CN114942421A - Omnidirectional scanning multiline laser radar autonomous positioning device and method - Google Patents

Omnidirectional scanning multiline laser radar autonomous positioning device and method Download PDF

Info

Publication number
CN114942421A
CN114942421A CN202210601988.1A CN202210601988A CN114942421A CN 114942421 A CN114942421 A CN 114942421A CN 202210601988 A CN202210601988 A CN 202210601988A CN 114942421 A CN114942421 A CN 114942421A
Authority
CN
China
Prior art keywords
laser
holder
laser point
radar
characteristic
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210601988.1A
Other languages
Chinese (zh)
Inventor
许录平
阎博
张华�
孙景荣
谈婉茹
秦庆国
陈宇
杨嘉宁
张波
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xidian University
Original Assignee
Xidian University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xidian University filed Critical Xidian University
Priority to CN202210601988.1A priority Critical patent/CN114942421A/en
Publication of CN114942421A publication Critical patent/CN114942421A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
    • G01S7/481—Constructional features, e.g. arrangements of optical elements
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
    • G01S17/02—Systems using the reflection of electromagnetic waves other than radio waves
    • G01S17/06—Systems determining position data of a target
    • G01S17/46—Indirect determination of position data
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/48—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S17/00
    • G01S7/483—Details of pulse systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Electromagnetism (AREA)
  • Optical Radar Systems And Details Thereof (AREA)

Abstract

The invention discloses an omnidirectional scanning multiline laser radar autonomous positioning device and method, which are characterized in that a connecting piece is used for aligning the mass center of a multiline laser radar to the center of a holder and vertically fixing the multiline laser radar on the holder, and the holder drives a radar to rotate to realize omnidirectional scanning in one height, so that the inaccuracy of data acquisition by combining the radar with the holder is avoided. The method comprises the steps of collecting data, preprocessing the data, solving a normal vector of laser points in the data, selecting a characteristic laser point set, carrying out iterative search on the characteristic laser point set locally and globally by using a wolf optimization algorithm to obtain an optimal angle distance variable quantity, splicing the characteristic laser point set through the optimal angle distance variable quantity, and completing autonomous positioning. The invention can be used for intelligent security and robot navigation.

