UWB / pdr fusion positioning method and system

Abstract

This invention discloses a UWB / PDR fusion positioning method and system, comprising: obtaining the distance difference between an ultra-wideband base station and an ultra-wideband tag carried by a pedestrian using a time difference of arrival (TDOA) positioning algorithm; calculating the first pedestrian position using the Chan's positioning algorithm based on the distance difference; acquiring pedestrian inertial correlation data to calculate the pedestrian's nonlinear step length and calculating the second pedestrian position using pedestrian dead reckoning; fusing the first and second pedestrian positions using an improved particle filter to obtain the final positioning result; constructing a particle transfer model based on the pedestrian's nonlinear step length to update the particle position at each moment, and dynamically updating the weight of each particle at the end of each step based on the Euclidean distance between each particle and the first pedestrian position at the end of each step; the particle position is generated based on the first pedestrian position at the initial moment. This invention has significant advantages such as stable fusion effect and high reliability.

Description

Technical Field

[0001] This invention belongs to the field of communication technology, and relates to indoor positioning technology, specifically to a UWB / PDR fusion positioning method and system. Background Technology

[0002] Ultra-wideband (UWB) positioning technology is renowned for its high accuracy and high data transmission rate, and is widely used in indoor positioning, asset tracking, and smart homes. UWB's high positioning accuracy makes it highly effective in applications requiring precise location data. Furthermore, UWB technology has strong anti-interference capabilities, enabling it to operate stably in complex indoor environments. However, UWB positioning technology also has some drawbacks. First, UWB signals are susceptible to interference from obstacles and environmental factors during propagation, potentially leading to signal attenuation and limited coverage. Second, due to the wide UWB spectrum, the associated hardware implementation costs are high, and its design and maintenance are relatively complex.

[0003] Pedestrian Dead Reckoning (PDR) algorithms utilize data from sensors such as accelerometers, gyroscopes, and magnetometers to track a user's gait and calculate their position. They can perform positioning indoors without relying on signal coverage and offer high accuracy. However, PDR algorithms are susceptible to the cumulative effects of sensor errors, leading to a decrease in accuracy over extended periods of use.

[0004] CN117739981A discloses a map-assisted UWB and PDR fusion positioning method, comprising: generating Gaussian-distributed particles within the radius of the initial position of the mobile module in the map-assisted UWB and PDR fusion positioning system, and using these particles as their initial positions; updating the state and position of the moving particles using a state equation; calculating and updating the particle weights based on the distance measurement between a single UWB base station and a UWB node; correcting the particle weights through map information matching; resampling and normalizing the particles; and statistically analyzing the information of all particles to obtain the positioning result. However, this patent relies solely on the distance measurement data of a single UWB base station and cannot directly achieve two-dimensional or three-dimensional positioning, requiring a high degree of dependence on PDR heading and step size estimation. Since the PDR algorithm is easily affected by the accumulation of sensor errors, the fusion effect significantly decreases over long-term use.

[0005] CN117870669A discloses a cable tunnel positioning method based on PDR, UWB, and map matching. Specifically, it includes: collecting motion information of inspection personnel using accelerometers, gyroscopes, and magnetometers; using a PDR algorithm to detect personnel gait and estimate step length and heading angle; performing UWB positioning based on the TDOA algorithm under conditions of minimal interference from non-line-of-sight factors; fusing the PDR and UWB positioning results using particle filtering; and correcting for partial and complete particle penetration through walls to obtain the personnel positioning result. However, this patent lacks dynamic evaluation of particle degradation, which may lead to premature or delayed resampling, resulting in low reliability of personnel positioning. Summary of the Invention

[0006] To address the shortcomings of existing technologies, such as a significant decrease in fusion effect and low reliability after long-term use, this invention provides a UWB / PDR fusion positioning method. By improving particle filtering, the coordinates calculated by PDR are fused with the UWB positioning coordinates, thereby improving positioning accuracy.

[0007] The present invention adopts the following technical solution.

