A vehicle cooperative localization method based on single snapshot arrival angle and power estimation

Abstract

The present invention relates to a vehicle cooperative positioning method based on single-snapshot direction-of-arrival (DOA) and power estimation, which includes receiving intelligent vehicle positioning signals by using uniform linear arrays in three cooperative base stations, determining the form of DOA estimation signals under each base station antenna array in a single-snapshot scenario; obtaining DOA estimation results through the generalized approximate message passing (GAMP) algorithm; dividing the spatial grid in detail based on the DOA estimation results, and obtaining power estimation results through the weighted l1-norm minimization method. A decision function is constructed, and the power estimation results and all DOA estimation results obtained by each base station are used for optimization. Finally, the position information of the intelligent vehicle is obtained based on the optimized DOA estimation results by using the triangulation method. Compared with the existing technical methods, the method of the present invention not only has low complexity, but also can provide higher resolution and positioning accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle positioning, and in particular to a vehicle collaborative positioning method based on single-snapshot arrival angle and power estimation. Background Art

[0002] Intelligent vehicle positioning is a very important issue in the research of autonomous driving. High-precision positioning technology not only makes it easier to track smart cars, but also enables the interconnection of vehicles to vehicles (V2V), vehicles to roadside infrastructure (V2I), and vehicles to urban networks based on location sharing, which is the only way to achieve intelligent transportation. Although the Global Navigation Satellite System (GNSS) is the most widely used and mature positioning technology today, in the field of autonomous driving, satellite signals are still easily affected by the complex surrounding environment, resulting in positioning failure, and the positioning update frequency based on GNSS is low, which cannot meet the problem of vehicle positioning at high speeds.

[0003] Direction of Arrival (DOA) estimation technology has high-precision positioning performance and is an important application in sonar, radar, telecommunications and other fields. Its goal is to locate one or more radiation sources in the propagation medium. Using DOA estimation technology to locate vehicles based on antenna arrays composed of advanced sensors is one of the mainstream solutions for the current research of intelligent vehicle positioning technology. Among these solutions, the single-snapshot DOA estimation positioning technology has become one of the positioning technologies with wide application value due to its advantages of high precision and low complexity. It can locate intelligent vehicles with high precision and real-time positioning. It plays a vital role in improving the accuracy and speed of vehicle positioning technology, and can make up for the shortcomings of GNSS-based positioning technology in a variety of application scenarios.

[0004] In recent years, some representative solutions for vehicle positioning methods based on DOA estimation technology have emerged. For example, Rui Zhang et al. proposed a vehicle positioning method based on TOA and DOA joint estimation of V2R communications in the paper "A Vehicle Positioning Method Based on Joint TOA and DOA Estimation with V2R Communications" in 2017; Huafei Wang et al. proposed a robust DOA estimation vehicle positioning technology based on SBL on three collaborative base stations in the paper "Assistant Vehicle Localization Based on Three Collaborative Base Stations via SBL-Based Robust DOAEstimation" in 2019. However, after analysis, it was found that the above two solutions have high computational complexity and require a certain number of snapshots to ensure the accuracy of positioning, making it difficult to achieve real-time positioning in actual applications. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide an intelligent vehicle positioning method based on single snapshot arrival angle and power estimation, which has low computational complexity, high positioning accuracy, real-time positioning and can ensure the robustness of vehicle positioning in single snapshot scenario applications.

[0006] The technical solution adopted by the present invention is a vehicle cooperative positioning method based on single snapshot arrival angle and power estimation, which comprises the following steps:

[0007] S1. Arrange three cooperative base stations in an equilateral triangle, divide the initial spatial grid, and form an initial overcomplete basis matrix A based on the initial spatial grid; set the positioning signals emitted by K vehicles to be incident on the arrays of the three cooperative base stations, and obtain a single snapshot reception signal of the arrays of the three cooperative base stations;

[0008] S2. Based on the initial overcomplete basis matrix A and the single snapshot received signal, a generalized approximate message passing algorithm is used to obtain arrival angle estimation values ​​of the three cooperative base stations, where the arrival angle estimation values ​​are the angular positions of the K vehicles corresponding to each cooperative base station;

[0009] S3, finely draw a spatial grid around each arrival angle estimate obtained in step S2 to form an updated complete basis matrix B i,k and its corresponding weight matrix, where i = 1, 2, 3, k = 1, 2, ..., K, according to the weight matrix, using weighted l 1 The signal power estimation is obtained by using the norm minimization optimization method;

[0010] S4. Based on the signal power estimation values ​​obtained from the three cooperative base stations, a decision function is constructed, and the arrival angle is optimized through the decision function to obtain the preferred arrival angle. Based on the preferred arrival angle, the vehicle positioning is completed using the triangulation positioning method.