Description

Omnidirectional scanning multiline laser radar autonomous positioning device and method
Technical Field
The invention belongs to the technical field of radars, and further relates to an omnidirectional scanning multi-line laser radar autonomous positioning device and method in the technical field of laser radars. The device can realize autonomous positioning by omni-directional scanning of the surrounding environment through the cloud deck fixing multi-line laser radar, acquire three-dimensional point cloud data of the surrounding environment according to the omni-directional scanning of the multi-line laser radar fixing cloud deck, and realize the purpose of autonomous positioning of the device by processing the three-dimensional point cloud data.
Background
With the improvement of science and technology and the progress of society, automatic control equipment and intelligent equipment are more and more widely applied in production and life of people. Lidar is similar to human eyes, and especially the perception of the environment around an automatic device during operation of the device is important. Laser radar can avoid illumination influence to accomplish the perception to surrounding environment through photoelectric detection, laser radar who uses commonly at present mainly has single line laser radar and these two types of multi-thread laser radar, compare with single line radar, multi-thread laser radar has a plurality of receiving and dispatching passageways, multi-thread laser radar who uses commonly can carry out three-dimensional detection, but it also can only realize 360 scanning on the horizontal plane, can only acquire 30 scanning on the face of bowing, can not carry out complete perception to too close object or too far object, how to expand multi-thread laser radar's every single move scanning angle, make its characteristic information of gathering more surrounding environment be one of the key problem that will solve. Meanwhile, when enough characteristic information is acquired, how to realize accurate autonomous positioning by processing the data of two adjacent frames to sense the change of the surrounding environment is also a key problem to be solved.
The patent document "a three-dimensional lidar" (application date: 2020, 10-8, application number: 202022434234.0, application publication number: CN 212131972U) of the beijing dahan positive science and technology ltd discloses a device for three-dimensionally and omni-directionally scanning the surrounding environment by using a multi-line lidar. The device disclosed in this patent application includes a main body, a lifting device, a fixing device, a tray, and a fixing plate. When the device is used for pulling the handle bar, one end of the handle bar is fixed on the outer surface of the telescopic rod through the fixing ring, the telescopic rod is driven to control the main body, the multi-line laser radar ascends and descends, the detection height adjusting function of the multi-line laser radar is achieved, and the longitudinal visual field is enlarged. The device has the defects that the elevation platform is controlled by the handlebar to adjust the self height of the radar to obtain an extended pitching scanning angle, so that the device can complete omnidirectional scanning on the surrounding environment at pitching positions only by carrying out data acquisition at multiple heights.
The applied patent document "lidar with high density point cloud data acquisition function" (applied date: 9/1/2020/202010906109.7, application publication number: CN 111913184 a) of Jiangsu Poddaty science and technology Co., Ltd discloses a device for acquiring surrounding environment point cloud by changing the orientation of a multi-line lidar. The device that this patent application discloses includes laser radar body, cloud platform bracing piece, cloud platform rotary part, position sensor, swinging boom and mount pad. Wherein the device passes through the mount pad with multi-thread laser radar and vertically installs on the swinging boom, drives the laser radar body through cloud platform rotary part and uses cloud platform rotary part as the center and carry out the rotation scanning, and position sensor carries out data acquisition, realizes the data acquisition of surrounding environment qxcomm technology. The device has the disadvantages that the laser radar body is arranged on the rotating arm, when the laser radar body is rotated and scanned, the laser radar body carries out mass center space motion, the relative position of the radar is always changed, a one-to-many relation exists between the horizontal angle information of the holder collected by the position sensor and the azimuth angle information of the laser radar body, the measured data is distorted, errors such as distortion and the like occur on the wall surface, the floor and the like of the measured surrounding environment, and the data can not be accurately collected.
A three-dimensional laser radar and IMU combined correction point cloud distortion and odometer completion robot autonomous positioning method is disclosed in a patent document applied by northeast university in the China 'rotating laser real-time positioning modeling system and method based on sensor fusion' (application date: 7/31/2021, application number: 202110835171.6, application publication number: CN 113570715A). The method comprises the following steps: firstly, completing coordinate conversion of a three-dimensional laser radar and IMU inertia measurement unit combination; secondly, carrying out distortion removal processing on point cloud data acquired by the three-dimensional laser radar by utilizing the calibration of an IMU inertial unit; thirdly, after feature extraction and feature matching are carried out on the point cloud data processed in the second step, iterative optimization approaching to zero is carried out on distance values between planes formed by the current frame feature points and the last frame nearest three feature points, and therefore the position and pose estimation of the odometer is completed; and fourthly, completing point cloud splicing based on pose estimation in the third step, and realizing real-time three-dimensional reconstruction. The method has the disadvantages that when the position and pose estimation of the odometer is carried out on the acquired three-dimensional point cloud data, the nearest neighbor point in a frame is continuously selected for iteration through the distribution relation of iteration based on the nearest neighbor distance standard until the iteration termination condition that the distance approaches to zero is met, however, when the nearest neighbor iteration point is updated through multiple iterations, the method is easy to get the local optimal solution, the sensing failure of the change of the surrounding environment is caused, the method is inconsistent with the actual positioning result, and the autonomous positioning cannot be realized.