[0008] This invention discloses a UWB / PDR fusion positioning method, comprising: The distance difference between the ultra-wideband base station and the ultra-wideband tag carried by the pedestrian is obtained by the time difference of arrival positioning algorithm; Based on the distance difference, the location of the first pedestrian is calculated using the Qian's positioning algorithm; Obtain pedestrian inertial correlation data, calculate the nonlinear step length of the pedestrian based on the pedestrian inertial correlation data, and use the pedestrian dead reckoning algorithm to obtain the second pedestrian position; The first pedestrian position and the second pedestrian position are fused using an improved particle filter to obtain the final positioning result. The improved particle filter constructs a particle transfer model based on the nonlinear step size of the pedestrian to update the particle position at the end of each step, and dynamically updates the weight of each particle at the end of each step based on the Euclidean distance between each particle and the first pedestrian position at the end of each step. The particle position is generated based on the first pedestrian position at the initial time.

[0009] More preferably, The pedestrian inertial correlation data includes heading angle, lateral acceleration, longitudinal acceleration, and vertical acceleration.

[0010] More preferably, The calculation of the pedestrian's nonlinear step length based on the pedestrian inertial correlation data is based on the maximum and minimum acceleration of each step of the pedestrian, as well as a preset nonlinear step length coefficient, to calculate the nonlinear step length of that step.

[0011] More preferably, The particle transfer model for updating the particle position at each moment is constructed based on the nonlinear step size of the pedestrian at each step, the heading angle of the step, the final positioning x-coordinate at the end of the previous step, and the final positioning y-coordinate at the end of the previous step. The first pedestrian position at the initial moment is taken as the final position at the initial moment.

[0012] More preferably, The dynamic update of the weight of each particle at the end of each step is also determined based on the environmental confidence factor at the end of that step. The environmental confidence factor corresponding to the end time of each step of the pedestrian is calculated by using the final location, the first pedestrian position, the second pedestrian position information, the heading angle, and the preset first and second adjustment coefficients at the end time of that step.

[0013] More preferably, The improved particle filter further determines the effective number of particles for each step based on the weight of each particle at the end of each step, and performs adaptive resampling based on the effective number of particles.

[0014] More preferably, The adaptive resampling process involves sorting all particles at that moment according to their corresponding particle weights from largest to smallest when the number of effective particles is greater than a preset effective particle threshold. Based on the number of effective particles, the particles at the top of the sorted list are selected to form a new sample set, which is used to determine the final location at that moment.

[0015] Another aspect of the present invention discloses a UWB / PDR fusion positioning system based on a UWB / PDR fusion positioning method, including a distance difference determination module, a first pedestrian position determination module, a second pedestrian position determination module, and a positioning fusion module: The distance difference determination module obtains the distance difference between the ultra-wideband base station and the ultra-wideband tag carried by the pedestrian through the time difference of arrival positioning algorithm; The first pedestrian location determination module calculates the first pedestrian's location using the Qian's positioning algorithm based on the distance difference; The second pedestrian position determination module is used to acquire pedestrian inertial correlation data, calculate the nonlinear step length of the pedestrian based on the pedestrian inertial correlation data, and use pedestrian dead reckoning to obtain the second pedestrian position; The positioning fusion module uses an improved particle filter to fuse the positions of the first and second pedestrians to obtain the final positioning result. The improved particle filter constructs a particle transfer model based on the nonlinear step size of the pedestrian to update the particle position at the end of each step, and dynamically updates the weight of each particle at the end of each step based on the Euclidean distance between each particle and the first pedestrian position at the end of each step. The particle position is generated based on the first pedestrian position at the initial time.

[0016] Another aspect of this application discloses an electronic device, including a processor and a storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the aforementioned UWB / PDR fusion positioning method.

[0017] This application also discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the UWB / PDR fusion positioning method.

[0018] The beneficial effects of this invention are compared with those of the prior art: This invention proposes a UWB / PDR fusion localization method. By improving the particle filter, it retains particles with high weights, prevents particle degradation, reduces the number of particles, and improves computational efficiency. By fusing the UWB localization results and the PDR algorithm's localization results using the improved particle filter, localization accuracy is improved.

[0019] The particle transfer model update mechanism based on the nonlinear step size of pedestrians proposed in this invention enables a deeper integration of UWB positioning results and PDR algorithm, further improving positioning accuracy.

