Long tunnel environment train positioning method based on UWB

Abstract

The invention discloses a long tunnel environment train positioning method based on UWB, and belongs to the technical field of train positioning, and the method comprises the steps: arranging UWB positioning base stations above a tunnel wall in a rectangular manner, and fixing UWB labels at the top end of a train; collecting RSSI data of the train and the UWB positioning base station at different distances to obtain preprocessed RSSI data; constructing a signal large-scale fading model, and performing parameter optimization on the signal large-scale fading model by using MEGNDO according to the preprocessed RSSI data to obtain a UWB ranging model suitable for a tunnel scene; in the running process of the train, the distance between the UWB positioning base station and the train is calculated by using a UWB distance measurement model according to RSSI data received by the UWB equipment; and according to the distance from the UWB positioning base station to the train, obtaining the position coordinates of the train by using a trilateral positioning method. The problems that satellite signals cannot be received in a tunnel environment and real-time positioning is difficult to carry out are solved.

Description

Technical Field

[0001] The invention belongs to the technical field of train positioning, and in particular relates to a UWB-based train positioning method in a long tunnel environment. Background Art

[0002] Due to the obstruction of buildings in the tunnel, the train cannot receive satellite signals, and the traditional train positioning methods such as track circuits and transponders have high maintenance costs. The common inertial navigation system will also cause the positioning results to drift in long tunnels due to long-term lack of correction. In the train control system, the realization of almost every sub-function requires the train's position information as one of the parameters. In order to ensure the accuracy of train positioning in satellite-blocked environments, and to promptly detect and prevent accidents such as train failures, collisions and derailments, UWB is introduced as a key technology for train positioning in tunnel environments, and the RSSI method that can provide a longer positioning distance is adopted. Summary of the invention

[0003] In view of the above-mentioned deficiencies in the prior art, the present invention provides a UWB-based train positioning method in a long tunnel environment, which solves the problem that satellite signals cannot be received in a tunnel environment and it is difficult to perform real-time positioning.

[0004] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a UWB-based train positioning method in a long tunnel environment, comprising the following steps:

[0005] S1. Arrange the UWB positioning base station in a rectangular shape above the tunnel wall, and fix the UWB tag on the top of the train;

[0006] S2, collect RSSI data between the train and the UWB positioning base station at different distances;

[0007] S3, performing threshold denoising and mean filtering on the RSSI data to obtain pre-processed RSSI data;

[0008] S4. Construct a signal large-scale fading model for describing the UWB signal propagation process in the tunnel scenario, and optimize the parameters of the signal large-scale fading model using the MEGNDO optimization algorithm based on the pre-processed RSSI data to obtain a UWB ranging model.

[0009] S5. When the train is running, RSSI data received by the UWB device is obtained, and the distance from the UWB positioning base station to the train is calculated using the UWB ranging model according to the RSSI data received by the UWB device;

[0010] S6. According to the distance from the UWB positioning base station to the train, the train position coordinates are obtained using the trilateral positioning method.

[0011] The beneficial effects of the present invention are as follows: the MEGNDO method proposed in the present invention fits the path loss of the tunnel environment according to the measured data, and improves the traditional RMSE-based objective function to solve the problem of large ranging errors caused by the appearance of the far-zone waveguide effect in the tunnel, thereby obtaining a UWB signal propagation model that is more suitable for the tunnel environment, thereby obtaining higher train positioning accuracy, and performing train integrity checks at the same time to prevent the carriages from being disconnected and stranded on the line or track in the travel section, thereby ensuring the safety of train operation and improving the transportation efficiency and reliability of the train. The present invention has high positioning accuracy and can locate the train in a tunnel environment, solving the problem of being unable to receive satellite signals and difficulty in real-time positioning in a tunnel environment, reducing the use of trackside equipment in traditional train positioning methods, improving the maintenance efficiency of train positioning equipment, reducing maintenance costs, checking train integrity, and ensuring train travel safety.

[0012] Furthermore, fixing the UWB tag on the top of the train in step S1 specifically involves fixing two UWB tags on the top of the center of the first carriage and the last carriage of the train respectively.

[0013] The beneficial effect of the above further solution is: installing UWB tags and receiving information in real time to realize the distance measurement between the train and the UWB positioning base station.