Disclosure of Invention
The invention aims to provide an omnidirectional scanning multiline laser radar autonomous positioning device and method aiming at the defects of the prior art, and aims to solve the problems that when the device is used for autonomous positioning, the radar has insufficient detection angle in pitching at the stage of scanning the surrounding environment, and the height of the radar needs to be changed for scanning for multiple times to obtain pitching omnidirectional laser points, and when the radar is combined with a holder to scan the surrounding environment, the relative position of the radar is changed, and the scanning of the surrounding environment is inaccurate; when the method is used for processing three-dimensional point cloud data, the solution of the nearest neighbor points is continuously updated during multiple iterations, which is easy to fall into local optimization, so that the perception of the change of the surrounding environment is failed, and the positioning result of the method is inconsistent with the actual result.
The device comprises a laser emitting end, a holder, a multi-line laser radar, a tripod head, a control device and a control device, wherein the laser emitting end of the multi-line laser radar is used for emitting laser beams in pitching to scan by changing the direction of the multi-line laser radar, the multi-line laser radar is aligned to the mass center of the tripod head, the tripod head rotates horizontally to drive the multi-line laser radar to emit the laser beams in the omnidirectional to scan, and the problems that the radar height scanning needs to be changed for many times when all information in pitching of the multi-line laser radar is obtained and the data acquisition is inaccurate when the radar is combined with the tripod head are solved. The method comprises the steps of firstly carrying out principal component analysis on acquired three-dimensional point cloud data to acquire a normal vector of the three-dimensional point cloud data, then selecting a laser point according to the change degree of the normal vector to carry out feature matching, then carrying out hunting on a wolf group through an improved wolf optimization algorithm to carry out global optimal search, locally controlling the search range, continuously carrying out iteration to balance the wolf group in global and local search, and solving the problem that the positioning is inconsistent with the reality due to failure in sensing the change of the surrounding environment.
The invention discloses an omnidirectional scanning multiline laser radar autonomous positioning device, which comprises: the device comprises a multi-line laser radar, a holder, a controller, a power module and a computer.
The multi-line laser radar is used for transmitting laser beams to a scanning area, transmitting returned echo laser beams to an object in the scanning area, and automatically calculating the interior to acquire scanned environment data;
the holder is used for driving the multi-line laser radar to horizontally rotate under the control of an internal steering engine;
the laser emitting end of the multi-line laser radar is in contact with the holder, the mass center of the multi-line laser radar is aligned with the center of the holder, the multi-line laser radar is fixed on the holder at an angle of 90 degrees in a vertical mode through a connecting piece, and the holder rotates to drive the multi-line laser radar to rotate so as to achieve accurate omnidirectional scanning of the surrounding environment;
the controller is used for transmitting the data of the holder and controlling the holder to start and stop;
the power supply module is used for supplying power to the multi-line laser radar, the controller and the computer;
and the computer is used for processing the multi-line laser radar data and the holder data.
The autonomous positioning method for the omnidirectional scanning of the multi-line laser radar comprises the following steps:
step 1, acquiring multi-line laser radar data and holder data of each rotation of a holder:
acquiring all frames of multi-line laser radar data obtained by scanning the radar for multiple circles each time the pan-tilt rotates for one circle and the horizontal azimuth angle of the pan-tilt corresponding to the data, and transmitting the data to a computer;
step 2, data preprocessing:
sequentially fusing, coordinate conversion, noise reduction and filtering the multi-line laser radar data and the corresponding holder data of each rotation of the holder;
step 3, solving normal vectors of all laser points in the three-dimensional point cloud data when the holder rotates for one circle:
taking each laser point in the three-dimensional point cloud data of each circle of rotation of the preprocessed tripod head as the center of a circle, solving a covariance matrix of each laser point in the neighborhood of the circle with the radius of 0.5cm, solving a characteristic vector and a characteristic value corresponding to the characteristic vector of the obtained covariance matrix of each laser point by adopting a singular value decomposition method, selecting the characteristic vector corresponding to the minimum characteristic value in the characteristic values of each laser point as a normal vector of the laser point, and completing the solving of normal vectors of all the laser points in the three-dimensional point cloud data of each circle of rotation of the tripod head;
and 4, selecting a characteristic laser point set of each rotation of the holder:
calculating the included angle between each laser point in each frame and each laser point in the neighborhood of each laser point in each circle of rotation of the tripod head, averaging all the included angles between each laser point in each frame and the neighborhood of each laser point, and forming a characteristic laser point set of each circle of rotation of the tripod head by the laser points with the average value larger than 5 in each frame in each circle of rotation of the tripod head;
