Meerkat optimization wireless positioning method based on improved Kalman filter

Abstract

The method comprises the following steps: S1, constructing a system TOA (Time of Arrival) ranging model in a non-line-of-sight environment, and initializing parameters of a Kalman filter; S2, predicting and observing a target node based on the system TOA ranging model, and optimizing the parameters of the Kalman filter according to the prediction and observation results; S3, correcting ranging information output by the system TOA ranging model by using the optimized Kalman filter, and solving a ranging information combination and a coarse position of the target node at a minimum positioning accuracy factor by using a positioning algorithm based on a linear position line; and S4, solving an accurate position of the target node by using a meerkat optimization algorithm. The method can improve positioning accuracy and is easy to implement.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of wireless positioning technology, in particular to a meerkat optimization wireless positioning method based on an improved Kalman filter. BACKGROUND

[0002] In recent years, with the rapid development of mobile communication technology and the popularity of mobile terminals, location services have gradually emerged and are widely used in human production activities. In the outdoor open environment, global positioning systems such as GPS and Beidou can effectively provide positioning services, but in some closed scenes or scenes with high positioning accuracy requirements, global positioning systems cannot provide positioning services that meet the needs. Therefore, positioning technology based on wireless communication has gradually developed to solve this problem. However, in real environments, there are usually complex structures of various obstacles, resulting in diffraction, reflection, and other situations of communication signals. Such non-line-of-sight propagation phenomena can greatly reduce the positioning accuracy of wireless communication.

[0003] The Chan algorithm proposed by Y.T.Chan and the linear location of opportunity (LLOP) algorithm proposed by Caffery are basic positioning algorithms for TDOA (Time Difference of Arrival) and TOA (Time of Arrival), respectively. When there is non-line-of-sight propagation, it is difficult to estimate accurate results with this algorithm, and in more serious cases, it can lead to positioning failure.

[0004] In order to eliminate the influence of non-line-of-sight propagation in wireless communication, AI-Jazzar in the document "ML and Bayesian TOA location estimators for NLOS environments" uses different scattering models to obtain the probability density function of different scattering models based on TOA signals, and then uses maximum likelihood estimation (ML) and Bayesian estimation to estimate the actual position of the target node. However, only when the actual channel is consistent with the assumed model, the positioning accuracy of the algorithm can be guaranteed. In the document "TOA positioning algorithm based on Kalman filter", researchers such as Li Feng use ordinary Kalman filter algorithm to filter out noise in the system, and the effect of suppressing the influence of non-line-of-sight error is limited. In the document "Consistent orthogonal cubature Kalman tracking algorithm based on TOA / TDOA", technical personnel such as Yan Leibing mix TDOA and TOA two kinds of measurement signals, and continuously update the system state quantity through orthogonal cubature Kalman filter to obtain positioning information. However, this method needs to observe two kinds of measurement signals at the same time, and the operation amount is large and complicated, which is not conducive to industrial implementation.

[0005] In summary, the existing wireless communication positioning method in non-line-of-sight environment generally has the problem of being difficult to achieve high-precision positioning simply and easily. SUMMARY

[0006] In order to solve the problems in the prior art, the application provides a meerkat optimization wireless positioning method based on an improved Kalman filter, which can improve positioning accuracy and is easy to implement.

[0007] In order to achieve the above-mentioned purpose, the application adopts the following specific scheme: the meerkat optimization wireless positioning method based on the improved Kalman filter comprises the following steps:

[0008] S1, constructing a system TOA (time of arrival) ranging model in a non-line-of-sight environment, and initializing parameters of the Kalman filter;

[0009] S2, predicting and observing the target node based on the system TOA ranging model, and optimizing the parameters of the Kalman filter according to the prediction and observation results;

[0010] S3, correcting the ranging estimation value output by the system TOA ranging model by using the optimized Kalman filter, and solving the ranging estimation value combination and the coarse position of the target node at the minimum positioning accuracy factor by using a positioning algorithm based on a linear position line;

[0011] S4, solving the accurate position of the target node by using a meerkat optimization algorithm.

[0012] As a further optimization of the above-mentioned meerkat optimization wireless positioning method based on the improved Kalman filter, the specific method of S2 comprises the following steps:

[0013] S21, taking the ranging estimation value and the first-order derivative of the system TOA ranging model as the state quantity of the system, and taking the state quantity as the input quantity of the Kalman filter;

[0014] S22, solving the ranging prediction value and the ranging observation value of the system TOA ranging model, and predicting the system covariance based on time change;

[0015] S23, adjusting the Kalman gain according to the ranging prediction value and the ranging observation value;

[0016] S24, constructing an update equation of the Kalman filter, the update equation being used to output the optimal value of the system state and the optimal estimation covariance matrix of the system according to the ranging prediction value and the ranging observation value of the system TOA ranging model.