[0014] Furthermore, the step S4 is specifically as follows:

[0015] S401, construct a large-scale signal fading model for describing the UWB signal propagation process in the tunnel scenario:

[0016] L s =L1+L2lgd+L3lgdH b +L4diffraction+L5lgdlgH b +L6H m +L7Clutter

[0017] Among them, L s is the path loss; L1 is the frequency correction factor; L2 is the distance attenuation factor; d is the distance from the UWB tag to the UWB positioning base station; L3 is the base station antenna height factor; H b is the antenna height of the UWB positioning base station; L4 is the diffraction loss correction factor; diffraction is the diffraction loss; L5 is the transmitting antenna and distance correction factor; L6 is the receiving antenna correction factor; H m is the height of the UWB tag antenna; L7 is the terrain loss factor; Clutter is the average weighted loss of the ground object; lg is the logarithmic function;

[0018] S402, by improving the population initialization strategy of the GNDO generalized normal distribution optimization algorithm, and adding periodic global mutation disturbance and multi-condition selection evolution strategy, the MEGNDO optimization algorithm is obtained;

[0019] S403: According to the pre-processed RSSI data, the MEGNDO optimization algorithm is used to optimize the parameters of the signal large-scale fading model to obtain a UWB ranging model.

[0020] The beneficial effects of the above further scheme are as follows: the MEGNDO optimization algorithm proposed in the present invention improves the shortcomings of slow convergence speed and premature maturity of traditional GNDO. At the same time, when performing SPM parameter optimization, the traditional RMSE-based objective function is modified, and the piecewise weighted idea is introduced to reduce the ranging error caused by the waveguide effect in the far zone through weight optimization.

[0021] Furthermore, the population initialization strategy for improving the GNDO generalized normal distribution optimization algorithm in step S402 is specifically: combining the Sine chaotic mapping with the piecewise linear chaotic mapping to construct a composite chaotic mapping strategy, and replacing the population initialization strategy of the GNDO generalized normal distribution optimization algorithm with the composite chaotic mapping strategy, thereby completing the improvement of the population initialization strategy of the GNDO generalized normal distribution optimization algorithm.

[0022] The beneficial effects of the above further scheme are: in order to address the defects of poor traversal and randomness of GNDO population initialization, a composite chaotic mapping method is used to traverse the solution space of the entire population, thereby enriching the diversity of the algorithm population and greatly improving the randomness of population initialization.

[0023] Furthermore, the expression of the composite chaotic mapping strategy is:

[0024]

[0025] Where f() is the composite chaotic mapping strategy function; mod is the modulus symbol; κ i is the result of chaotic mapping of the ith individual; α and β are both chaotic parameters; sin is the sine function; π is the pi; rand is a random number between (0,1);.

[0026] The beneficial effects of the above further scheme are: in order to address the defects of poor traversal and randomness of GNDO population initialization, a composite chaotic mapping method is used to traverse the solution space of the entire population, thereby enriching the diversity of the algorithm population and greatly improving the randomness of population initialization.

[0027] Furthermore, the expression of the periodic global variation disturbance in step S402 is:

[0028]

[0029]

[0030]

[0031] Among them, v new is the updated tracking vector; mod is the modulus symbol; is the tracking vector of the i-th individual at the t-th iteration; is the position of the i-th individual at the t-th iteration; is an adjustment parameter with a value between 0 and 1; λ3 and λ4 are both random numbers that obey the standard normal distribution; v1 and v2 are both tracking vectors; and is the position of three randomly selected individuals at the tth iteration; Fit() is the fitness function; V is the disturbance factor; k is the global shrinkage factor; p is the reverse factor; x max is the maximum value at the tth iteration; x min is the minimum value at the tth iteration; rand is a random number between (0,1); T c is the global disturbance criterion. When it is 1, the global mutation disturbance strategy is executed, otherwise, the global mutation disturbance strategy is not executed; t is the number of iterations; T is the maximum number of iterations.

[0032] The beneficial effects of the above further scheme are: proposing a periodic global variation perturbation strategy, which solves the defect that GNDO is prone to fall into local optimization for complex problems, improves the global optimization performance of GNDO, and jumps out of local optimization in time. At the same time, the periodic idea is used for strategy embedding to ensure the continuity of problem solving, effectively avoid local optimization and improve the optimization accuracy of the algorithm.