[0020] The particle weight update in this invention introduces environmental feature recognition and dynamically adjusts the weight of each particle to solve the problem of sudden changes in UWB positioning error in NLOS (non-line-of-sight) environment. It realizes unsupervised environment adaptive fusion, which not only further improves positioning accuracy, but also overcomes the problem of insufficient robustness of fixed weight allocation strategy in complex indoor environment in the prior art.

[0021] The present invention also has the outstanding advantages of stable fusion effect and high reliability during long-term use. Attached Figure Description

[0022] Figure 1 This is a schematic flowchart illustrating the overall implementation of the present invention; Figure 2 This is a schematic diagram of the TDOA positioning algorithm. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, other embodiments obtained by those skilled in the art without creative effort are all within the protection scope of this invention.

[0024] This application discloses a UWB / PDR fusion positioning method, see appendix. Figure 1 ,include: The distance difference between the ultra-wideband base station and the ultra-wideband tag carried by the pedestrian is obtained by the time difference of arrival positioning algorithm; Based on the distance difference, the location of the first pedestrian is calculated using the Qian's positioning algorithm; Obtain pedestrian inertial correlation data, calculate the nonlinear step length of the pedestrian based on the pedestrian inertial correlation data, and use the pedestrian dead reckoning algorithm to obtain the second pedestrian position; The pedestrian inertial correlation data includes heading angle, lateral acceleration, longitudinal acceleration, and vertical acceleration.

[0025] The calculation of the pedestrian's nonlinear step length based on the pedestrian inertial correlation data is based on the maximum and minimum acceleration of each step, as well as a preset nonlinear step length coefficient, to calculate the nonlinear step length of that step, as shown in the following formula: ; in, This represents the non-linear step size of a pedestrian's step. These are nonlinear step size coefficients; This represents the maximum acceleration during the walking process; This represents the minimum acceleration during the walking process.

[0026] The first pedestrian position and the second pedestrian position are fused using an improved particle filter to obtain the final positioning result. The improved particle filter constructs a particle transfer model based on the nonlinear step size of the pedestrian to update the particle position at the end of each step, and dynamically updates the weight of each particle at the end of each step based on the Euclidean distance between each particle and the first pedestrian position at the end of each step. The particle position is generated based on the first pedestrian position at the initial time.

[0027] The particle transfer model for updating the particle position at each moment, based on the pedestrian's nonlinear step size, is constructed by taking into account the nonlinear step size of each step, the heading angle of that step, the final x-coordinate of the previous step, and the final y-coordinate of the previous step, as shown below: ; in, This is the particle transfer model corresponding to step u; Let u be the nonlinear step size of the pedestrian at step u. Let be the heading angle of the pedestrian at step u; The final x-coordinate of the pedestrian at the end of step u-1; The final ordinate of the pedestrian at the end of step u-1 is determined; the position of the first pedestrian at the initial moment is taken as the final position at the initial moment.

[0028] Based on the particle transfer model corresponding to each step and the position of each particle at the end of the previous step, the position of the particle at the end of the current step is calculated; each particle is generated at the initial time based on the position of the first pedestrian at the initial time, as shown in the following formula: ; in, For the first The particle was in the pedestrian's... The position at the end of the step; For the first The particle was in the pedestrian's... -1 is the position at the end of the step.

[0029] The dynamic update of the weight of each particle at the end of each step is also determined based on the environmental confidence factor at the end of that step. Specifically, the weight of each particle at the end of each step is dynamically updated based on the Euclidean distance between each particle and the position of the first pedestrian at the end of each step, and the environmental confidence factor at the end of that step, as shown in the following formula: ; in, For pedestrians The first step at the end time Individual particle weights; Indicates the first One particle; Indicates the pedestrian's position The Euclidean distance between the r-th particle and the position of the first pedestrian at the end of the step; Indicates pedestrian number The environmental confidence factor corresponding to the end time of the step.