[0011] Compared with the prior art, the invention has the following advantages: the method of the invention is based on the GAMP algorithm and 1 The norm minimization method obtains DOA estimation results and signal power estimation results. By constructing a decision function, two cooperative base stations with better estimation effects are selected from three cooperative base stations for vehicle positioning, which effectively improves the accuracy of DOA estimation, power estimation and matching intelligent vehicle positioning. In addition, the method of the present invention is suitable for the situation of unknown positioning signal probability distribution, and obtains low-complexity DOA estimation through a scalar estimation function, which effectively guarantees the real-time requirements of intelligent vehicle positioning. After obtaining the DOA estimation results provided by the three cooperative base stations, the method provided by the present invention selects the preferred DOA estimation result through auxiliary information and then performs vehicle positioning, which effectively avoids the influence of unstable DOA on vehicle positioning performance.

[0012] Preferably, the specific process of step S1 includes the following steps:

[0013] S1.1, let the coordinates of the three cooperative base stations be (0,0)m, (D,0)m and (D / 2, )m, m represents meters, and the spatial angle range [-90°, 90°] is evenly divided into N grids Forming an initial overcomplete basis matrix based on N grids in, Where n = 1, 2, ..., N, M represents an antenna array with M equally spaced elements, λ and d represent the carrier wavelength and the element spacing, respectively;

[0014] S1.2. Assume that the positioning signals emitted by K vehicles are incident on the arrays of three cooperative base stations. After analog-to-digital sampling, the single-snapshot received signals of the arrays of the three cooperative base stations are obtained. The single-snapshot array received signal y of the array of the i-th cooperative base station is i (t) is expressed as:

[0015]

[0016] in, represents the K sparse signal vector, n i (t) represents the Gaussian white noise vector, t represents the sampling time.

[0017] Preferably, the specific process of step S3 includes the following steps:

[0018] S3.1. Set the arrival angle estimation values ​​of the three cooperative base stations obtained in step S2 Respectively expressed as: Where i = 1, 2, 3, in each The surrounding space grid is finely divided, and the The surrounding fine-grained spatial grid is represented as: According to the finely divided spatial grid, an updated complete basis matrix B for the array antennas of the three cooperative base stations is formed. i,k , where i = 1, 2, 3, k = 1, 2, ..., K, the updated complete basis matrix B i,k The specific expression is:

[0019]

[0020] S3.2. Constructing the weight matrix W i,k , its specific expression is:

[0021] S3.3, based on the weight matrix W obtained in step S3.2 i,k Construct the optimization objective function:

[0022] Among them, C i =[B i,1 ,...B i,K ], λ represents a parameter for balancing sparsity and accuracy, and power estimation values ​​corresponding to the three cooperative base stations are obtained by solving the optimization objective function.

[0023] Preferably, the specific process of step S4 includes the following steps:

[0024] S4.1. For the positioning signal transmitted by the kth vehicle, set the estimated angle of arrival obtained in step S2 to The power estimate obtained in step S3.3 is p i,k , construct the decision function in is the noise power,

[0025] S4.2, respectively substitute the arrival angle estimation value obtained in step S2 and the power estimation value obtained in step S3.3 into the decision function, and respectively obtain the decision values ​​of the three base stations regarding the K arrival angles; according to the obtained decision values, if the sizes of any two of the three decision values ​​are inconsistent, then select the arrival angles of the two base stations with smaller decision values ​​as the preferred arrival angles and use them for vehicle positioning; if the sizes of the three decision values ​​are consistent, then randomly select the arrival angles of two base stations as the preferred arrival angles and use them for vehicle positioning;

[0026] S4.3, let the preferred arrival angle obtained in step S4.2 be ω k ,β k , if the coordinates of the corresponding preferred base stations are (D / 2, )m, (0,0)m, then the corresponding vehicle coordinates (L x ,L y )for:

[0027]

[0028] If the coordinates of the corresponding preferred base station are (0,0)m and (D,0)m, then the corresponding vehicle coordinates (L x ,L y )for:

[0029]

