Figure 1 As shown, the present invention provides a method for locating an unmanned system in a closed space under external interference, including: Step 1, constructing a model, considering multi-source external interference, introducing a flag signal conforming to a Gaussian distribution, and constructing a state space model for the location problem of an unmanned system in a closed space.

[0052] In a specific embodiment, step 1, model construction, is specifically implemented as follows: Step 1-1: To model the positioning problem of an unmanned system in an actual underground enclosed space under unknown interference, a binary variable following a Bernoulli distribution is introduced. and δ k This indicates whether a denial-of-service attack or spoofed data interference has occurred, from which the standard state-space model is derived:

[0053] x k =F k x k-1 +w k ;

[0054]

[0055] Where, x k z represents the system state at time k. k F represents the sensor measurement at time k. k and H k Represents the state transition matrix and observation matrix, and the process noise w. k Measurement noise v k They are independent and their corresponding nominal covariance matrices are Q k and R k ω k This represents the probability of injected spoofed data interference. and the probability of denial-of-service disturbance All are unknowns. To avoid loss of generality, the interference from false data is modeled as a covariance matrix η with zero mean. k Non-Gaussian noise.

[0056] Step 1-2: Introduce a flag signal u independent of measurement noise. k As an extension term of the measurement model, its mean is the zero variance matrix B. k The pseudo-measurement model considering multi-source disturbances is expressed as follows:

[0057]

[0058] Where the pseudo-measurement matrix H j,k With H k Same, ω j,k This indicates unknown external interference.

[0059] Step 2: Interference detection, obtaining the predicted system state at time k-1 based on time updates. and prediction error covariance P k|k-1 The novelty of the flag signal is calculated using one-step prediction information, and the presence of interference is detected based on the chi-square statistic of the novelty. Corresponding measures are then taken based on the detection results.

[0060] In a specific embodiment, step 2, interference detection, is specifically implemented as follows: Step 2-1: Based on the positioning results of the unmanned system at the previous moment... and P k-1|k-1 Update the time:

[0061]

[0062] P k|k-1 =F k P k-1|k-1 F k T +Q k ;

[0063] in, and P k|k-1 These are the one-step state prediction of the system at time k and its prediction error covariance matrix, respectively.

[0064] Step 2-2: Calculate the chi-square standard value based on the innovation of the extended term according to the prediction results, and perform outlier detection by comparing it with the threshold.

[0065]

[0066] Where, γ k It is the residual based on the extended term. It is the expected covariance information of the sensor when it is not subject to external interference.

[0067] It should be noted that the novelty refers to the difference between the actual observed value and the predicted observed value in Kalman filtering, reflecting new information in the observation.

[0068] Steps 2-3: Standardize the innovation based on the chi-square distribution, introduce a chi-square threshold with confidence level ζ, and determine whether interference exists by comparing it with the threshold:

[0069]

[0070] Where θ is a threshold obtained from the chi-square distribution table based on the degrees of freedom of the measurement dimension and the confidence level.

[0071] Step 3: Measurement Update. The proposed generalized statistics are used to measure the similarity between expected and actual quantities, and the system state variables are calculated. With covariance matrix P k|k The state is updated based on the posterior estimation results.

[0072] In one specific embodiment, step 3, measurement update, is implemented as follows: Step 3-1: Define generalized statistics:

[0073]

[0074] Where β1 and β2 are user-defined kernel function mixing coefficients, and σ and ε are kernel bandwidth and degrees of freedom parameters, respectively.

[0075] It should be noted that the kernel function mixing coefficients are the coefficients used when linearly combining multiple kernel functions to form a mixed kernel function. These coefficients determine the proportion of each kernel function in the mixed kernel function. By adjusting the mixing coefficients, the advantages of different kernel functions can be combined, enabling the mixed kernel function to better adapt to various complex data distributions and problem types. Kernel bandwidth is a key parameter in a kernel function, determining the "width" or "smoothness" of the kernel function. Degrees of freedom parameters typically appear in kernel functions or statistical models with specific distributions, used to describe the shape and characteristics of the distribution, and to adjust the complexity and robustness of the model.

[0076] Step 3-2: Define the cost function based on generalized statistics, maximize the lower bound of the cost function using gradient descent, and obtain the optimal posterior estimate of the forward filtering using a fixed-point iterative algorithm.