[0030] The environmental confidence factor corresponding to the end time of each step of the pedestrian is calculated based on the final location within a preset first time period, the location information of the first pedestrian, the location information of the second pedestrian, the heading angle, and preset first and second adjustment coefficients at the end time of that step, as shown in the following formula: ; Wherein, sigmod(·) represents an S-shaped function that transforms the numbers within the parentheses into probabilities between 0 and 1; Indicates pedestrian number The variance of the first pedestrian's position within a preset first time period before the end of the step (the first time period can be set by those skilled in the art according to the actual situation, and will not be elaborated here. Assuming that the first time period includes 3 sampling times, the variance of the first pedestrian's position within the preset first time period before the end of the step). At the end of the step, it is based on the first step. -1st step end time, the first -2-step end time, the first -3. At the end of each time step, calculate the Euclidean distance between the first pedestrian's position at each time step and the final position at the corresponding time step. Use this variance as the variance of the first pedestrian's position within the first time period. Indicates pedestrian number The variance of the second pedestrian's position within the first time period preset before the end of the step is calculated in a similar way to the variance of the first pedestrian's position, and will not be repeated here. Indicates pedestrian number Within a preset first time period before the end of the step, the angular velocity variance of the heading angle; α represents the first adjustment coefficient, whose preferred value range is [0.3, 0.5]; β represents the second adjustment coefficient, whose preferred value range is [0.1, 0.3]; those skilled in the art can set the duration of the first time period according to the actual situation, preferably, the first time period includes 3 to 5 sampling times. Within the first time period at the start of the fusion positioning, the following is set =1.

[0031] The improved particle filtering further determines the effective number of particles for each step based on the weight of each particle at the end of each step, and then uses this effective number of particles as a basis for further analysis. Perform adaptive resampling.

[0032] The adaptive resampling process involves sorting all particles at a given moment according to their corresponding particle weights from largest to smallest when the number of effective particles exceeds a preset effective particle threshold, and selecting the particles ranked highest in the sorted list. The particles form a new sample set to determine the final location at that moment.

[0033] This application also discloses a UWB / PDR fusion positioning system based on the UWB / PDR fusion positioning method, including a distance difference determination module, a first pedestrian position determination module, a second pedestrian position determination module, and a positioning fusion module.

[0034] The distance difference determination module obtains the distance difference between the ultra-wideband base station and the ultra-wideband tag carried by the pedestrian through the time difference of arrival positioning algorithm; The first pedestrian location determination module calculates the first pedestrian's location using the Qian's positioning algorithm based on the distance difference; The second pedestrian position determination module is used to acquire pedestrian inertial correlation data, calculate the nonlinear step length of the pedestrian based on the pedestrian inertial correlation data, and use pedestrian dead reckoning to obtain the second pedestrian position; The positioning fusion module uses an improved particle filter to fuse the positions of the first and second pedestrians to obtain the final positioning result. The improved particle filter constructs a particle transfer model based on the nonlinear step size of the pedestrian to update the particle position at the end of each step, and dynamically updates the weight of each particle at the end of each step based on the Euclidean distance between each particle and the first pedestrian position at the end of each step. The particle position is generated based on the first pedestrian position at the initial time.

[0035] Example 1 A UWB / PDR fusion positioning method.

[0036] S1: See Appendix Figure 2 By obtaining the measured TDOA (Time Difference of Arrival) value (i.e., the time difference between the signal transmitted by the same UWB tag arriving at two different base stations), the distance difference between the UWB tag (Ultra-Wideband tag) and multiple UWB base stations (Ultra-Wideband base stations) is obtained; based on the distance difference obtained from the TDOA measurement values ​​of the UWB tag carried by the pedestrian to different base stations at the same time, a set of hyperbolic equations about the location of the pedestrian's UWB positioning tag is constructed; a TDOA positioning model is established based on this set of hyperbolic equations.

[0037] S2: Based on the TDOA measurement value, the Chan positioning algorithm is used to solve the aforementioned hyperbolic equations to obtain the specific location of the UWB tag, which is denoted as the first pedestrian position.

[0038] The exact location of the label is calculated using the following method: Establish coordinate axes for calculating the specific location of the label; Based on TDOA data, we can obtain ,

[0039] in For base stations The actual distance to the label. For base stations The actual distance to the label. For the first The coordinates of each base station For the first The x-coordinate of each base station coordinate. For the first The ordinate of each base station coordinate; The coordinates of the unknown UWB label are: x is the x-coordinate of the label, and y is the y-coordinate of the label. For base stations The actual distance to the tag and the base station The distance difference between the actual distance to the label and the actual distance to the label.