[0030] If the coordinates of the corresponding preferred base station are (D / 2, )m, (D,0)m, then the corresponding vehicle coordinates (L x ,L y )for:

[0031] BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a power spectrum diagram of a vehicle cooperative positioning method based on single snapshot arrival angle and power estimation and a GAMP algorithm for DOA estimation proposed in a specific embodiment of the present invention;

[0033] Figure 2 is a comparison diagram of the mean square error of positioning according to the signal-to-noise ratio of a vehicle cooperative positioning method based on single-snapshot arrival angle and power estimation proposed in a specific embodiment of the present invention;

[0034] Figure 3 This is a flow chart of a vehicle collaborative positioning method based on single snapshot arrival angle and power estimation proposed by the present invention. DETAILED DESCRIPTION

[0035] The invention will be further described below with reference to the accompanying drawings and in combination with specific implementations, so that those skilled in the art can implement the invention with reference to the description. The protection scope of the invention is not limited to the specific implementations.

[0036] The embodiment of the present invention provides a vehicle cooperative positioning method based on single snapshot arrival angle and power estimation, and the implementation steps are as follows:

[0037] S1. Arrange three cooperative base stations in an equilateral triangle, divide the initial spatial grid, and form an initial overcomplete basis matrix A based on the initial spatial grid; set the positioning signals emitted by K vehicles to be incident on the arrays of the three cooperative base stations, and obtain a single snapshot reception signal of the arrays of the three cooperative base stations;

[0038] In this embodiment, the specific process of determining the DOA estimation signal form under each cooperative base station antenna array includes the following steps:

[0039] S1.1, let the coordinates of the three cooperative base stations be (0,0)m, (D,0)m and (D / 2, )m, divide the space [-90°, 90°] angle range evenly into N grids Forming an initial overcomplete basis matrix based on N grids Take an antenna array with M equally spaced elements as an example. Expressed as

[0040]

[0041] Where λ and d are the carrier wavelength and array element spacing, respectively, and the superscript T represents the transpose operation.

[0042] S1.2. Assume that the positioning signals emitted by K vehicles are incident on the arrays of three cooperative base stations. After analog-to-digital sampling, the single-snapshot received signals of the arrays of the three cooperative base stations are obtained. The single-snapshot array received signal y of the array of the i-th cooperative base station is i (t) is expressed as:

[0043]

[0044] in, is a K-sparse signal vector. If the angle of the incident signal is Then i (t) is non-zero, and the other elements are zero, is a Gaussian white noise vector, t represents the sampling time;

[0045] S2, based on A and the single snapshot received signal of the base station antenna array, the arrival angle estimation values ​​corresponding to the three cooperative base stations are obtained by using the generalized approximate message passing (GAMP) algorithm;

[0046] In this embodiment, the specific process of obtaining the arrival angle estimation values ​​corresponding to the three cooperative base stations by using the generalized approximate message passing (GAMP) algorithm based on A and the single snapshot received signal of the base station antenna array is as follows: Construct an overcomplete matrix A, a Bernoulli Gaussian input distribution model, a Gaussian output distribution model, determine the GAMP input function and output function, and run the GAMP algorithm to obtain the arrival angle estimation values ​​corresponding to the three cooperative base stations;

[0047] S3, finely draw a spatial grid around the estimated value of the arrival angle in step S2 to form an updated complete basis matrix B i,k and the corresponding weight matrix, and based on the weighted l 1 The norm minimization optimization method is used to obtain the arrival angle and signal power estimation;

[0048] In this embodiment, a spatial grid is drawn around the estimated arrival angle value in step S2, the complete basis matrix and the corresponding weight matrix are updated, and the weight matrix is ​​calculated based on the weighted matrix. 1 The specific process of obtaining arrival angle and signal power estimation by norm minimization optimization method includes the following steps:

[0049] S3.1, suppose the arrival angle groups corresponding to base stations 1 to 3 obtained in step S2 are and Then respectively in The surrounding fine space grid is:

[0050]

[0051] Then, an updated complete basis matrix B for the array antennas of base stations 1 to 3 is formed. i,k , where i = 1, 2, 3, k = 1, 2, ..., K, B i,k The specific expression is:

[0052]

[0053] S3.2. Constructing the weight matrix W i,k , its specific expression is:

[0054]

[0055] S3.3, based on step S3.2, the weight matrix W obtained i,k Constructing the optimization objective function Among them, C i =[B i,1 ,...B i,K ], λ is a parameter that balances sparsity and accuracy. After solving, we get the power estimation values ​​corresponding to the three cooperative base stations.