[0033] Furthermore, the multi-condition selection evolution strategy in step S402 specifically comprises first performing condition judgment, and iteratively evolving the individual positions that meet the conditions; the expressions for condition judgment and iteratively evolving the individual positions are respectively:

[0034]

[0035] x new =E f ×v(index(randperm(3,1)))

[0036] E f =abs(randn) / e 2 +1 / 2

[0037] Among them, Judgment is a conditional judgment, which needs to satisfy both condition 1Con1 and condition 2Con2; Fit() is the fitness function; x new is the updated individual position; x i is the position of the ith individual; Sort() is the sorting; Npop is the population size; E f is the evolution factor; v(index(i)) is the position of the individual with selection sorting i; index is the sorting function; randperm(3,1) is to select 1 individual from 3 individuals; abs is the absolute value; randn is a random number that conforms to the normal distribution; e is a natural constant.

[0038] The beneficial effects of the above further scheme are: it proposes to use evolutionary strategies to selectively iterate and evolve individual positions, which improves the optimization accuracy and convergence speed of GNDO and the superiority of the optimization cluster. Considering that the solution performance is closely related to the average performance of the population, it is extremely important to avoid the emergence of poor individuals. An individual evolution strategy is proposed, and a multi-condition selection strategy is used to avoid evolving better individuals, which greatly increases the optimization accuracy and convergence speed of GNDO, making the algorithm optimization performance even better.

[0039] Furthermore, the objective function of parameter optimization in step S4 is:

[0040]

[0041]

[0042] w=σ(d / 100)

[0043] σ(d / 100)=1 / (1+e -d / 100 )

[0044] Among them, J new (d) is the objective function for parameter optimization; d is the distance from the UWB tag to the UWB positioning base station; n is the number of samples; e j is the error between the predicted data and the actual measurement of the jth sample; w j is the weight of the jth sample; e is the error between the RSSI data predicted by the large-scale fading model of the signal and the actual measured RSSI data; L dB It is the actual measured RSSI data; is the RSSI data predicted by the large-scale signal fading model; Pos is the optimal parameter; w is the weight vector; σ() is the sigmoid function; e is a natural constant.

[0045] The beneficial effects of the above further scheme are: using the S-shaped function to adjust the impact of the error on different distances solves the problem of large distance calculation errors in the far field. When using MEGNDO for optimization, the traditional RMSE-based objective function is adaptively improved to make the search results more consistent with the propagation characteristics of UWB signals in tunnel scenarios.

[0046] Furthermore, the expression of the three-side positioning method in step S6 is:

[0047]

[0048] Among them, x is the horizontal coordinate of the UWB tag; y is the vertical coordinate of the UWB tag; z is the vertical coordinate of the UWB tag; x1 is the horizontal coordinate of the first selected UWB positioning base station; x2 is the horizontal coordinate of the second selected UWB positioning base station; x3 is the horizontal coordinate of the third selected UWB positioning base station; x4 is the horizontal coordinate of the fourth selected UWB positioning base station; y1 is the vertical coordinate of the first selected UWB positioning base station; y2 is the vertical coordinate of the second selected UWB positioning base station; y3 is the vertical coordinate of the third selected UWB positioning base station; y4 is the horizontal coordinate of the fourth selected UWB positioning base station. The vertical coordinate of the selected UWB positioning base station is z1; the vertical coordinate of the first selected UWB positioning base station is z2; the vertical coordinate of the second selected UWB positioning base station is z3; the vertical coordinate of the third selected UWB positioning base station is z4; the vertical coordinate of the fourth selected UWB positioning base station is d1; the distance from the first selected UWB positioning base station to the UWB tag is d2; the distance from the second selected UWB positioning base station to the UWB tag is d3; the distance from the third selected UWB positioning base station to the UWB tag is d4; the distance from the fourth selected UWB positioning base station to the UWB tag.

[0049] The beneficial effect of the above further solution is that the position information of the train in the tunnel environment is obtained according to the trilateration method. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 The present invention is a flow chart of the method.

[0051] Figure 2 It is a schematic diagram of a train positioning unit in a tunnel environment according to the present invention.