[0077]

[0078] in, and P k|k It is the posterior state estimate and its estimation error covariance matrix, while K k This is the forward filter gain, and its expression is as follows:

[0079]

[0080] Among them, B p,k|k-1 and B v,k It is the Cholesky decomposition of the prediction covariance and the measurement noise covariance, I nx It is an identity matrix with the same dimension as the state vector, Π k It is an adjustment matrix.

[0081] Step 4: Backward smoothing. Using the estimation results from forward filtering, and within the proposed framework based on generalized statistics, the recursive estimation results of the system state are calculated. and P k|N This yields the positioning results corrected by the unmanned system.

[0082] In one specific embodiment, step 4, backward smoothing, is specifically implemented as follows: Step 4-1: Input the initial value of the forward filter. P k|k .

[0083] Step 4-2: Define the cost function based on the generalized statistics, and solve it using fixed-point iteration to obtain and output the optimal smoothing estimation result:

[0084]

[0085] in, and P k|N It is the smoothed estimate and its covariance information at the current moment, while G k It is the smoothing gain, and its expression is as follows:

[0086]

[0087] Among them, B w,k|k+1 and B p,k|k Q k+1 and P k|k Cholesky decomposition yields Θ k+1 It is the adjustment weight matrix for smoothing estimation, and β is the mixing coefficient of the generalized statistic.

[0088] It should be noted that the Cholesky decomposition is a method for decomposing a positive definite matrix into the product of a lower triangular matrix and its conjugate transpose.

[0089] The proposed method for unmanned system positioning in enclosed spaces under external interference, based on RTS smoothing, achieves accurate detection and forward filtering positioning even in complex and harsh environments where process and measurement noise are non-Gaussian and cross-coupled with unknown external interference. Furthermore, robust RTS smoothing technology further enhances the accuracy of the filtered positioning, meeting the demands of applications requiring extremely high positioning accuracy, such as high-precision equipment positioning in underground mines. This enables unmanned systems to stably and accurately determine their location even in challenging environments, greatly expanding their application scenarios.

[0090] The effectiveness of the method proposed in this invention was verified by using two-dimensional plane unmanned system positioning based on distance and speed sensors.

[0091] The state transition matrix is ​​chosen as follows:

[0092]

[0093] Where T = 1s is the sampling time, I2 is a 2×2 identity matrix, and O2 is a 2×2 zero matrix.

[0094] The observation matrix for the linear system is chosen as follows:

[0095] H k =[I202],

[0096] H j,k =[I202].

[0097] The system state vector is chosen as follows:

[0098] x k =[p x,k p y,k v x,k v y,k ] T ,

[0099] Where (p) x,k ,p y,k ) and (v x,k ,v y,k ) are the position and velocity in the x and y directions, respectively.

[0100] The initial state is chosen as follows:

[0101]

[0102] The initial covariance matrix is ​​chosen as follows:

[0103] P0 = diag[100m] 2 100m 2 10m 2 s -2 10m 2 s -2 ] T .

[0104] The mean root mean square error (ARMSE) is used to calculate the estimation errors of the proposed method, existing maximum entropy filtering and its smoothing method, existing variational Bayesian smoothing method based on Student's t-distribution, and classical (extended) Kalman filtering and its smoothing method. The formula for calculating the mean root mean square error of position is:

[0105]

[0106] RMSE pos The formula for calculating (k) is:

[0107]

[0108] Where the subscript 's' is the index of the Monte Carlo simulation experiment, The position detected by sensor k The estimated number of Monte Carlo simulation experiments, M c =500, the root mean square error of speed (ARMSE) vel The calculation method is similar to the method described above.

[0109] refer to Figure 2 , 3The presents the method proposed in this invention, existing maximum entropy filtering and smoothing methods, and classical Kalman filtering and smoothing methods for obtaining the average root mean square error of position and the average root mean square error of velocity in a simulated closed and complex scene. The comparison shows that the method proposed in this invention is significantly more accurate than the other existing methods.

[0110] refer to Figure 4 In a closed underground tunnel, the unmanned system uses sensors to measure the distance and position of the tag on the equipment. Based on the filtering algorithm proposed in this invention, the data is estimated and predicted to achieve accurate positioning, thereby obtaining the real motion information of the unmanned system.