[0040] set up For location information where the label is unknown. Let R be the specific location of the tag in the coordinate system, where R is the distance from the tag to the reference base station (usually base station number 1), therefore R = R1. Indicates the specific position of the label's x-axis; Indicates the specific position of the label's vertical axis; Indicates the specific distance from the tag to the reference base station; based on the aforementioned settings Establish error equations ,in The first error vector is represented by h; the first distance difference observation matrix is ​​represented by h. Represents the first positioning coefficient matrix; h and The definitions are as follows: ,

[0041] in, ; This represents the square of the modulus of the coordinates of the i-th base station; The x-coordinate represents the coordinates of the i-th base station; Let be the ordinate of the i-th base station. For the first The difference in the x-coordinate between the coordinates of each base station and the reference base station; For the first The difference in the ordinate between the coordinates of each base station and the reference base station; The x-coordinate of the reference base station coordinates; The ordinate is the reference base station coordinate. This is the difference between the actual distance from base station 2 to the tag and the actual distance from the reference base station to the tag; This is the difference between the actual distance from base station 3 to the tag and the actual distance from the reference base station to the tag; For base stations The difference between the actual distance to the tag and the actual distance from the reference base station to the tag; The square of the modulus of the coordinates of the first base station; The square of the modulus of the coordinates of the second base station; The square of the modulus of the coordinates of the third base station; This is the difference in the x-coordinate between the second base station and the reference base station. This is the difference in the x-coordinate between the third base station and the reference base station. This is the difference in the ordinate between the coordinates of the second base station and the reference base station; Let be the difference between the ordinates of the third base station and the reference base station. Assume the variance of the measurement error between base station i and reference base station 1 in the TDOA positioning algorithm is... Then the covariance matrix of the measurement error is Where Q represents the first covariance matrix; diag{·} represents the diagonal matrix operator; This represents the variance of the measurement error between base station 2 and reference base station 1; This represents the variance of the measurement error between base station 3 and reference base station 1; This represents the variance of the measurement error between base station p and reference base station 1. It is expressed using the first covariance matrix. Replace the first error vector The solution obtained by weighted least squares method is as follows. The position is shown in the following formula: in, It takes the minimum value within the parentheses; the first error vector is shown in the following formula: ; in, According to Determine the actual distance from the second base station to the pedestrian; According to Determine the actual distance from the third base station to the pedestrian; According to Determine the actual distance from the Pth base station to the pedestrian; This represents the propagation speed of electromagnetic waves. The solution obtained in the above process As .

[0042] The above process uses replace This increases the error. To reduce the error, weighted least squares method is used for further optimization. The position is shown in the following formula. The positioning error formula is as follows: ; in, This represents the increment of the first distance difference observation matrix. = , In a specific location The first distance difference observation matrix is ​​calculated below; This represents the increment of the first positioning coefficient matrix. = , for the initial value The first positioning coefficient matrix is ​​calculated below; Depend on Substituting into the above formula, we get: ; in, This represents an increment indicating the location information of an unknown label; United and This gives the increment of the label's specific position in the coordinate system. and its covariance matrix for:

[0043] ; Where E[·] represents the expectation within the parentheses; Will Divided into x-axis components , ordinate components Base station distance components There are three components in total: , , ,in These are the errors for the horizontal axis component, the vertical axis component, and the base station distance component, respectively. and Subtract respectively and ,right Squaring the components

[0044] The above formula is abbreviated as: ; in, Indicates the second error vector; This represents the second distance difference observation matrix. = ; This represents the second positioning coefficient matrix. = ; This represents the position residual vector of the tag relative to the reference base station. = .

[0045] The covariance matrix is ,

[0046] in, It is based on Calculated The x value in; It is based on Calculated The y-value in; It is based on The calculated actual distance from the reference base station to the tag.

[0047] right Perform least squares estimation

[0048] The final position of the label is the position of the first pedestrian, denoted as . As shown in the following formula: .

[0049] S3: Obtain pedestrian inertial correlation data using inertial sensors carried by the pedestrian; calculate the pedestrian's nonlinear step length based on the pedestrian inertial correlation data, and use the pedestrian dead reckoning to obtain the second pedestrian position based on the pedestrian inertial correlation data; the pedestrian inertial correlation data includes pedestrian heading angle data, x-axis acceleration (horizontal axis acceleration), y-axis acceleration (vertical axis acceleration), and z-axis acceleration (vertical axis acceleration) data. The pedestrian heading angle refers to the clockwise angle between the projection of the pedestrian's forward direction onto the local horizontal plane and geographic north (or geomagnetic north).