step 5, dividing the cubic cell:
the three-dimensional corridor environment where the autonomous positioning device is located is divided into cubic cells with the side length of dcm, and each cell comprises each laser point in a characteristic laser point set when a holder rotates for one circle;
step 6, calibrating the cubic cell by using the characteristic laser point set of the selected pan-tilt rotating for one circle:
selecting a characteristic laser point set which does not select the pan-tilt for one circle of rotation, if the distance between a laser point and the center of the cubic unit in the characteristic laser point set is less than the center distance of the cubic unit
Figure BDA0003669671830000041
Scaling the cubic cell as a presence object;
and 7, acquiring the optimal angle distance variation between one rotation of the selected holder and the next circle of characteristic laser point set by using a gray wolf optimization algorithm:
step 7.1, generating the angle distance variation of each laser point in each dimension in the next circle of characteristic laser point set of the selected pan-tilt rotating for one circle:
step 7.2, sorting the angle distance variable quantity of each laser point in an ascending order, and selecting the angle distance variable quantities of the first 3 laser points;
and 7.3, updating the angle distance variation of each laser point in each dimension:
step 7.4, sorting the angle distance variable quantity of each updated laser point in an ascending order, deleting the maximum angle distance variable quantity, and supplementing randomly to generate a mean value of 0 and a variance of
Figure BDA0003669671830000042
Selecting the minimum angle distance variable quantity;
step 7.5, judging whether the distance between the laser point corresponding to the minimum angular distance variable and the center of the calibrated unit is less than 0.1cm, if so, recording the angular distance variable as the optimal angular distance variable and executing step 8, otherwise, executing step 7.2;
step 8, judging whether the characteristic laser point set is selected completely, if so, executing step 9, otherwise, executing step 6;
and 9, finishing autonomous positioning:
and splicing the characteristic laser point sets of each circle of rotation of the holder by using the optimal angle distance variation set among all the characteristic laser point sets of the rotation of the holder, and performing map construction in real time to finish autonomous positioning.
Compared with the prior art, the invention has the following advantages:
firstly, the scanning azimuth of the multi-line laser radar is changed in the device, so that the laser transmitting end scans in the pitching direction, and the horizontal azimuth of the laser transmitting end is obtained by combining the holder, thereby overcoming the defect that the prior art needs to change the radar height to acquire point cloud information to acquire all laser points in the pitching direction, and ensuring that the device can acquire surrounding environment point cloud data in an omnidirectional scanning manner at a height position.
Secondly, because the center of mass of the multi-line laser radar is aligned and fixed with the center of the holder in the device, the relative position of the radar is not changed all the time when the scanning is finished, the defects of distortion and inaccurate measurement of images of the surrounding environment caused by extending the rotating arm by using the holder in the prior art are overcome, and the accurate measurement of the surrounding environment can be finished by the device.
Thirdly, because the gray wolf optimization algorithm is adopted in the method, the global search is realized by catching the gray wolf group, the local search range is controlled, and the defect that the local optimal solution is easily caused by selecting the nearest neighbor point in the multiple iterations in the prior art is overcome, so that the method can balance the global search and the local search in the multiple iterations, complete the perception of the change of the surrounding environment and obtain accurate positioning.
Drawings
FIG. 1 is a schematic view of the apparatus of the present invention;
FIG. 2 is a schematic view of a combination of a multi-line laser radar and a pan/tilt head in the apparatus of the present invention;
FIG. 3 is a flow chart of the method of the present invention;
FIG. 4 is a simulation of the present invention.
Detailed Description
The invention is further described below with reference to the figures and examples.
The omni-directional scanning multiline lidar apparatus of the present invention is further described with reference to fig. 1 and the examples.
The omnidirectional scanning multi-line laser radar device comprises a multi-line laser radar, a holder, a controller, a power supply module and a computer. The mass center of the multi-line laser radar is aligned to the center of the holder, the holder is fixedly connected with the holder by a connecting piece, the holder controls the multi-line laser radar to rotate through an internal steering engine, the multi-line laser radar transmits laser beams to a scanning area, returned echo laser beams are transmitted through an object in the scanning area, the internal self-calculation is carried out to obtain data of the multi-line laser radar, and the holder drives the multi-line laser radar to rotate through self horizontal rotation, so that the accurate omnidirectional scanning of the surrounding environment by the multi-line laser radar is completed at one height; the controller comprises a Bluetooth and a single chip microcomputer, the Bluetooth sends holder data and receives control signals, and the single chip microcomputer sends the control signals to a steering engine in the holder to realize the start-stop control of the holder; the power supply module is realized by an external direct current power supply, and realizes power supply to the multi-line laser radar, the controller and the computer through voltage reduction processing; and the computer processes the data of the multi-line laser radar and the data of the holder through an autonomous positioning algorithm program running in the computer. Furthermore, the controller is connected with the fixedly connected multi-line laser radar and the holder through a line, so that effective transmission of control signals is guaranteed; the multi-line laser radar transmits information with the computer through the router interface and is used for feeding back data of the multi-line laser radar; the controller carries out wireless transmission with the computer through the mode of bluetooth connection for the data of feedback cloud platform.