[0017] As a further optimization of the above-mentioned meerkat optimization wireless positioning method based on the improved Kalman filter, the specific method of S3 comprises the following steps:

[0018] S31, obtain the ranging observation value corresponding to the target node output by the system TOA ranging model, and correct it by using the optimized Kalman filter;

[0019] S32, divide the optimized ranging estimation value into multiple groups, and the number of ranging estimation values in each group is three;

[0020] S33, calculate the coarse position of the target node and the positioning accuracy factor corresponding to each group of ranging estimation values by using the positioning algorithm based on the linear position line;

[0021] S34, keep the group of ranging estimation values corresponding to the smallest positioning accuracy factor and the coarse position of the target node;

[0022] S35, arrange the remaining ranging estimation values from small to large, and keep the first N of them.

[0023] As a further optimization of the meerkat optimization wireless positioning method based on the improved Kalman filter described above: the specific method of S4 includes the following steps:

[0024] S41, determine the fitness function of the meerkat optimization algorithm;

[0025] S42, set the key parameters of the meerkat optimization algorithm, including the maximum number of iterations;

[0026] S43, determine the search space of the meerkat optimization algorithm by using the reserved N+3 ranging information and the coarse position of the target node, and randomize the initial positions of each individual in the search space;

[0027] S44, calculate the fitness value of each individual in the current population by the fitness function;

[0028] S45, update the density factor and the search direction;

[0029] S46, update the position of each individual in the population by using the excavation stage and the attraction stage of the meerkat optimization algorithm;

[0030] S47, update the fitness value of each individual in the population by the fitness function;

[0031] S48, if the current number of iterations reaches the maximum number of iterations or there is an adaptive value that reaches the preset threshold, then S49, otherwise return to S45;

[0032] S49, take the position of the individual with the largest fitness value as the accurate position of the target node.

[0033] As a further optimization of the meerkat optimization wireless positioning method based on the improved Kalman filter described above: in S41, the fitness function is

[0034]

[0035] wherein d i represents the ranging estimation value of the individual from the base station i, x, y are the positions of the individual, x i , y i are the positions of the base station i, and N represents the number of base stations.

[0036] As a further optimization of the meerkat optimization wireless positioning method based on the improved Kalman filter: in S42, the key parameters further include the population number N, the meerkat hunting ability, and the density factor update coefficient C.

[0037] Beneficial effects: the present application adopts the Kalman gain adaptive change Kalman filtering algorithm, corrects the ranging observation value of the system TOA ranging model in the correct direction, and then iteratively optimizes the position of the target node based on the meerkat algorithm, which can effectively suppress the NLOS error and improve the positioning accuracy, and the measurement signal is single and easy to realize. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 is the overall flowchart of the present application;

[0039] Figure 2 is the flowchart of optimizing the parameters of the Kalman filter;

[0040] Figure 3 is the flowchart of the positioning algorithm based on the linear position line;

[0041] Figure 4 is the result graph of the simulation experiment in the specific embodiment. DETAILED DESCRIPTION

[0042] The technical solutions in the embodiments of the present application will be described clearly and completely in the embodiments of the present application combined with the drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0043] Please refer to Figures 1 to 3 , the meerkat optimization wireless positioning method based on the improved Kalman filter, which comprises S1 to S4.

[0044] S1, construct a system TOA (time of arrival) ranging model in a non-line-of-sight environment, and initialize the parameters of the Kalman filter. For a wireless communication system with multiple base stations, the wireless communication devices therein can be regarded as a target node. In a non-line-of-sight environment, the system TOA ranging model can be represented by the following formula:

[0045] m k (t i )=r k (t i )+n k (t i )+nlos k (t i );

[0046] Where m k (t i ) is the target node and the kth base station at t i The measured distance at the moment r k (t i ) is the target node and the kth base station, at t i The actual distance at the moment; n k (t i ) is the systematic measurement error, which can be considered to be independently distributed, with mean zero and variance δ k Gaussian white noise; nlos k (t i ) is the kth base station at t i The non-line-of-sight error caused by the measurement process of time.

[0047] S2. Predict and observe the target node based on the system TOA ranging model, and optimize the parameters of the Kalman filter according to the prediction and observation results. The specific method of S2 includes S21 to S24.

[0048] S21, the distance estimation value of the system TOA distance measurement model and its first-order derivative are used as the state quantity of the system, and the state quantity is used as the input quantity of the Kalman filter. Specifically, Represents the state quantity of the system, where l k (t i ) is the base station k at t i The estimated distance between the time and the target node, Indicates l k (t i ) is the first-order reciprocal of .