[0052] Figure 3 This is a diagram of the positioning test results of the present invention in a tunnel environment. DETAILED DESCRIPTION

[0053] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0054] like Figure 1 As shown, in one embodiment of the present invention, a UWB-based train positioning method in a long tunnel environment includes the following steps:

[0055] S1. Arrange the UWB positioning base station in a rectangular shape above the tunnel wall, and fix the UWB tag on the top of the train;

[0056] S2, collect RSSI data between the train and the UWB positioning base station at different distances;

[0057] S3, performing threshold denoising and mean filtering on the RSSI data to obtain pre-processed RSSI data;

[0058] S4. Construct a signal large-scale fading model for describing the UWB signal propagation process in the tunnel scenario, and optimize the parameters of the signal large-scale fading model using the MEGNDO optimization algorithm based on the pre-processed RSSI data to obtain a UWB ranging model.

[0059] S5. When the train is running, RSSI data received by the UWB device is obtained, and the distance from the UWB positioning base station to the train is calculated using the UWB ranging model according to the RSSI data received by the UWB device;

[0060] S6. According to the distance from the UWB positioning base station to the train, the train position coordinates are obtained using the trilateral positioning method.

[0061] In this embodiment, the present invention provides a method for positioning a train in a long tunnel environment based on UWB, where UWB refers to a signal with an absolute bandwidth exceeding 500MHz, or a relative bandwidth greater than or equal to 0.2. UWB technology has a high transmission rate, low cost, strong anti-multipath interference capability, and high resolution, and is very suitable for high-precision positioning systems. UWB positioning methods related to position parameters are divided into the following categories: arrival angle estimation, received signal strength estimation, and arrival time estimation. Compared with arrival time and arrival angle positioning methods, RSSI-based positioning methods do not require additional hardware devices, have low power consumption, low cost, and can provide a larger ranging range. RSSI information can be used in combination with mathematical modeling methods to locate trains in tunnel environments.

[0062] The working principle of the present invention is: the signal large-scale fading model can reflect the relationship between the signal reception strength and the signal transmission distance. Therefore, the MEGNDO algorithm proposed in the present invention is used to correct the signal large-scale fading model in a tunnel environment, and then the signal reception strength and the corrected model are used to estimate the distance between the UWB positioning base station and the train in the tunnel environment, and finally the train position information is determined according to the three-sided positioning method.

[0063] The step S1 of fixing the UWB tag on the top of the train specifically involves fixing two UWB tags on the top of the center of the first carriage and the last carriage of the train respectively.

[0064] In this embodiment, according to the size and shape of the tunnel, the UWB positioning base station is arranged in a rectangular shape throughout the tunnel and fixed on the top of the tunnel wall. The UWB operating frequency is 3244-4695MHz. The UWB tag is fixed on the top of the train. When the train is running, the RSSI data between the train and the positioning base station at different distances is obtained at a communication rate of 110Kbps.

[0065] In this embodiment, the RSSI data received by the UWB device is preprocessed. First, threshold denoising is performed, and the processing principle mainly considers distance filtering and receiving power filtering; when filtering distance, the sampling data within 1m and beyond 600m from the UWB positioning base station during measurement is filtered out; according to the filtering principle of the receiving power intensity of the sampling point, RSSI data with a field strength lower than -120dBm and higher than -40dBm is filtered out. When wireless signals are transmitted in a tunnel environment, they are affected by polarization effects, multipath effects, etc. Interference noise affects the validity of the measurement data and the accuracy of the measurement results. In order to reduce the impact of the above interference on the results, the present invention uses mean filtering to improve the signal quality of RSSI data.

[0066] The step S4 is specifically as follows:

[0067] S401, construct a large-scale signal fading model for describing the UWB signal propagation process in the tunnel scenario:

[0068] L s =L1+L2lgd+L3lgdH b +L4diffraction+L5lgdlgH b +L6H m +L7Clutter

[0069] Among them, L s is the path loss; L1 is the frequency correction factor; L2 is the distance attenuation factor; d is the distance from the UWB tag to the UWB positioning base station; L3 is the base station antenna height factor; H bis the antenna height of the UWB positioning base station; L4 is the diffraction loss correction factor; diffraction is the diffraction loss; L5 is the transmitting antenna and distance correction factor; L6 is the receiving antenna correction factor; H m is the height of the UWB tag antenna; L7 is the terrain loss factor; Clutter is the average weighted loss of the ground object; lg is the logarithmic function;

[0070] S402, by improving the population initialization strategy of the GNDO generalized normal distribution optimization algorithm, and adding periodic global mutation disturbance and multi-condition selection evolution strategy, the MEGNDO optimization algorithm is obtained;

[0071] S403: According to the pre-processed RSSI data, the MEGNDO optimization algorithm is used to optimize the parameters of the signal large-scale fading model to obtain a UWB ranging model.