The connection of the multi-line lidar to the pan/tilt head of the present invention is further described with reference to fig. 2.
The multi-line laser radar in fig. 2 is a cylinder, the laser emitting end of the multi-line laser radar is the side surface of the cylinder, and the tripod head is a single-degree-of-freedom tripod head controlled by a steering engine and can rotate 360 degrees horizontally. The laser transmitting end of the multi-line laser radar is contacted with the holder, the mass center of the laser transmitting end is aligned to the center of the holder, and the multi-line laser radar is fixed on the holder at an angle of 90 degrees vertically by using a connecting piece. When the holder drives the multi-line laser radar to rotate, the holder horizontally rotates for a circle, the radar rotates for multiple circles to emit laser beams to scan the surrounding environment, and omnidirectional scanning of the surrounding environment is achieved.
The autonomous positioning method of the multiline lidar omnidirectional scanning of the present invention is further described with reference to fig. 3 and the embodiment.
Step 1, acquiring multi-line laser radar data and holder data of each rotation of the holder.
Step 1.1, acquiring multi-line laser radar data and holder data of each rotation of the holder.
And the multi-line laser radar rotates and scans for a circle in the horizontal direction to obtain a frame of multi-line laser radar data, and analyzes a frame of data to obtain a radar horizontal azimuth angle, a radar pitching azimuth angle and a distance between the radar and an object in the surrounding environment.
And the controller records horizontal azimuth data of the holder corresponding to the multi-line laser radar data when the holder rotates for one week.
And 1.2, transmitting data to the computer.
The multiline laser radar uses a router interface, and multiline laser radar data obtained by scanning for a plurality of circles of radar rotation of each circle of a holder is continuously sent to a computer through a User Datagram Protocol (UDP).
Bluetooth in the controller connected with the holder is automatically paired with computer Bluetooth after being electrified, and horizontal azimuth data of the holder is transmitted to the computer.
And 2, preprocessing data.
And 2.1, fusing the data of the multi-line laser radar and the data of the corresponding holder for each rotation of the holder.
The pitching azimuth angle in the multiline laser radar data of each frame and the horizontal azimuth angle of the corresponding cloud platform are added and fused to form the horizontal azimuth angle of the three-dimensional point cloud data, the horizontal azimuth angle in the multiline laser radar data is the pitching azimuth angle of the three-dimensional point cloud data, the distance between the radar in the multiline laser radar data and the surrounding environment is the distance of the three-dimensional point cloud data, and therefore the three-dimensional point cloud data of each rotation of the cloud platform under polar coordinates is obtained.
Step 2.2, converting the three-dimensional point cloud data of the polar coordinate system into data under an XYZ space coordinate system by using the following formula:
Figure BDA0003669671830000071
wherein the content of the first and second substances,
Figure BDA0003669671830000072
representing the three-dimensional position of the ith radar laser point in the mth frame of three-dimensional point cloud data under XYZ space coordinates, wherein M is 1,2, the.
Figure BDA0003669671830000073
Representing the distance between the ith radar laser point in the mth frame of three-dimensional point cloud data and an object in the surrounding environment,
Figure BDA0003669671830000074
representing the horizontal azimuth angle measured by the ith radar laser point in the mth frame of three-dimensional point cloud,
Figure BDA0003669671830000075
and the pitch azimuth angle measured by the ith radar laser point in the mth frame of three-dimensional point cloud is represented, and multiplication operation is represented.
And 2.3, removing isolated noise points around the three-dimensional point cloud data in the XYZ space coordinate system by using a median filter, and finishing noise reduction and filtering processing.
And 3, solving normal vectors of all laser points in the three-dimensional point cloud data when the holder rotates for one circle.
Taking each laser point in the three-dimensional point cloud data of each circle of rotation of the preprocessed tripod head as the center of a circle, solving a covariance matrix of each laser point in the neighborhood of the circle with the radius of 0.5cm, solving a characteristic vector and a characteristic value corresponding to the characteristic vector of the obtained covariance matrix of each laser point by adopting a singular value decomposition method, selecting the characteristic vector corresponding to the minimum characteristic value in the characteristic values of each laser point as a normal vector of the laser point, and completing the solving of all the normal vectors of the laser points in the three-dimensional point cloud data of each circle of rotation of the tripod head.
And 4, selecting a characteristic laser point set of each rotation of the holder.
And calculating the included angle between each laser point in each frame and each laser point in the neighborhood of each laser point in each rotation circle of the tripod head, averaging all the included angles between each laser point in each frame and the neighborhood of each laser point, and forming a characteristic laser point set of each rotation circle of the tripod head by the laser points with the average value larger than 5 in each frame in each rotation circle of the tripod head.
And 5, dividing the cubic cell.
The three-dimensional corridor environment where the autonomous positioning device is located is divided into cubic cells with the side length of 1cm, and each cell comprises each laser point in the characteristic laser point set when the holder rotates for one circle.
