CN-115752464-B - Unmanned aerial vehicle autonomous charging navigation method, device, equipment and medium

CN115752464BCN 115752464 BCN115752464 BCN 115752464BCN-115752464-B

Abstract

The invention belongs to the technical field of electric power, and particularly discloses an unmanned aerial vehicle autonomous charging navigation method, device, equipment and medium. The method comprises the steps of scanning environment information to obtain laser radar data, identifying a charging pile road sign from the laser radar data, establishing a coordinate system, processing the coordinate system through an ICP method to obtain pose gain, obtaining odometer information corresponding to the pose gain, fusing the odometer information and the pose gain through an extended Kalman filtering method to obtain an estimated pose, and completing navigation by taking the estimated pose as input of a PID control method. According to the invention, by applying the road sign recognition and positioning navigation technology, the traditional manual charging mode is replaced, the occupation of human resources is greatly reduced, the charging cost of the unmanned aerial vehicle is reduced, and the charging efficiency is improved.

Inventors

LIU ZHEN
QI XIANGHE
REN ZHIGANG
CAO ZHENBO
ZHANG XIJIA
DAI TIANZE
LI YUE

Assignees

国网北京市电力公司
国家电网有限公司

Dates

Publication Date: 20260505
Application Date: 20221031

Claims (4)

1. The unmanned aerial vehicle autonomous charging navigation method is characterized by comprising the following steps of: scanning environment information to obtain laser radar data; Identifying a charging pile marker from laser radar data, establishing a local coordinate system by taking the identified charging pile marker as a reference, wherein the charging pile marker is an isosceles triangle, and an origin of the local coordinate system is established at an intersection point of two waists of the isosceles triangle, wherein clustering is carried out on the laser radar data, a breakpoint in one frame of laser radar data is found, the laser radar data is divided into a plurality of clusters, the points in each cluster are connected into a line, and a segmentation-merging algorithm is adopted to group the clusters As input, divide into Each subset of Fitting all points in each point set S into a straight line by using a least square method, so that the distance between all points in each point set S and the straight line fitted by the point set does not exceed a first preset value, and dividing the laser radar into a plurality of clusters according to the Euclidean distance between adjacent points in laser radar data And a first preset value Classifying the size relation between the two; Clustering and linear fitting are carried out on laser radar data to identify a charging pile road sign, wherein the linear fitting adopts a segmentation-merging algorithm, and concretely comprises the steps of A1, taking a point set S as input, and A2, taking two ending points p1 and p2 in the point set S to obtain an initialization endpoint set A3, adopting a least square method to carry out Fitting straight line L to all points between the points, and calculating the points Sum point Distance between the points to line L A4, judging Maximum value of (2) A magnitude relation with a first preset value, if Greater than a first preset value, inserting a point pk in the initialization endpoint set E, Repeating the A3 segmentation And Interval point up to Less than or equal to a preset value; Obtaining pose gain under a coordinate system according to the pose transformation relation of the landmark template data and the laser radar data, wherein the landmark template data is ideal data obtained by scanning a charging pile landmark at a preset reference pose by the laser radar; The method comprises the steps of obtaining odometer information corresponding to pose gain, carrying out fusion by adopting an extended Kalman filtering method according to the odometer information and the pose gain to obtain an estimated pose of the unmanned aerial vehicle under a local coordinate system, taking the estimated pose as input of a PID control method, and controlling the unmanned aerial vehicle to move to a target charging position; when the estimated pose is obtained by fusion of the odometer information and the pose gain through an extended Kalman filtering method, the estimated pose comprises pose prediction and pose updating, wherein the pose prediction obtains observation information by calculating a jacobian matrix of a motion model and updating a covariance matrix of the pose state of the unmanned aerial vehicle, and the pose updating obtains an expression form of Kalman gain through calculating the jacobian matrix of the observation model and combining with covariance, and obtains pose estimation according to the Kalman gain and the observation information.
2. An unmanned aerial vehicle autonomous charging navigation device for implementing the unmanned aerial vehicle autonomous charging navigation method of claim 1, comprising: the laser radar is used for scanning the environmental information to obtain laser radar data; the landmark identification module is used for identifying the charging pile landmarks from the laser radar data and establishing a coordinate system; The pose gain calculation module is used for processing the coordinate system through an ICP method to obtain pose gain; The odometer is used for acquiring odometer information corresponding to the pose gain; the estimated pose module is used for obtaining an estimated pose by fusing the information of the odometer and the pose gain through an extended Kalman filtering method; and the control module is used for taking the estimated pose as the input of the PID control method to finish navigation.
3. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements a method for autonomous charging navigation of a drone according to claim 1.
4. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements a unmanned aerial vehicle autonomous charging navigation method according to claim 1.