[0072] In this embodiment, the current research on train positioning of railway wireless communication mostly uses the WINNERⅡ model with indoor characteristics. However, the tunnel scene is quite different from the indoor environment. In order to more accurately express the UWB signal propagation process in the tunnel scene, a more flexible SPM model is selected and corrected in the tunnel scene. The SPM model expression is:

[0073] L s =L1+L2lgd+L3lgdH b +L4diffraction+L5lgdlgH b +L6H m +L7Clutter

[0074] In this embodiment, the GNDO algorithm converges slowly and is prone to fall into local optimum. The present invention proposes a MEGNDO optimization algorithm, which improves the global search capability and convergence efficiency by improving the population initialization strategy, adding periodic global variation disturbances, and multi-condition selection evolution strategies. At the same time, in view of the problem that the waveguide effect in the tunnel environment causes the far-field UWB signal fading to slow down, and the noise has a great interference on the prediction results, the idea of ​​piecewise weighting is proposed. When using MEGNDO to optimize the SPM parameters, the traditional RMSE is improved, and the far-field ranging error is reduced based on weight optimization. In addition, the present invention can also use other optimization algorithms, such as the seagull optimization algorithm, the particle swarm optimization algorithm, etc.

[0075] The population initialization strategy for improving the GNDO generalized normal distribution optimization algorithm in step S402 is specifically: combining the Sine chaotic mapping with the piecewise linear chaotic mapping to construct a composite chaotic mapping strategy, and replacing the population initialization strategy of the GNDO generalized normal distribution optimization algorithm with the composite chaotic mapping strategy, thereby completing the improvement of the population initialization strategy of the GNDO generalized normal distribution optimization algorithm.

[0076] The expression of the composite chaotic mapping strategy is:

[0077]

[0078] Where f() is the composite chaotic mapping strategy function; mod is the modulus symbol; κ i is the result of chaotic mapping of the ith individual; α and β are both chaotic parameters; sin is the sine function; π is the pi; rand is a random number between (0,1).

[0079] In this embodiment, the composite chaotic mapping population initialization strategy: in order to improve the ergodicity and randomness of the initial population, it is proposed to combine the Sine chaotic mapping suitable for low-dimensional mapping with the piecewise linear chaotic mapping with strong randomness and ergodicity advantages to construct a composite chaotic mapping strategy. The specific formula is expressed as follows:

[0080]

[0081] Among them, t is a random number between (0,1), and when α, β∈(0,1), the function is in a chaotic state. In view of the defects of poor traversal and randomness of GNDO population initialization, a composite chaotic mapping method is used to traverse the solution space of the entire population, thereby enriching the diversity of the algorithm population and greatly improving the randomness of population initialization.

[0082] The expression of the periodic global variation disturbance in step S402 is:

[0083]

[0084]

[0085]

[0086] Among them, v new is the updated tracking vector; mod is the modulus symbol; is the tracking vector of the i-th individual at the t-th iteration; is the position of the i-th individual at the t-th iteration; is an adjustment parameter with a value between 0 and 1; λ3 and λ4 are both random numbers that obey the standard normal distribution; v1 and v2 are both tracking vectors; and is the position of three randomly selected individuals at the tth iteration; Fit() is the fitness function; V is the disturbance factor; k is the global shrinkage factor; p is the reverse factor; x max is the maximum value at the tth iteration; x minis the minimum value at the tth iteration; rand is a random number between (0,1); T c is the global disturbance criterion. When it is 1, the global mutation disturbance strategy is executed, otherwise, the global mutation disturbance strategy is not executed; t is the number of iterations; T is the maximum number of iterations.

[0087] In this embodiment, in order to solve the defect that GNDO is prone to local optimization for complex problems, a periodic global variation perturbation strategy is proposed. This strategy combines the reverse idea with the perturbation idea. At the same time, in order to ensure the continuity of solving the problem, the periodic idea is used for strategy embedding. The periodic strategy uses ten iterations as a cycle to perform T c Assign a value.

