CN-121185277-B - Unmanned aerial vehicle co-location method and system under satellite refusing environment

CN121185277BCN 121185277 BCN121185277 BCN 121185277BCN-121185277-B

Abstract

The invention belongs to the technical field of navigation positioning, and particularly discloses a method and a system for co-positioning unmanned aerial vehicles in a satellite refusing environment. The invention designs a cluster unmanned aerial vehicle co-location system architecture based on relative information assistance by utilizing the relative distance information between the cluster unmanned aerial vehicles, through the architecture, the communication between an inertial navigation system and the unmanned aerial vehicle clusters can provide more reliable location service, and the accuracy of location is optimized by combining a selective correction self-adaptive particle filtering algorithm. According to the method, the self positions of the auxiliary unmanned aerial vehicles are obtained by using the inertial navigation system, the relative distance between the auxiliary unmanned aerial vehicles and the unmanned aerial vehicles to be assisted is measured by using the external source relative distance sensor, the observation positions of the unmanned aerial vehicles to be assisted are obtained, the result is brought into the improved particle filter for filtering, and the drift error of the inertial navigation system is corrected, so that the positioning precision of the collaborative positioning in the rejection environment is improved.

Inventors

YANG LEI
Li Chilong
WANG DONG
LI WEIXIN
WANG DANDAN

Assignees

济南大学

Dates

Publication Date: 20260508
Application Date: 20250908
Priority Date: 20250512

Claims (7)