[0056] S4. Optimize the angle of arrival through a decision function based on the arrival angle and power estimation results obtained from the three cooperative base stations, and complete the vehicle positioning using the triangulation positioning method based on the optimal arrival angle.

[0057] In this embodiment, based on the arrival angle and power estimation results obtained by the three cooperative base stations, the arrival angle is optimized through the decision function, and the specific process of completing the vehicle positioning by using the triangulation positioning method based on the optimal arrival angle includes the following steps:

[0058] S4.1. For the positioning signal transmitted by the kth vehicle, set the estimated angle of arrival obtained in step S2 to The power estimate obtained in step S3.3 is p i,k , construct the decision function in is the noise power,

[0059] S4.2, respectively substitute the arrival angle estimation value obtained in step S2 and the power estimation value obtained in step S3.3 into the function decision function, and respectively obtain the decision values ​​of the three base stations regarding the K arrival angles; according to the obtained decision values, if the sizes of any two of the three decision values ​​are inconsistent, then select the arrival angles of the two base stations with smaller decision values ​​as the preferred arrival angles and use them for vehicle positioning; if the sizes of the three decision values ​​are consistent, then randomly select the arrival angles of two of the base stations as the preferred arrival angles and use them for vehicle positioning;

[0060] S4.3, let the preferred arrival angle obtained in step S4.2 be ω k ,β k , if the coordinates of the corresponding preferred base stations are (D / 2, )m, (0,0)m, then the corresponding vehicle coordinates (L x ,L y )for:

[0061]

[0062] If the coordinates of the corresponding preferred base station are (0,0)m and (D,0)m, then the corresponding vehicle coordinates (L x ,L y )for:

[0063]

[0064] If the coordinates of the corresponding preferred base station are (D / 2, )m, (D,0)m, then the corresponding vehicle coordinates (L x ,L y )for:

[0065]

[0066] The following simulation experiments are used to analyze the positioning performance and computational effectiveness of a vehicle collaborative positioning method based on single snapshot arrival angle and power estimation proposed in an embodiment of the present invention. The simulation process is carried out using MATLAB software, and the computer configuration is a ThinkpadX1 laptop computer with i5-10210UCPU and 8GDDR4RAM.

[0067] Simulation experiment 1: The DOA estimation is simulated by using the method proposed in the embodiment of the present invention and the GAMP method in the prior art; the number of array elements in each cooperative base station is the same and is 80, the number of sampling snapshots is 1, the spatial grid search range is -30° to 30°, and the search step is 0.1°. Consider the following scenario: under the condition of 0 signal-to-noise ratio, the incoming wave directions are -8° and 10°, and the power of the two positioning signals is set to 0.99. The simulation results are as follows: Figure 1 As shown. Figure 1 It can be seen that the method of the present invention can achieve a relatively accurate power estimation result. On the other hand, due to the unknown specific distribution information of the positioning signal and the noise signal, it is difficult for the GAMP algorithm to obtain accurate results in power estimation, which fully demonstrates the effectiveness and excellence of the method of the present invention.

[0068] Simulation experiment 2: The relationship between the mean square error (MMSE, unit: meter (m)) and the signal-to-noise ratio of vehicle positioning is simulated by using the method proposed in the embodiment of the present invention and the GAMP method in the prior art. The positions of the three 5G cooperative base stations are B and 1 (50m, ),B 2 (0m,0m),B 3 (100m, 0m), the number of array elements in each cooperative base station is the same and is 80, the number of sampling snapshots is 1, the spatial grid search range is -30°~30°, and the search step is 0.1°. The coordinates of the intelligent vehicle are located at S1 (50m, 13.3975m). Consider the following scenario. The received signal-to-noise ratio of the cooperative base stations is different. The received signal-to-noise ratio of the first cooperative base station is 5 dB lower than that of the second cooperative base station, and also 5 dB lower than that of the third cooperative base station. The simulation results when the received signal-to-noise ratio of the second cooperative base station changes from -3 dB to 9 dB are as follows Figure 2 As shown. Figure 2 It can be seen that the positioning performance of the method of the present invention is better than the GAMP average method in all aspects. At the same time, it can achieve centimeter-level positioning accuracy, which fully demonstrates the effectiveness and excellence of the method of the present invention.