[0049] S22, solve the ranging prediction value and ranging observation value of the system TOA ranging model, and predict the system covariance based on the moment-to-moment change. The specific transformation formula is:

[0050]

[0051]

[0052] in Indicates that base station k is at t i-1the optimal estimation result of the ranging prediction value of the target node at time t denotes the measurement matrix of base station k, and i the ranging prediction value of the target node at time t is the state transition matrix of base station k, and denotes the measurement matrix of base station k, and i the predicted covariance of the target node at time t denotes the measurement matrix of base station k, and i-1 the optimal estimation covariance of the ranging observation value of the target node at time t k is the covariance of the predicted noise. The ranging estimation value is generated based on the ranging prediction value and the ranging observation value, which belongs to the prior art and will not be described here.

[0053] S23, adjusting the Kalman gain based on the ranging prediction value and the ranging observation value. The adjustment method of the Kalman gain is:

[0054]

[0055] wherein β is a scaling factor greater than 1, H k denotes the measurement matrix of base station k, and k is the covariance of the measurement noise of base station k, and k (t i ) is the ranging observation value sequence of base station k. Considering the existence of non-line-of-sight error in the non-line-of-sight environment, the ranging observation value is often larger than the true value. By adjusting the Kalman gain, the ranging observation value can be corrected in the correct direction, thereby improving the positioning accuracy.

[0056] S24, constructing an update equation of the Kalman filter. The update equation is used to output the optimal value of the system state and the optimal estimation covariance matrix of the system according to the ranging prediction value and the ranging observation value of the system TOA ranging model. The adaptive Kalman gain solving formula is:

[0057]

[0058]

[0059] wherein I is an identity matrix.

[0060] S3, correcting the ranging observation value output by the system TOA ranging model by using the optimized Kalman filter to obtain the ranging estimation value, and solving the ranging estimation value combination and the coarse position of the target node at the minimum positioning accuracy factor by using the positioning algorithm based on the linear position line. The specific method of S3 includes S31 to S35.

[0061] S31, obtain the ranging observation value corresponding to the target node output by the system TOA ranging model, and correct it using the optimized Kalman filter.

[0062] S32, divide the optimized ranging estimation value into multiple groups, and the number of ranging estimation values in each group is three.

[0063] S33, calculate the coarse position of the target node and the positioning accuracy factor corresponding to each group of ranging estimation values using a positioning algorithm based on a linear position line.

[0064] S34, keep the group of ranging estimation values corresponding to the smallest positioning accuracy factor and the coarse position of the target node.

[0065] S35, arrange the remaining ranging estimation values from small to large, and keep the first N of them.

[0066] S4, solve the accurate position of the target node using the meerkat optimization algorithm. The specific method of S4 includes S41 to S49.

[0067] S41, determine the fitness function of the meerkat optimization algorithm. The fitness function is

[0068]

[0069] where d i represents the ranging estimation value of the individual from base station i, x and y are the positions of the individual, x i and y i are the positions of base station i, N represents the number of base stations, and the fitness value represents the average error value of the distance between the individual and each base station and the ranging result.

[0070] S42, set the key parameters of the meerkat optimization algorithm, including the maximum number of iterations t max , the population size N, the meerkat hunting ability γ, and the density factor update coefficient C.

[0071] S43, determine the search space of the meerkat optimization algorithm using the N+3 ranging information retained and the coarse position of the target node, and randomize the initial positions of each individual in the population in the search space. Specifically, the search space size is determined by drawing a ball with the target node coarse position as the center.

[0072] S44, calculate the fitness value of each individual in the current population through the fitness function.

[0073] S45, update the density factor and the search direction. In the meerkat optimization algorithm, the smell intensity I i of the prey is related to the prey concentration and the distance between meerkat individuals, and the calculation formula is as follows:

[0074]

[0075] where S is the source intensity, x prey is the position of the prey, which is also the position of the optimal individual, d i is the distance between the prey and the individual of honey badger. The formula for updating the density factor a is as follows:

[0076]

[0077] S46, the position of each individual in the population is updated in the mining stage and the attraction stage of the honey badger optimization algorithm. The mining stage performs a cardioid motion, and the formula is as follows:

[0078] x new = x prey +F x y x I x prey +F x r3 x a x d i x |cos(2p r4) x [1-cos(2p r5)];

[0079] where r3, r4, and r5 are random numbers between 0 and 1, F is a parameter for controlling the search direction, r6 is a random number between 0 and 1. In the attraction stage, the honey badger follows the guide to reach the hive, which can be expressed by the following formula:

[0080] x new = x prey +F x r7 x a x d i ;

[0081] where r7 is a random number between 0 and 1.