1. The unmanned aerial vehicle co-location method under the satellite refusing environment is characterized by comprising the following steps: step 1, aiming at each unmanned aerial vehicle to be assisted in the unmanned aerial vehicle group, acquiring an initial position of the unmanned aerial vehicle, wherein the process is as follows: Firstly, three auxiliary unmanned aerial vehicles are selected from the auxiliary unmanned aerial vehicles, and the self position of each auxiliary unmanned aerial vehicle is obtained respectively; the unmanned aerial vehicle group comprises at least three auxiliary unmanned aerial vehicles with first precision inertial navigation systems and at least one unmanned aerial vehicle to be assisted with second precision inertial navigation systems, and a distance sensor is arranged on the auxiliary unmanned aerial vehicles; step2, after the position of the unmanned aerial vehicle to be assisted is obtained in the step 1, the flight track of the unmanned aerial vehicle to be assisted is obtained, the flight track of the unmanned aerial vehicle to be assisted is used as observation information, filtering processing is carried out on the self inertial navigation prediction track based on a selective correction self-adaptive particle filtering algorithm, inertial navigation errors are fed back and corrected, and finally the cooperative positioning of the unmanned aerial vehicle to be assisted is realized; Wherein the accuracy of the first accuracy inertial navigation system is higher than the accuracy of the second accuracy inertial navigation system; The selective correction self-adaptive particle filter algorithm improves the estimation precision by introducing a high-weight particle fine adjustment strategy between the weight calculation and resampling steps, relieves the problems of particle degradation and depletion, and simultaneously optimizes the calculation efficiency; The processing procedure of the high-weight particle fine tuning strategy is as follows: I. first, the screening threshold R and the adjustment factor of the high-weight particles are given ; And II, after the weight of the particle state is updated through the weight calculation step, executing the following operations: Screening high-weight particles, the particles satisfying the following conditions are regarded as high-weight particles: setting high weight particle index set The method comprises the following steps: wherein, the method comprises the steps of, Is the particle weight; II.2 calculating the center position of the high weight particles The formula is as follows: ; Wherein, the The state of the particles is indicated, Is the number of high weight particles; II.3 setting low weight particle index set The method comprises the following steps: ; The low weight particles are close to the high weight center position, and the calculation formula for updating the particle state is as follows: , ; Wherein, the Is the updated particle state; when the algorithm is repeated next time, the updated particle state is used as an initial value to carry out state propagation; In the processing process of the high-weight particle fine-tuning strategy, an adaptive screening threshold value is calculated based on an adaptive screening threshold value feedback algorithm To replace the screening threshold R for a given high weight particle; Wherein the method comprises the steps of The residual error change rate of the observation position and the filtering estimation position can be adaptively adjusted, and the formula is as follows: ; ; ; Wherein the method comprises the steps of 、 Respectively is 、 The difference between the Euclidean distance of the estimated position obtained by the time particle filtering and the position obtained by the least square method is the residual, As a rate of change of the residual error, For the learning rate, the step size used to control the adjustment, Representing the filtered estimated position of the object, The position of the observation is indicated, 、 Respectively represent 、 A threshold value for time; the first formula is used for obtaining Euclidean distance between the observed position and the filter estimated position as residual error, and the second formula is used for using the previous time Is the residual of (2) and the moment As a residual rate of change, a third formula for taking the difference between the residuals at the previous time To adjust the time by adding a product of the learning rate and the residual rate of change to the threshold value of (2) Is set to a threshold value of (2).
2. The method for co-locating a drone in a satellite rejection environment according to claim 1, In the step 2, in the high-weight particle fine tuning policy processing process, for the update formula of the particle state in ii.3, a direct-proportion distance fine tuning method based on radial weighted shrinkage is further provided, for updating the particle state, where the formula is: , ; , ; ; Wherein, the Indicating Euclidean distance between the ith low weight particle and the center position of the high weight particle at the moment t, Indicating the state of the low weight particles at time t i.e. the state of the particles before modification, The maximum euclidean distance between the center positions of all low-weight particles and high-weight particles at the time t is shown, As a function of the weight of the material, In order to adjust the factor(s), Representing the state of the low-weight particles at the time t+1, namely the modified low-weight particle state; Is a very small positive number, and is used for preventing the situation that the denominator is 0; the first formula is used for calculating Euclidean distances between the center positions of all low-weight particles and high-weight particles respectively as a weighting reference, the second formula is a weight function, the physical meaning of the formula is that different weights are given by different calculated Euclidean distances and used for subsequent weighting, and the third formula is that the physical meaning of the third formula is that the adjusted low-weight particle state is obtained by adding the product of the adjustment coefficient, the weight and the difference between the distances to the particle state before adjustment.
3. The method for co-locating a drone in a satellite rejection environment according to claim 1, In the step 2, the processing procedure of the selectively correcting adaptive particle filter algorithm is as follows: Step 2.1, randomly extracting N particles from prior distribution at an initial time t=0, and giving the same weight to each particle to represent various position states of unmanned opportunities, wherein N is a natural number; Step 2.2. For each time step According to the state of the system, carrying out state propagation on each particle, namely deducing the next flight positioning according to the previous position information and the current motion state; step 2.3, updating the weight of each particle according to the obtained observation value of the unmanned aerial vehicle to be assisted and the state of each particle, and normalizing the updated weight of each particle; step 2.4, introducing a high-weight particle fine adjustment strategy to carry out particle state fine adjustment; step 2.5, resampling the particles after the state fine adjustment, and re-extracting N particles according to the weight of the particles, so that the particles with large weight are duplicated for a plurality of times, and the particles with small weight are eliminated; calculating a state estimation value of the system in a weighted average mode according to the particle set after the updating, namely after the resampling in the step 2.5, and calculating the most probable position of the unmanned aerial vehicle; and 2.7, increasing t by 1, and repeating the steps 2.2 to 2.6 until all time steps are completed, so as to obtain the accurate position of each time step after filtering, and further obtain the flight track of the unmanned aerial vehicle to be assisted.
4. The method for co-locating a drone in a satellite rejection environment according to claim 3, In said step 2.1, the actual position of the unmanned aerial vehicle to be assisted is known The method comprises the following steps: the initial state of each particle is generated by adding a random variable near the true position; For the ith particle, the three components of its initial position along the three directions x, y, z are: ; Wherein the method comprises the steps of As a result of the random variable, 、、 Indicating the position of the unmanned aerial vehicle to be assisted at time 1, Representing the position of the three-dimensional ith particle at time 1; In the step 2.2, the state propagation process for each particle is as follows: ; Wherein the method comprises the steps of The position of the ith particle in the third dimension at the t time; 、、 is the position of the three-dimensional ith particle at the t-1 time; 、、 speed information at time t-1; In the step 2.3, the observation value obtained by the auxiliary unmanned aerial vehicle is set as The particles are Updating the weight formula of each particle, wherein the calculation formula is as follows: ; Wherein, the As the weight of the material to be weighed, To the extent of similarity between the observed value and the particle, here, Calculation by Euclidean distance And The distance between the two plates is set to be equal, To measure the noise standard deviation.
5. The method for co-locating a drone in a satellite rejection environment according to claim 4, In the step 2.6, a calculation formula for obtaining the state estimation value of the system by adopting a weighted average mode is as follows: ; Wherein, the Indicating the weight of the ith particle at time t, Is a state estimate of the system.
6. The method for co-locating a drone in a satellite rejection environment according to claim 1, In the step 1, the distance sensor calculates the distance between each auxiliary unmanned aerial vehicle and the unmanned aerial vehicle to be assisted according to the signal arrival time, and obtains the distance from the unmanned aerial vehicle to be assisted to three auxiliary unmanned aerial vehicles according to the Euclidean distance formula; and solving by adopting a least square method, and finally calculating the position of the unmanned aerial vehicle to be assisted.
7. A co-location system of unmanned aerial vehicle under satellite refusing environment comprises a ground control center and an unmanned aerial vehicle group; The unmanned aerial vehicle group comprises at least three auxiliary unmanned aerial vehicles with first precision inertial navigation systems and at least one unmanned aerial vehicle to be assisted with second precision inertial navigation systems, and a distance sensor is arranged on the auxiliary unmanned aerial vehicles; Wherein the accuracy of the first accuracy inertial navigation system is higher than the accuracy of the second accuracy inertial navigation system; the ground control center is provided with computer equipment, wherein the position information of the auxiliary unmanned aerial vehicle and the distance information between the auxiliary unmanned aerial vehicle and the unmanned aerial vehicle to be assisted are transmitted to the computer equipment at the ground control center; the computer device includes a memory and one or more processors; The memory has executable code stored therein, wherein the processor, when executing the executable code, is adapted to implement the steps of the method for co-locating a drone in a satellite rejection environment according to any one of claims 1 to 6.