Claims

1. A vehicle cooperative localization method based on single snapshot arrival angle and power estimation, Features: The method comprises the following steps: S1. Arrange three cooperative base stations in an equilateral triangle, divide the initial spatial grid, and form an initial overcomplete basis matrix A based on the initial spatial grid; set the positioning signals emitted by K vehicles to be incident on the arrays of the three cooperative base stations, and obtain a single snapshot reception signal of the arrays of the three cooperative base stations; S2. Based on the initial overcomplete basis matrix A and the single snapshot received signal, a generalized approximate message passing algorithm is used to obtain arrival angle estimation values ​​of the three cooperative base stations, where the arrival angle estimation values ​​are the angular positions of the K vehicles corresponding to each cooperative base station; S3, finely draw a spatial grid around each arrival angle estimate obtained in step S2 to form an updated complete basis matrix B i,k and its corresponding weight matrix, where i = 1, 2, 3, k = 1, 2, ..., K, according to the weight matrix, using weighted l 1 The signal power estimation value is obtained by the norm minimization optimization method; the specific process is as follows: S3.

1. Set the arrival angle estimation values ​​of the three cooperative base stations obtained in step S2 Respectively expressed as: and Where i = 1, 2, 3, in each The surrounding space grid is finely divided, and the The surrounding fine-grained spatial grid is represented as: According to the finely divided spatial grid, an updated complete basis matrix B for the array antennas of the three cooperative base stations is formed. i,k , where i = 1, 2, 3, k = 1, 2, ..., K, the updated complete basis matrix B i,k The specific expression is: S3.

2. Constructing the weight matrix W i,k , its specific expression is: S3.3, based on the weight matrix W obtained in step S3.2 i,k Construct the optimization objective function: Among them, C i =[B i,1 ,...B i,K ], λ represents a parameter for balancing sparsity and accuracy, and power estimation values ​​corresponding to the three cooperative base stations are obtained by solving the optimization objective function; S4. Based on the signal power estimation values ​​obtained by the three cooperative base stations, a decision function is constructed, and the optimal angle of arrival is optimized through the decision function to obtain the optimal angle of arrival. Based on the optimal angle of arrival, the vehicle positioning is completed using the triangulation positioning method. The specific process is as follows: S4.

1. For the positioning signal transmitted by the kth vehicle, set the estimated angle of arrival obtained in step S2 to The power estimate obtained in step S3.3 is p i,k , construct the decision function in is the noise power, S4.2, respectively substitute the arrival angle estimation value obtained in step S2 and the power estimation value obtained in step S3.3 into the function decision function, and respectively obtain the decision values ​​of the three base stations regarding the K arrival angles; according to the obtained decision values, if the sizes of any two of the three decision values ​​are inconsistent, then select the arrival angles of the two base stations with smaller decision values ​​as the preferred arrival angles and use them for vehicle positioning; if the sizes of the three decision values ​​are consistent, then randomly select the arrival angles of two of the base stations as the preferred arrival angles and use them for vehicle positioning; S4.3, let the preferred arrival angle obtained in step S4.2 be ω k ,β k , if the coordinates of the corresponding preferred base stations are (D / 2, )m, (0,0)m, then the corresponding vehicle coordinates (L x ,L y )for: If the coordinates of the corresponding preferred base station are (0,0)m and (D,0)m, then the corresponding vehicle coordinates (L x ,L y )for: If the coordinates of the corresponding preferred base station are (D / 2 , )m, (D,0)m, then the corresponding vehicle coordinates (L x ,L y )for:

2. The vehicle cooperative positioning method based on single snapshot arrival angle and power estimation according to claim 1, Features: The specific process of step S1 includes the following steps: S1.1, let the coordinates of the three cooperative base stations be (0,0)m, (D,0)m and (D / 2, )m, m represents meters, and the spatial angle range [-90°, 90°] is evenly divided into N grids Forming an initial overcomplete basis matrix based on N grids in, Where n = 1, 2, K, N, M represents an antenna array with M equally spaced elements, λ and d represent the carrier wavelength and the element spacing respectively; S1.

2. Assume that the positioning signals emitted by K vehicles are incident on the arrays of three cooperative base stations. After analog-to-digital sampling, the single-snapshot received signals of the arrays of the three cooperative base stations are obtained. The single-snapshot array received signal y of the array of the i-th cooperative base station is i (t) is expressed as: in, represents the K sparse signal vector, n i (t) represents the Gaussian white noise vector, t represents the sampling time.

Images