[0050] Those skilled in the art should know that inertial sensors include gyroscopes, magnetic sensors, and accelerometers. Gyroscopes and magnetic sensors can obtain heading angle data; accelerometers can obtain x-axis (horizontal axis), y-axis (vertical axis), and z-axis (vertical axis) acceleration data.

[0051] The nonlinear step size of each step taken by a pedestrian is calculated using the following formula: ; Where represents the nonlinear step size of a pedestrian's step; The nonlinear step size coefficient was obtained by linear fitting using the least squares method based on multiple sets of experimental results. and These represent the maximum and minimum accelerations during the walking process, i.e., the accelerations of the smart device, including the inertial sensor, during the walking process. The maximum and minimum accelerations on the axis are given. The step size coefficient is calculated using the least squares method, and the calculation process is as follows: 1) Detect the maximum acceleration of each complete step during a pedestrian's walking process. and the minimum acceleration for each complete step And use a measuring tape to measure the actual distance of each complete step. .

[0052] in, This indicates the maximum acceleration in step 1; This indicates the maximum acceleration in step 2; This represents the maximum acceleration at step n. This represents the minimum acceleration in step 1; This represents the minimum acceleration in step 2; This represents the minimum acceleration at step n; n represents the total number of steps taken by the pedestrian during this detection process.

[0053] 2) Substituting the maximum acceleration, minimum acceleration, and step size for each step into the step size calculation formula, we have:

[0054] in, For the first The nonlinear step size of the step. For the first Maximum acceleration of the step, For the first The minimum acceleration of a step.

[0055] 3) Using the least squares method to Group data When fitted to a straight line, the coefficients The solution formula is: .

[0056] S4: The final positioning result is obtained by fusing the first pedestrian position obtained by UWB positioning and the second pedestrian position obtained by PDR positioning using improved particle filtering.

[0057] The initial position of the first pedestrian is taken as the final location at the initial moment, along with the system state. ;in, The x-coordinate represents the pedestrian's position at the initial moment; The ordinate represents the pedestrian's initial position; the UWB positioning value at the end of each step is used as the observation value at the end of that step, and u represents the th step the pedestrian has taken since the initial time. Step, the first Observations at the end of the step ;in, Indicates the first The x-coordinate of the observation at the end of the step; Indicates the first The ordinate of the observation at the end of the step; A particle transfer model update mechanism based on the nonlinear step size of pedestrians is introduced. This mechanism constructs a particle transfer model based on the nonlinear step size of each step, the heading angle of that step, the final x-coordinate of the previous step, and the final y-coordinate of the previous step, as follows: ; in, This is the particle transfer model corresponding to step u; Let u be the nonlinear step size of the pedestrian at step u. Let be the heading angle of the pedestrian at step u; The final x-coordinate of the pedestrian at the end of step u-1; The final vertical coordinate of the pedestrian at the end of step u-1 is determined.

[0058] Those skilled in the art should know that the system state in particle filtering refers to a probability density function approximated by a set of random samples that propagate in the state space and are used to estimate the system state. These samples are particles; each particle represents a possible system state, and the system state can be estimated by randomly generating, updating, and weighting these particles.

[0059] An improved particle filter is used to fuse the first pedestrian position obtained from UWB localization and the second pedestrian position obtained from PDR localization, and the output of the improved particle filter is used as the final localization result. The particle filter algorithm process is as follows: Initialization, as follows: The initial UWB positioning position is the initial position of the first pedestrian. As a truth value, where This represents the x-coordinate of the first pedestrian's position at the initial moment; This represents the ordinate of the first pedestrian's position at the initial moment, generated using a Gaussian white noise distribution model to create the initial particle set. and corresponding weights The particles represent candidate position hypotheses generated based on the UWB positioning location at the initial moment. Each particle represents a possible positional state. Those skilled in the art should know how to generate the initial particle set according to a Gaussian white noise distribution model, which will not be elaborated here.