[0111] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for locating an unmanned system in a closed space under external interference, characterized in that, include: Step 1: Construct a model, considering multi-source external interference, introduce a flag signal that conforms to a Gaussian distribution, and construct a state-space model for the localization problem of an unmanned system in a closed space. Step 2: Interference detection, obtaining the system status based on time updates. Predicted value at time and prediction error covariance The information of the flag signal is calculated using one-step prediction information, and the presence of interference is detected based on the chi-square statistic of the information. Corresponding measures are then taken based on the detection results. Step 3: Measurement Update. The proposed generalized statistics are used to measure the similarity between expected and actual quantities, and the system state variables are calculated. With covariance matrix The state is updated based on the posterior estimation results; Step 3, measurement update, is implemented as follows: Step 3-1: Define generalized statistics: ； in, and These are user-defined kernel function mixing coefficients. and These are the kernel bandwidth and degrees of freedom parameters, respectively. Step 3-2: Define the cost function based on generalized statistics, maximize the lower bound of the cost function using gradient descent, and obtain the optimal posterior estimate of the forward filtering using a fixed-point iterative algorithm. ； ； in, and It is the posterior state estimate and its estimation error covariance matrix, while This is the forward filter gain, and its expression is as follows: ； ； in, and It is the Cholesky decomposition of the prediction covariance and the measurement noise covariance. It is an identity matrix with the same dimensions as the state vector. It is an adjustment matrix; Step 4: Backward smoothing. Using the estimation results from forward filtering, and within the proposed framework based on generalized statistics, the recursive estimation results of the system state are calculated. and This yields the positioning results corrected by the unmanned system.

2. The method for locating an unmanned system in a closed space under external interference as described in claim 1, characterized in that, Step 1, building the model, is described in the following method: Step 1-1: To model the localization problem of an unmanned system in a real-world enclosed space under unknown disturbances, a binary variable following a Bernoulli distribution is introduced. and This indicates whether a denial-of-service attack or spoofed data interference has occurred, from which the standard state-space model is derived: ； ； in, Indicates in The system state at any given moment. Indicates in The sensor's measurement value at that moment. and Representing the state transition matrix and observation matrix, process noise Measurement noise They are independent and their corresponding nominal covariance matrices are respectively and , This represents the probability of injected spoofed data interference. and the probability of denial-of-service disturbance All variables are unknown. To avoid loss of generality, the interference from false data is modeled as a zero-mean covariance matrix. Non-Gaussian noise; Step 1-2: Introduce a flag signal independent of measurement noise. As an extension of the measurement model, its mean is a zero-variance matrix. The pseudo-measurement model considering multi-source disturbances is expressed as follows: ； The pseudo-measurement matrix and same, This indicates unknown external interference.

3. The method for locating an unmanned system in a closed space under external interference as described in claim 1, characterized in that, Step 2, interference detection, is performed as follows: Step 2-1: Based on the positioning results of the unmanned system at the previous moment and Update the time: ; ； in, and The system is in One-step state prediction at time t and its prediction error covariance matrix; Step 2-2: Calculate the chi-square standard value based on the innovation of the extended term according to the prediction results, and perform outlier detection by comparing it with the threshold. ； ； in, It is the residual based on the extended term. It is the expected covariance information of the sensor when it is not subject to external interference; Steps 2-3: Standardize the innovation based on the chi-square distribution and introduce a confidence level of... The chi-square threshold is used to determine whether interference exists by comparing it with the threshold: ； in, The threshold is obtained from the chi-square distribution table based on the degrees of freedom and confidence level of the measurement dimension.

4. The method for locating an unmanned system in a closed space under external interference as described in claim 1, characterized in that, Step 4, backward smoothing, is performed as follows: Step 4-1: Input the initial values ​​for the forward filter , ; Step 4-2: Define the cost function based on the generalized statistics, and solve it using fixed-point iteration to obtain and output the optimal smoothing estimation result: ； ； in, and It is the smoothed estimate and its covariance information at the current moment, while It is the smoothing gain, and its expression is as follows: ； ； in, and Each by and Cholesky decomposition yielded It is the adjustment weight matrix for smoothing estimation. It is the mixing coefficient of the generalized statistic.

Images