Description

Unmanned aerial vehicle co-location method and system under satellite refusing environment Technical Field The invention belongs to the technical field of navigation positioning, and relates to a method and a system for co-positioning unmanned aerial vehicles in a satellite refusing environment. Background In recent years, unmanned aerial vehicle collaborative navigation technology is applied in different fields. The unmanned aerial vehicle is used as a self-organizing intelligent body which has small volume and low cost and can interact with surrounding individuals, and is widely applied to the fields of agriculture, rescue, military and the like to execute tasks such as detection, tracking, commercial performance and the like. Collaborative navigation is a navigation technology in a multi-agent system, and aims to improve the navigation precision, robustness and task execution efficiency of the whole system through information sharing, data fusion and collaboration among a plurality of agents (such as unmanned aerial vehicles, robots and the like). Compared with single-machine navigation, collaborative navigation can realize higher-precision positioning, path planning and collaborative task execution in complex environments (such as limited GPS signals, interference or dynamic obstacles, and the like) through means of inter-machine communication, distributed computation, environment perception, and the like. Co-location is the core of co-navigation, which enables accurate determination of relative positions among multiple agents, providing a key basis for co-navigation. And the co-location algorithm is a key technical means for realizing co-location and completing subsequent navigation tasks. The existing algorithms include Kalman filtering algorithm, particle filtering algorithm, factor graph, machine learning type and the like. Kalman filtering is an optimal estimation method based on a linear system state space model. The method uses the dynamic equation and the observation equation of the system to continuously estimate and correct the state variable through two steps of prediction and updating. However, the algorithm has higher accuracy requirements on the system model, and when the actual system has larger deviation from the model, the filtering performance can be reduced. Particle filtering is a nonlinear filtering algorithm based on Monte Carlo simulation. The probability distribution is represented by randomly sampling a large number of particles in a state space, updating the weights of the particles according to the measured values of the sensors, and then obtaining an estimated value of the state through resampling and other operations. The algorithm needs a large amount of particles to ensure the accuracy of estimation, has relatively poor real-time performance, and can cause particle degradation when applied in a high-dimensional state space. Under the environment of rejection of the global navigation satellite system, the requirement of high-precision positioning is difficult to meet by only relying on an inertial navigation system. Disclosure of Invention The invention aims to provide a co-positioning method of unmanned aerial vehicles in a satellite rejection environment, which is characterized in that prediction information and observation information of unmanned aerial vehicles to be assisted are respectively obtained through an on-board endogenous inertia and exogenous relative distance sensor of an unmanned aerial vehicle group, and filtering processing is carried out through a selective correction self-adaptive particle filtering algorithm, so that the co-positioning precision is improved. In order to achieve the above purpose, the invention adopts the following technical scheme: a method for co-locating unmanned aerial vehicles in a satellite refusing environment comprises the following steps: step 1, aiming at each unmanned aerial vehicle to be assisted in the unmanned aerial vehicle group, acquiring an initial position of the unmanned aerial vehicle, wherein the process is as follows: Firstly, three auxiliary unmanned aerial vehicles are selected from the auxiliary unmanned aerial vehicles, and the self position of each auxiliary unmanned aerial vehicle is obtained respectively; the unmanned aerial vehicle group comprises at least three auxiliary unmanned aerial vehicles with first precision inertial navigation systems and at least one unmanned aerial vehicle to be assisted with second precision inertial navigation systems, and a distance sensor is arranged on the auxiliary unmanned aerial vehicles; step2, after the position of the unmanned aerial vehicle to be assisted is obtained in the step 1, the flight track of the unmanned aerial vehicle to be assisted is obtained, the flight track of the unmanned aerial vehicle to be assisted is used as observation information, filtering processing is carried out on the self inertial navigation prediction track based on a selective correcti