[0060] The initial weights of the particles are ,satisfy .

[0061] in, For the first The position of each particle at the initial moment, wherein the position includes a horizontal coordinate and a vertical coordinate. For the first The initial weights of each particle. The total number of particles is preferably 100 to 500.

[0062] Updated, the formula is as follows: At the pedestrian Step end time according to formula Perform location update to obtain pedestrian number The position at the end of the step is given by the following formula: ; in, For the first The particle was in the pedestrian's... The position at the end of the step; For the first The particle was in the pedestrian's... -1 is the position at the end of the step.

[0063] Calculate the updated number The Euclidean distance between each particle and the position of the first pedestrian is given by the following formula:

[0064] in, Indicates the pedestrian's position Step End Time The Euclidean distance between each particle and the position of the first pedestrian; For the first The particle was in the pedestrian's... The x-axis coordinate at the end of the step. For the first The particle was in the pedestrian's... The ordinate of the step's end time; For pedestrians The x-coordinate of the first pedestrian's position observed at the end of the step; For pedestrians The vertical coordinate of the first pedestrian's position observed at the end of the step. The unit of measurement for coordinates in this invention is meters.

[0065] Particle weight update by The decision has been made, and the update method is as follows: ; in, For pedestrians The first step at the end time Individual particle weights; Indicates the first One particle; Indicates the pedestrian's position The Euclidean distance between the r-th particle and the position of the first pedestrian at the end of the step; Indicates pedestrian number The environmental confidence factor corresponding to the end time of the step is used by this invention to introduce environmental feature recognition, as shown in the following formula: ; Where sigmod(·) represents an S-shaped function that transforms the numbers within the parentheses into probabilities between 0 and 1; Indicates pedestrian number The variance of the first pedestrian's position within the preset first time period before the end of the step (assuming the first time period includes 3 sampling times, the current time is the first time). At the end of the step, it is based on the first step. -1st step end time, the first -2-step end time, the first -At the end of step 3, the Euclidean distance between the first pedestrian position at each moment and the final position obtained through this invention at the corresponding moment is used to calculate the variance of the first pedestrian position within the preset first time period. The calculation methods for the variance of the second pedestrian position and the variance of the angular velocity of the heading angle are similar to those for the first pedestrian position (and will not be repeated here). Indicates pedestrian number The variance of the second pedestrian's position during the first time period preset before the end of the step; Indicates pedestrian number Within a preset first time period before the end of the step, the angular velocity variance of the heading angle; α represents the first adjustment coefficient, whose preferred value range is [0.3, 0.5]; β represents the second adjustment coefficient, whose preferred value range is [0.1, 0.3]; those skilled in the art can set the duration of the first time period according to the actual situation, preferably, the first time period includes 3 to 5 sampling times. Within the first time period at the start of the fusion positioning, the following is set =1.

[0066] When the UWB signal is stable ( Hour, The value tends towards 2, enhancing the weight of UWB observations and improving positioning accuracy; when the UWB signal is interfered with ( When the PDR is large and stable, The value tends to be 0.5, reducing the UWB weight and relying on PDR inertial navigation.

[0067] This invention introduces environmental feature recognition to update particle weights and dynamically adjusts the weights of each particle, solving the problem of abrupt changes in UWB positioning errors in NLOS (Non-Line-of-Sight) environments. This achieves unsupervised adaptive fusion of environmental features and overcomes the insufficient robustness of fixed weight allocation strategies in complex indoor environments in existing technologies. Perform adaptive resampling, i.e., adaptively adjust the number of particles: ; in, The effective number of particles, To round up. Set the effective particle threshold to [value]. ,like Then, sort the particles by weight from largest to smallest, and select the top... A new sample set is formed by selecting particles with large weights, and then the state estimate is output; otherwise, a resampling step is performed. Threshold Set as ; Specifically, by adaptively adjusting the number of particles, the computational efficiency is improved by retaining the effective particles with high weights to prevent particle degradation and reducing the number of particles.

[0068] Output state estimate: Calculate the expected value of the resampled particle set: ; in, Indicates the pedestrian's position The final positioning at the end of the step; when the previous step is selected. When a new sample set is formed by particles with large weights, N in the formula should be replaced with... .