[0088] In this embodiment, in order to improve the global optimization performance of GNDO and jump out of local optimization in time, a periodic global variation perturbation strategy is proposed, which combines the reverse idea and the perturbation idea. At the same time, in order to ensure the continuity of solving the problem, the periodic idea is used for strategy embedding, which effectively avoids local optimization while improving the optimization accuracy of the algorithm.

[0089] The multi-condition selection evolution strategy in step S402 specifically comprises first performing condition judgment, and then iteratively evolving the individual positions that meet the conditions; the expressions for condition judgment and iteratively evolving the individual positions are respectively:

[0090]

[0091] x new =E f ×v(index(randperm(3,1)))

[0092] E f =abs(randn) / e 2 +1 / 2

[0093] Among them, Judgment is a conditional judgment, which needs to satisfy both condition 1Con1 and condition 2Con2; Fit() is the fitness function; x new is the updated individual position; x i is the position of the ith individual; Sort() is the sorting; Npop is the population size; E f is the evolution factor; v(index(i)) is the position of the individual with selection sorting i; index is the sorting function; randperm(3,1) is to select 1 individual from 3 individuals; abs is the absolute value; randn is a random number that conforms to the normal distribution; e is a natural constant.

[0094] In this embodiment, in order to further improve the optimization accuracy and convergence speed of GNDO, the present invention proposes to use an evolutionary strategy to selectively iteratively evolve individual positions, thereby improving the superiority of the optimization cluster. Condition 1 is v new The fitness value is less than v i Fitness value, condition 2 is v new The fitness value ranking does not belong to the top 70% of the population. Considering that the solution performance is closely related to the average performance of the population, it is extremely important to avoid the emergence of poor individuals. An individual evolution strategy is proposed, and a multi-condition selection strategy is used to avoid evolving better individuals, which greatly increases the optimization accuracy and convergence speed of GNDO, making the algorithm optimization performance more superior.

[0095] The objective function of parameter optimization in step S4 is:

[0096]

[0097]

[0098] w=σ(d / 100)

[0099] σ(d / 100)=1 / (1+e -d / 100 )

[0100] Among them, J new (d) is the objective function for parameter optimization; d is the distance from the UWB tag to the UWB positioning base station; n is the number of samples; e j is the error between the predicted data and the actual measurement of the jth sample; w j is the weight of the jth sample; e is the error between the RSSI data predicted by the large-scale fading model of the signal and the actual measured RSSI data; L dB It is the actual measured RSSI data; is the RSSI data predicted by the large-scale signal fading model; Pos is the optimal parameter; w is the weight vector; σ() is the sigmoid function; e is a natural constant.

[0101] In this embodiment, the proposed MEGNDO optimization method is used to refine the parameters of SPM. It is important to note that this optimization process has a seven-dimensional nature, that is, the optimal parameters L1 to L7 of SPM are the focus of the search. Pos = [Pos1, ..., Pos N ], N = 7. Therefore, as the distance increases, the RSSI predicted by SPM is It is consistent with the actual measured RSSI The error between In this optimization problem, RMSE is defined as the objective function. The attenuation of UWB signals in the tunnel environment slows down in the far field, but SPM shows a logarithmic curve. In this area, slight noise interference may cause significant prediction errors in the distance. In order to solve the problem of large distance calculation errors in the far field, the S-shaped function is used to adjust the impact of the error on different distances. The sigmoid function is expressed as σ(x)=1 / (1+e -x ). Therefore, the weight vector for this task is expressed as w = σ(d / 100). Subsequently, a new objective function can be formulated as follows:

[0102]

[0103] The expression of the trilateral positioning method in step S6 is:

[0104]

[0105] Among them, x is the horizontal coordinate of the UWB tag; y is the vertical coordinate of the UWB tag; z is the vertical coordinate of the UWB tag; x1 is the horizontal coordinate of the first selected UWB positioning base station; x2 is the horizontal coordinate of the second selected UWB positioning base station; x3 is the horizontal coordinate of the third selected UWB positioning base station; x4 is the horizontal coordinate of the fourth selected UWB positioning base station; y1 is the vertical coordinate of the first selected UWB positioning base station; y2 is the vertical coordinate of the second selected UWB positioning base station; y3 is the vertical coordinate of the third selected UWB positioning base station; y4 is the horizontal coordinate of the fourth selected UWB positioning base station. The vertical coordinate of the selected UWB positioning base station is z1; the vertical coordinate of the first selected UWB positioning base station is z2; the vertical coordinate of the second selected UWB positioning base station is z3; the vertical coordinate of the third selected UWB positioning base station is z4; the vertical coordinate of the fourth selected UWB positioning base station is d1; the distance from the first selected UWB positioning base station to the UWB tag is d2; the distance from the second selected UWB positioning base station to the UWB tag is d3; the distance from the third selected UWB positioning base station to the UWB tag is d4; the distance from the fourth selected UWB positioning base station to the UWB tag.