[0082] S47, the fitness value of each individual in the population is updated by the fitness function.

[0083] S48, if the current iteration number reaches the maximum iteration number or there is an adaptive value that reaches the preset threshold, then S49 is performed, otherwise, S45 is returned.

[0084] S49, the position of the individual with the largest fitness value is taken as the accurate position of the target node.

[0085] The method of the present application is verified by simulation experiments.

[0086] The simulation experiment condition is set as utilizing 4 base stations to participate in positioning, the base station coordinates are BS1 (0, 0), BS2 (0, 3464), BS3 (3000, 1732), BS4 (3000, -1732), and the non-line-of-sight (NLOS) error is subjected to a Gaussian distribution with a mean of 150 m, the target node moves at a constant speed from the starting point (1500, 600) to the end point (1698, 798) along a straight line for 100 s, the positioning is sampled once per second, and the root mean square error of each positioning is solved.

[0087] The actual positioning effect of the positioning method is as shown in Figure 4 Compared with various classical positioning algorithms, the present application has the best positioning performance, and the average root mean square error is reduced by 89% and 87% compared with LLOP algorithm and Chan algorithm, respectively.

[0088] The above description of disclosed embodiments enables those skilled in the art to carry out or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. The honey badger optimal wireless positioning method based on the improved Kalman filter is characterized by: The steps include: S1. Build a TOA ranging model for the system in a non-line-of-sight environment and initialize the parameters of the Kalman filter. S2. Predict and observe the target node based on the system TOA ranging model, and optimize the parameters of the Kalman filter according to the prediction and observation results; S3. Use the optimized Kalman filter to correct the ranging estimate output by the system TOA ranging model, and use the positioning algorithm based on the linear position line to solve the ranging information combination with the minimum positioning precision factor and the rough position of the target node; S4. Use the honey badger optimization algorithm to solve the exact location of the target node.

2. The honey badger optimal wireless positioning method based on the improved Kalman filter according to claim 1 is characterized in that: The specific method of S2 includes the following steps: S21. Using the ranging estimation value of the system TOA ranging model and its first-order derivative as the state quantity of the system, and using the state quantity as the input quantity of the Kalman filter; S22, solving the ranging prediction value and ranging observation value of the system TOA ranging model, and predicting the system covariance based on the time-to-time change; S23, adjusting the Kalman gain according to the ranging prediction value and the ranging observation value; S24. Construct an update equation for the Kalman filter, where the update equation is used to output the optimal value of the system state and the optimal estimated covariance matrix of the system according to the ranging prediction value and the ranging observation value of the system TOA ranging model.

3. The honey badger optimal wireless positioning method based on the improved Kalman filter according to claim 1 is characterized in that: The specific method of S3 includes the following steps: S31, obtaining the ranging observation value corresponding to the target node output by the system TOA ranging model, and correcting it using the optimized Kalman filter; S32, dividing the optimized distance estimation values ​​into multiple groups, with each group having three distance estimation values; S33, using a positioning algorithm based on a linear position line to calculate the rough position and positioning precision factor of the target node corresponding to each set of ranging estimation values; S34, retaining a set of ranging estimation values ​​corresponding to the minimum positioning precision factor and the rough position of the target node; S35. Arrange the remaining distance estimation values ​​from small to large, and retain the first N of them.

4. The honey badger optimal wireless positioning method based on the improved Kalman filter as claimed in claim 3 is characterized in that: The specific method of S4 includes the following steps: S41, determining the fitness function of the honey badger optimization algorithm; S42, setting key parameters of the honey badger optimization algorithm, the key parameters including the maximum number of iterations; S43, using the retained N+3 ranging information and the rough position of the target node to determine the search space of the honey badger optimization algorithm, and randomizing the initial position of each individual in the population in the search space; S44, calculating the fitness value of each individual in the current population through the fitness function; S45, updating the density factor and search direction; S46, using the mining phase and attraction phase of the honey badger optimization algorithm to update the position of each individual in the population; S47, updating the fitness value of each individual in the population through the fitness function; S48. If the current number of iterations reaches the maximum number of iterations or a fitness value reaches a preset threshold, proceed to S49; otherwise, return to S45; S49. The position of the individual with the largest fitness value is used as the precise position of the target node.

5. The honey badger optimal wireless positioning method based on the improved Kalman filter as claimed in claim 4 is characterized in that: In S41, the fitness function is ； in Indicates that the individual is away from the base station The distance estimate of 、 is the individual's position, 、 For base stations location, Indicates the number of base stations.

6. The honey badger optimal wireless positioning method based on the improved Kalman filter as claimed in claim 4 is characterized in that: In S42, key parameters also include population size N, honey badger hunting ability, and density factor update coefficient C.

Images