[0069] The expected value of the particle set is the optimal position for fusion localization.

[0070] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0071] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0072] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0073] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A UWB / PDR fusion positioning method, characterized in that, include: The distance difference between the ultra-wideband base station and the ultra-wideband tag carried by the pedestrian is obtained by the time difference of arrival positioning algorithm; Based on the distance difference, the location of the first pedestrian is calculated using the Qian's positioning algorithm; Obtain pedestrian inertial correlation data, calculate the nonlinear step length of the pedestrian based on the pedestrian inertial correlation data, and use the pedestrian dead reckoning algorithm to obtain the second pedestrian position; The first pedestrian position and the second pedestrian position are fused using an improved particle filter to obtain the final positioning result. The improved particle filter constructs a particle transfer model based on the nonlinear step size of the pedestrian to update the particle position at the end of each step, and dynamically updates the weight of each particle at the end of each step based on the Euclidean distance between each particle and the first pedestrian position at the end of each step. The particle position is generated based on the first pedestrian position at the initial time.

2. The UWB / PDR fusion positioning method according to claim 1, characterized in that: The pedestrian inertial correlation data includes heading angle, lateral acceleration, longitudinal acceleration, and vertical acceleration.

3. The UWB / PDR fusion positioning method according to claim 1 or 2, characterized in that: The calculation of the pedestrian's nonlinear step length based on the pedestrian inertial correlation data is based on the maximum and minimum acceleration of each step of the pedestrian, as well as a preset nonlinear step length coefficient, to calculate the nonlinear step length of that step.

4. The UWB / PDR fusion positioning method according to claim 1 or 2, characterized in that: The particle transfer model for updating the particle position at each moment is constructed based on the nonlinear step size of the pedestrian at each step, the heading angle of the step, the final positioning x-coordinate at the end of the previous step, and the final positioning y-coordinate at the end of the previous step. The first pedestrian position at the initial moment is taken as the final position at the initial moment.

5. The UWB / PDR fusion positioning method according to claim 1, characterized in that: The dynamic update of the weight of each particle at the end of each step is also determined based on the environmental confidence factor at the end of that step. The environmental confidence factor corresponding to the end time of each step of the pedestrian is calculated by using the final location, the first pedestrian position, the second pedestrian position information, the heading angle, and the preset first and second adjustment coefficients at the end time of that step.

6. The UWB / PDR fusion positioning method according to claim 1, characterized in that: The improved particle filter further determines the effective number of particles for each step based on the weight of each particle at the end of each step, and performs adaptive resampling based on the effective number of particles.

7. The UWB / PDR fusion positioning method according to claim 6, characterized in that: The adaptive resampling process involves sorting all particles at that moment according to their corresponding particle weights from largest to smallest when the number of effective particles is greater than a preset effective particle threshold. Based on the number of effective particles, the particles at the top of the sorted list are selected to form a new sample set, which is used to determine the final location at that moment.

8. A UWB / PDR fusion positioning system utilizing the UWB / PDR fusion positioning method according to any one of claims 1-7, characterized in that, It includes a distance difference determination module, a first pedestrian location determination module, a second pedestrian location determination module, and a location fusion module: The distance difference determination module obtains the distance difference between the ultra-wideband base station and the ultra-wideband tag carried by the pedestrian through the time difference of arrival positioning algorithm; The first pedestrian location determination module calculates the first pedestrian's location using the Qian's positioning algorithm based on the distance difference; The second pedestrian position determination module is used to acquire pedestrian inertial correlation data, calculate the nonlinear step length of the pedestrian based on the pedestrian inertial correlation data, and use pedestrian dead reckoning to obtain the second pedestrian position; The positioning fusion module uses an improved particle filter to fuse the positions of the first and second pedestrians to obtain the final positioning result. The improved particle filter constructs a particle transfer model based on the nonlinear step size of the pedestrian to update the particle position at the end of each step, and dynamically updates the weight of each particle at the end of each step based on the Euclidean distance between each particle and the first pedestrian position at the end of each step. The particle position is generated based on the first pedestrian position at the initial time.

9. A terminal, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the UWB / PDR fusion positioning method according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the steps of the UWB / PDR fusion positioning method according to any one of claims 1-7.

Images