And 6, calibrating the cubic cell by using the characteristic laser point set of the selected pan-tilt rotating for one circle.
Selecting a characteristic laser point set which does not select the pan-tilt for one circle of rotation, if the distance between a laser point and the center of the cubic unit in the characteristic laser point set is less than the center distance of the cubic unit
Figure BDA0003669671830000076
The cubic cell is labeled as a presence object.
And 7, acquiring the optimal angle distance variation between the characteristic laser point set of one circle of rotation of the selected holder and the next circle of rotation by utilizing a gray wolf optimization algorithm.
Step 7.1, according to the following formula, generating the angle distance variation S of each laser point in each dimension in the next circle characteristic laser point set of the selected pan-tilt rotating for one circle ju ：
Figure BDA0003669671830000081
Wherein j represents the jth laser point in the characteristic laser point set of the next circle of the rotation of the selected pan-tilt, and u is shown asThe u-th dimension, j being 1, 2., K, u being 1, 2.., 6, K representing the total number of laser points in the next characteristic set of laser points for one rotation of the selected head,
Figure BDA0003669671830000082
representing a mean of 0 and a variance of
Figure BDA0003669671830000083
Random number of (2), r in the present embodiment max ＝{10,10,10,20,20,20}。
And 7.2, sorting the angle distance variable quantities of each laser point in an ascending order, and selecting the angle distance variable quantities of the first 3 laser points.
Step 7.3, updating the angle distance variation of each laser point in each dimension according to the following formula:
Figure BDA0003669671830000084
wherein S is j ' u Indicating the amount of angular distance change, S, in the u-th dimension of each updated laser spot pu The angular distance variation of the u-th dimension of the p-th laser point is represented, p is 1,2 and 3, epsilon is linearly decreased from 2 to 0 along with the increase of the iteration number, and tau represents [0,1]The random number of (2).
Step 7.4, sorting the angle distance variable quantity of each updated laser point in an ascending order, deleting the maximum angle distance variable quantity, and supplementing randomly to generate a mean value of 0 and a variance of
Figure BDA0003669671830000085
The minimum angular distance variation is selected.
And 7.5, judging whether the distance between the laser point corresponding to the minimum angular distance variable and the center of the calibrated unit is less than 0.1cm, if so, recording the angular distance variable as the optimal angular distance variable and executing the step 8, otherwise, executing the step 7.2.
And 8, judging whether the characteristic laser point set is selected completely, if so, executing a step 9, otherwise, executing a step 6.
And 9, finishing the autonomous positioning.
And splicing the characteristic laser point sets of each circle of rotation of the holder by using the optimal angle distance variation set among all the characteristic laser point sets of the rotation of the holder, and performing map construction in real time to finish autonomous positioning.
The effects of the present invention will be further explained by the following experiments:
1. conditions of the experiment:
the hardware device of the simulation experiment of the invention is as follows: an omnidirectional scanning device consisting of a 16-line laser radar of Velodyne company and a PWM single-degree-of-freedom holder, and a controller consisting of a stm32 singlechip and an HC05 Bluetooth.
The experimental data processing platform of the invention is as follows: windows 10 operating system and Matlab2020 b.
The experimental scene of the invention is as follows: a 40 meter long corridor, two moving pedestrian targets, target 1, target 2, respectively.
2. And (3) analyzing the experimental content and the result:
the experiment of the invention is that the device of the invention is placed on the ground of a corridor, the device is manually moved, the corridor environment with the length of 40 meters, the width of 10 meters and the height of 3 meters is scanned in an omnidirectional way to obtain point cloud data, and then the point cloud data obtained by scanning is processed by utilizing an autonomous positioning method (a Grey wolf optimization algorithm) to obtain a result graph of the autonomous positioning of the device, as shown in figure 4.
In the experiment, the adopted autonomous positioning method refers to an optimized solving method provided by Mirjalili et al in Grey Wolf Optimizer [ J ] Advances in Engineering Software,2014,69:46-61, which is called a Wolf optimization algorithm for short.
The invention is further described below in conjunction with the results chart of fig. 4.
Fig. 4(a) is a diagram of a real-time positioning result of the present apparatus using the autonomous positioning method, and fig. 4(b) is a diagram of an autonomous mapping result of the present apparatus using the autonomous positioning method.
Fig. 4(a) shows that the real-time positioning result of the sirius optimization algorithm in the autonomous positioning method, the circle represents the real-time positioning completed by the device, and the method of the line sent by the circle represents the forward direction of the multi-line laser radar in the device.
Fig. 4(b) shows the result of autonomous mapping of the self-localization method grayish wolf optimization algorithm, wherein the X-axis represents the length of the corridor environment in which the device of the present invention is located, the Y-axis represents the width of the corridor environment in which the device of the present invention is located, and the Z-axis represents the height of the corridor environment in which the device of the present invention is located.
The experimental results show that the device composed of the multi-line laser radar and the holder can complete omnidirectional scanning, the gray wolf optimization method is used for solving the position and posture variation quantity between adjacent frames, real-time positioning and autonomous map building can be completed, and even if iteration is carried out for multiple times, the change of the surrounding environment can still be perceived to realize accurate autonomous positioning.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (7)