Description

Unmanned aerial vehicle autonomous charging navigation method, device, equipment and medium Technical Field The invention belongs to the technical field of electric power, and particularly relates to an unmanned aerial vehicle autonomous charging navigation method, device, equipment and medium. Background Along with the development of unmanned aerial vehicle technology, the unmanned aerial vehicle has higher and higher requirements on autonomy and independence, and the unmanned aerial vehicle is required to work autonomously for a long time under application scenes such as industry, family service and the like. The traditional manual intervention charging mode not only prevents the intelligent unmanned aerial vehicle, but also consumes human resources, so that the unmanned aerial vehicle can safely, quickly and efficiently realize autonomous charging under the unmanned aerial vehicle intervention, and the unmanned aerial vehicle intelligent charging method is a key technology for realizing unmanned aerial vehicle intelligent. Disclosure of Invention The invention aims to provide an unmanned aerial vehicle autonomous charging navigation method, device, equipment and medium, so as to solve the technical problem that the existing unmanned aerial vehicle charging method needs human intervention and cannot be charged autonomously. In order to achieve the above purpose, the invention adopts the following technical scheme: In a first aspect, an unmanned aerial vehicle autonomous charging navigation method includes the steps of: scanning environment information to obtain laser radar data; Identifying a charging pile road sign from laser radar data, and establishing a coordinate system; processing the coordinate system by an ICP method to obtain pose gain; acquiring odometer information corresponding to pose gain; Fusion is carried out by adopting an extended Kalman filtering method according to the odometer information and the pose gain to obtain an estimated pose; and taking the estimated pose as an input of a PID control method to control the unmanned aerial vehicle to move to a target charging position. The invention further improves the method that before the charging pile mark is identified from the laser radar data, the laser radar data is clustered, and the clustering specifically comprises the following steps: Clustering the laser radar data to find out a break point in one frame of laser radar data, and dividing the laser radar data into a plurality of clusters; Connecting the points in each cluster into a line, and dividing the cluster into n subsets { S 1,...,Sn }, by adopting a segmentation-merging algorithm, and taking a clustering grouping point set S as an input; All points in each point set S are fitted into a straight line by a least square method, so that the distances between all points in each point set S and the straight line fitted by the point set are not more than a first preset value. The invention further improves the classification according to the size relation between the Euclidean distance d of the adjacent point in the laser radar data and the first preset value threshold when the laser radar is divided into a plurality of clusters. The invention further improves that when the fitting is carried out to form a straight line, the method specifically comprises the following steps: a1, taking a point set S as input; a2, taking two ending points p1 and p2 in the point set S to obtain an initialization endpoint set E= { p 1,pN }; A3, fitting all points between the points (p i,pi+1) to a straight line L by adopting a least square method, and calculating the distance d m between the points p i and p i+1 and the straight line L; and A4, judging the magnitude relation between the maximum value d k in d m and the first preset value, if d k is larger than the first preset value, inserting points pk, E= {. P i,pk,pi+1, & gt} in the initialization endpoint set E, and repeating the A3 segmentation (p i,pk) and (p k,pi+1) until d k is smaller than or equal to a preset value. The invention further improves that when the position and orientation gain is obtained by processing the coordinate system through an ICP method, the method specifically comprises the following steps: Obtaining road sign template data; Scanning and matching are carried out according to the landmark template data and the laser radar data, so that the pose transformation relation of the landmark template data and the laser radar data is obtained; And obtaining the pose gain under the coordinate system according to the pose transformation relation of the landmark template data and the laser radar data. The invention further improves that the road sign template data are ideal data obtained by scanning the charging pile road sign by the laser radar at a preset reference pose. The method is further improved in that when the estimated pose is obtained by fusion according to the odometer information and the pose gain through an extended