[0106] In this embodiment, UWB tags are installed in both the front and rear carriages. The data collected by the UWB tags at the two locations can be used to calculate the front and rear positions respectively through the method flow of the present invention, thereby performing an integrity check on the train.

[0107] In this embodiment, the onboard control system calculates and processes the positioning data and then feeds it back to the train control system center and the train dispatching center to determine the integrity of the train and issue driving instructions to the train according to the vehicle operation status. The UWB-based train positioning method in a long tunnel environment, the deployment of UWB equipment in a tunnel environment and the schematic diagram of the positioning unit are shown in the figure. Figure 2 As shown, the test results are as follows Figure 3 shown.

Claims

1. A UWB-based train positioning method in a long tunnel environment, characterized in that: The following steps are involved: S1. Arrange the UWB positioning base station in a rectangular shape above the tunnel wall, and fix the UWB tag on the top of the train; S2, collect RSSI data between the train and the UWB positioning base station at different distances; S3, performing threshold denoising and mean filtering on the RSSI data to obtain pre-processed RSSI data; S4. Construct a signal large-scale fading model for describing the UWB signal propagation process in the tunnel scenario, and optimize the parameters of the signal large-scale fading model using the MEGNDO optimization algorithm based on the pre-processed RSSI data to obtain a UWB ranging model. S5. When the train is running, RSSI data received by the UWB device is obtained, and the distance from the UWB positioning base station to the train is calculated using the UWB ranging model according to the RSSI data received by the UWB device; S6. According to the distance from the UWB positioning base station to the train, the train position coordinates are obtained using the trilateral positioning method.

2. The UWB-based train positioning method in a long tunnel environment according to claim 1 is characterized in that: The step S1 of fixing the UWB tag on the top of the train specifically involves fixing two UWB tags on the top of the center of the first carriage and the last carriage of the train respectively.

3. The UWB-based train positioning method in a long tunnel environment according to claim 1 is characterized in that: The step S4 is specifically as follows: S401, construct a large-scale signal fading model for describing the UWB signal propagation process in the tunnel scenario: L s =L1+L2lgd+L3lgdH b +L4diffraction+L5lgdlgH b +L6H m +L7Clutter Among them, L s is the path loss; L1 is the frequency correction factor; L2 is the distance attenuation factor; d is the distance from the UWB tag to the UWB positioning base station; L3 is the base station antenna height factor; H b is the antenna height of the UWB positioning base station; L4 is the diffraction loss correction factor; diffraction is the diffraction loss; L5 is the transmitting antenna and distance correction factor; L6 is the receiving antenna correction factor; H m is the height of the UWB tag antenna; L7 is the terrain loss factor; Clutter is the average weighted loss of the ground object; lg is the logarithmic function; S402, by improving the population initialization strategy of the GNDO generalized normal distribution optimization algorithm, and adding periodic global mutation disturbance and multi-condition selection evolution strategy, the MEGNDO optimization algorithm is obtained; S403: According to the pre-processed RSSI data, the MEGNDO optimization algorithm is used to optimize the parameters of the signal large-scale fading model to obtain a UWB ranging model.

4. The UWB-based train positioning method in a long tunnel environment according to claim 3 is characterized in that: The population initialization strategy for improving the GNDO generalized normal distribution optimization algorithm in step S402 is specifically: combining the Sine chaotic mapping with the piecewise linear chaotic mapping to construct a composite chaotic mapping strategy, and replacing the population initialization strategy of the GNDO generalized normal distribution optimization algorithm with the composite chaotic mapping strategy, thereby completing the improvement of the population initialization strategy of the GNDO generalized normal distribution optimization algorithm.