1. An omnidirectional scanning multi-line laser radar autonomous positioning device comprises a multi-line laser radar, a holder, a controller, a power supply module and a computer; the multi-line laser radar system is characterized in that the multi-line laser radar is fixedly connected with the holder through a connecting piece, and the holder rotates to drive the multi-line laser radar to rotate so as to realize accurate omnidirectional scanning on the surrounding environment; wherein:
the multi-line laser radar is used for transmitting laser beams to a scanning area, transmitting returned echo laser beams to an object in the scanning area, and automatically calculating the interior to acquire scanned environment data;
the holder is used for driving the multi-line laser radar to horizontally rotate under the control of an internal steering engine;
the laser transmitting end of the multi-line laser radar is in contact with the holder, the mass center of the multi-line laser radar is aligned to the center of the holder, the multi-line laser radar is fixed on the holder at an angle of 90 degrees vertically by using a connecting piece, and the holder rotates to drive the multi-line laser radar to rotate so as to realize accurate omnidirectional scanning on the surrounding environment;
the controller is used for transmitting the data of the holder and controlling the holder to start and stop;
the power supply module is used for supplying power to the multi-line laser radar, the controller and the computer;
and the computer is used for processing the multi-line laser radar data and the holder data.
2. The method for autonomously positioning the multi-line lidar in omnidirectional scanning according to the claim 1, wherein the gray wolf optimization algorithm is used to obtain the optimal angular distance variation between one rotation of the selected pan-tilt and the next characteristic laser point set; the autonomous positioning method comprises the following steps:
step 1, acquiring multi-line laser radar data and holder data of each rotation of a holder:
acquiring all frames of multi-line laser radar data obtained by scanning for multiple circles of radar rotation every time the holder rotates for one circle and transmitting the data and the corresponding horizontal azimuth angle of the holder to a computer;
step 2, data preprocessing:
sequentially fusing, coordinate conversion, noise reduction and filtering the multi-line laser radar data and the corresponding holder data of each rotation of the holder;
step 3, solving normal vectors of all laser points in the three-dimensional point cloud data of each rotation of the holder:
taking each laser point in the three-dimensional point cloud data of each circle of rotation of the pre-processed tripod head as the center of a circle, solving a covariance matrix of each laser point in the neighborhood of the circle with the radius of 0.5cm, solving a characteristic vector and a characteristic value corresponding to the characteristic vector of the covariance matrix of each obtained laser point by adopting a singular value decomposition method, selecting the characteristic vector corresponding to the minimum characteristic value in the characteristic values of each laser point as a normal vector of the laser point, and completing the solving of all the normal vectors of the laser points in the three-dimensional point cloud data of each circle of rotation of the tripod head;
and 4, selecting a characteristic laser point set of each rotation of the holder:
calculating the included angle between each laser point in each frame and each laser point in the neighborhood of each laser point in each circle of rotation of the tripod head, averaging all the included angles between each laser point in each frame and the neighborhood of each laser point, and forming a characteristic laser point set of each circle of rotation of the tripod head by the laser points with the average value larger than 5 in each frame in each circle of rotation of the tripod head;
step 5, dividing the cubic cell:
the three-dimensional corridor environment where the autonomous positioning device is located is divided into cubic cells with the side length of dcm, and each cell comprises each laser point in a characteristic laser point set when a holder rotates for one circle;
step 6, calibrating the cubic cell by using the characteristic laser point set of the selected pan-tilt rotating for one circle:
selecting a characteristic laser point set which does not select the pan-tilt for one circle of rotation, if the distance between a laser point and the center of the cubic unit in the characteristic laser point set is less than the center distance of the cubic unit
Figure FDA0003669671820000021
Scaling the cubic cell as a presence object;
and 7, acquiring the optimal angle distance variation between one rotation of the selected holder and the next circle of characteristic laser point set by using a gray wolf optimization algorithm:
step 7.1, generating the angle distance variation of each laser point in each dimension in the next circle of characteristic laser point set of the selected pan-tilt rotating for one circle:
step 7.2, sorting the angle distance variable quantity of each laser point in an ascending order, and selecting the angle distance variable quantities of the first 3 laser points;
step 7.3, updating the angle distance variable quantity of each laser point on each dimension;
step 7.4, sorting the angle distance variable quantity of each updated laser point in an ascending order, deleting the maximum angle distance variable quantity, and supplementing randomly to generate a mean value of 0 and a variance of
Figure FDA0003669671820000022
Selecting the minimum angle distance variable quantity;
step 7.5, judging whether the distance between the laser point corresponding to the minimum angular distance variable quantity and the center of the calibrated unit is less than 0.1cm, if so, recording the angular distance variable quantity as the optimal angular distance variable quantity and executing step 8, otherwise, executing step 7.2;
step 8, judging whether the characteristic laser point set is selected completely, if so, executing step 9, otherwise, executing step 6;
and 9, finishing autonomous positioning:
and splicing the characteristic laser point sets of each circle of rotation of the holder by using the optimal angle distance variation set among all the characteristic laser point sets of the rotation of the holder, and performing map construction in real time to finish autonomous positioning.
3. The omni-directional scanning multiline lidar self-positioning method according to claim 2, wherein the fusion in step 2 is: and adding the pitching azimuth angle in the multi-line laser radar data of each frame and the horizontal azimuth angle of the corresponding holder to fuse the pitching azimuth angle and the horizontal azimuth angle into the horizontal azimuth angle of the three-dimensional point cloud data, wherein the horizontal azimuth angle in the multi-line laser radar data is the pitching azimuth angle of the three-dimensional point cloud data, and the distance between the radar in the multi-line laser radar data and the surrounding environment is the distance of the three-dimensional point cloud data.
4. The omni-directional scanning multi-line lidar self-positioning method according to claim 2, wherein the coordinate transformation in step 2 is to convert the three-dimensional point cloud data of the fused polar coordinate system into data in XYZ spatial coordinate system by using the following formula:
Figure FDA0003669671820000031
wherein the content of the first and second substances,
Figure FDA0003669671820000032
the three-dimensional position of the ith radar laser point in the mth frame of three-dimensional point cloud data under XYZ space coordinates is represented, wherein m is1,2, wherein M, i is 1,2, N, M represents the total number of frames obtained by the tripod head every rotation, N represents the total number of laser points in a frame of three-dimensional point cloud data,
Figure FDA0003669671820000033
representing the distance measured by the ith radar laser point in the mth frame of three-dimensional point cloud data and an object in the surrounding environment,
Figure FDA0003669671820000034
representing the horizontal azimuth angle measured by the ith radar laser point in the mth frame of three-dimensional point cloud,
Figure FDA0003669671820000035
and the pitch azimuth angle measured by the ith radar laser point in the mth frame of three-dimensional point cloud is represented, and multiplication operation is represented.
5. The omni-directional scanning multiline lidar self-localization method according to claim 2, wherein the denoising and filtering process in step 2 is to remove isolated noise points existing around the coordinate-converted three-dimensional point cloud data by using a median filter.
6. The method according to claim 2, wherein the step 7.1 of generating the angle distance variation of each laser point in each dimension in the next round of characteristic laser point set of the selected pan/tilt head rotating for one round is generated by the following formula:
Figure FDA0003669671820000041
wherein S is ju Representing the angle distance variation of the jth laser point in the jth dimension in the generated characteristic laser point set of the next circle of rotation of the selected tripod head, j representing the jth laser point in the characteristic laser point set of the next circle of rotation of the selected tripod head, u representing the jth dimension,j 1,2, K, u 1,2, 6, K representing the total number of laser points in the set of characteristic laser points for the next revolution of the selected head,
Figure FDA0003669671820000042
representing a mean of 0 and a variance of
Figure FDA0003669671820000043
The random number of (2).
7. The method as claimed in claim 6, wherein the step 7.3 of updating the angular distance variation of each laser spot in each dimension is implemented by the following formula:
Figure FDA0003669671820000044
wherein S is j ' u Represents the angular distance variation, S, of the updated jth laser point in the u-th dimension pu The angular distance variation of the u-th dimension of the p-th laser point is represented, p is 1,2 and 3, epsilon is linearly decreased from 2 to 0 along with the increase of the iteration number, and tau represents [0,1]The random number of (2).
CN202210601988.1A 2022-05-30 2022-05-30 Omnidirectional scanning multiline laser radar autonomous positioning device and method Pending CN114942421A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210601988.1A CN114942421A (en) 2022-05-30 2022-05-30 Omnidirectional scanning multiline laser radar autonomous positioning device and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210601988.1A CN114942421A (en) 2022-05-30 2022-05-30 Omnidirectional scanning multiline laser radar autonomous positioning device and method