5. The UWB-based train positioning method in a long tunnel environment according to claim 4 is characterized in that: The expression of the composite chaotic mapping strategy is: Where f() is the composite chaotic mapping strategy function; mod is the modulus symbol; κ i is the result of chaotic mapping of the ith individual; α and β are both chaotic parameters; sin is the sine function; π is the pi; rand is a random number between (0,1).

6. The UWB-based train positioning method in a long tunnel environment according to claim 3 is characterized in that: The expression of the periodic global variation disturbance in step S402 is: Among them, v new is the updated tracking vector; mod is the modulus symbol; is the tracking vector of the i-th individual at the t-th iteration; is the position of the i-th individual at the t-th iteration; is an adjustment parameter with a value between 0 and 1; λ3 and λ4 are both random numbers that obey the standard normal distribution; v1 and v2 are both tracking vectors; and is the position of three randomly selected individuals at the tth iteration; Fit() is the fitness function; V is the disturbance factor; k is the global shrinkage factor; p is the reverse factor; x max is the maximum value at the tth iteration; x min is the minimum value at the tth iteration; rand is a random number between (0,1); T c is the global disturbance criterion. When it is 1, the global mutation disturbance strategy is executed, otherwise, the global mutation disturbance strategy is not executed; t is the number of iterations; T is the maximum number of iterations.

7. The UWB-based train positioning method in a long tunnel environment according to claim 3 is characterized in that: The multi-condition selection evolution strategy in step S402 specifically comprises first performing condition judgment, and then iteratively evolving the individual positions that meet the conditions; the expressions for condition judgment and iteratively evolving the individual positions are respectively: Judgment: x new =E f ×v(index(randperm(3,1))) E f =abs(randn) / e 2 +1 / 2 Among them, Judgment is a conditional judgment, which needs to satisfy both condition 1Con 1 and condition 2Con 2; Fit() is the fitness function; x new is the updated individual position; x i is the position of the ith individual; Sort() is the sorting; Npop is the population size; E f is the evolution factor; v(index(i)) is the position of the individual with selection sorting i; index is the sorting function; randperm(3,1) is to select 1 individual from 3 individuals; abs is the absolute value; randn is a random number that conforms to the normal distribution; e is a natural constant.

8. The UWB-based train positioning method in a long tunnel environment according to claim 1, characterized in that: The objective function of parameter optimization in step S4 is: w=σ(d / 100) σ(d / 100)=1 / (1+e -d / 100 ) Among them, J new (d) is the objective function for parameter optimization; d is the distance from the UWB tag to the UWB positioning base station; n is the number of samples; e j is the error between the predicted data and the actual measurement of the jth sample; w j is the weight of the jth sample; e is the error between the RSSI data predicted by the large-scale fading model of the signal and the actual measured RSSI data; L dB It is the actual measured RSSI data; is the RSSI data predicted by the large-scale signal fading model; Pos is the optimal parameter; w is the weight vector; σ() is the sigmoid function; e is a natural constant.

9. The UWB-based train positioning method in a long tunnel environment according to claim 1, characterized in that: The expression of the trilateral positioning method in step S6 is: Among them, x is the horizontal coordinate of the UWB tag; y is the vertical coordinate of the UWB tag; z is the vertical coordinate of the UWB tag; x1 is the horizontal coordinate of the first selected UWB positioning base station; x2 is the horizontal coordinate of the second selected UWB positioning base station; x3 is the horizontal coordinate of the third selected UWB positioning base station; x4 is the horizontal coordinate of the fourth selected UWB positioning base station; y1 is the vertical coordinate of the first selected UWB positioning base station; y2 is the vertical coordinate of the second selected UWB positioning base station; y3 is the vertical coordinate of the third selected UWB positioning base station; y4 is the horizontal coordinate of the fourth selected UWB positioning base station. The vertical coordinate of the selected UWB positioning base station is z1; the vertical coordinate of the first selected UWB positioning base station is z2; the vertical coordinate of the second selected UWB positioning base station is z3; the vertical coordinate of the third selected UWB positioning base station is z4; the vertical coordinate of the fourth selected UWB positioning base station is d1; the distance from the first selected UWB positioning base station to the UWB tag is d2; the distance from the second selected UWB positioning base station to the UWB tag is d3; the distance from the third selected UWB positioning base station to the UWB tag is d4; the distance from the fourth selected UWB positioning base station to the UWB tag.

Images