Publications (1)

Publication Number Publication Date
CN114942421A true CN114942421A (en) 2022-08-26

Family

ID=82908481

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210601988.1A Pending CN114942421A (en) 2022-05-30 2022-05-30 Omnidirectional scanning multiline laser radar autonomous positioning device and method

Country Status (1)

Country Link
CN (1) CN114942421A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116068572A (en) * 2022-12-09 2023-05-05 中建材凯盛机器人（上海）有限公司 System, method, device, processor and computer readable storage medium for realizing vehicle body contour detection processing based on laser radar

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116068572A (en) * 2022-12-09 2023-05-05 中建材凯盛机器人（上海）有限公司 System, method, device, processor and computer readable storage medium for realizing vehicle body contour detection processing based on laser radar

Similar Documents

Publication Publication Date Title
CN110356976B (en) Port unmanned truck crane accurate alignment system and method based on template matching
CN108508430B (en) Laser radar rotation control method for target detection
CN109356614B (en) Tunnel injection machine control equipment and method and tunnel injection machine
CN112605033B (en) Photovoltaic panel posture recognition and cleaning regulation and control device and method
CN110641721B (en) Boarding bridge parking method
CN110065074A (en) A kind of the visual servo laser orientation system and method for picking robot
JP2023133088A (en) Mobile flexible measurement system and measurement method for large complex component
WO2022016754A1 (en) Multi-machine cooperative vehicle washing system and method based on unmanned vehicle washing device
CN112327898B (en) Unmanned aerial vehicle well inspection navigation method and device and unmanned aerial vehicle
CN113066120B (en) Intelligent pole and tower inclination detection method based on machine vision
CN112363176B (en) Elevator hoistway inspection and modeling method and device and inspection and modeling system
CN113155047B (en) Long-distance hole distance measuring device and method, storage medium, equipment and rail vehicle
CN108564628A (en) A kind of cutterhead vision positioning orientation system towards development machine automation
CN114942421A (en) Omnidirectional scanning multiline laser radar autonomous positioning device and method
CN108196538B (en) Three-dimensional point cloud model-based field agricultural robot autonomous navigation system and method
CN110322462B (en) Unmanned aerial vehicle visual landing method and system based on 5G network
CN108876862A (en) A kind of noncooperative target point cloud position and attitude calculation method
CN106643521B (en) A kind of detection method and device of corps canopy height
CN113177918B (en) Intelligent and accurate inspection method and system for electric power tower by unmanned aerial vehicle
CN112257536B (en) Space and object three-dimensional information acquisition and matching equipment and method
CN112478968A (en) Elevator hoistway inspection control method, device and system and storage medium
WO2022078437A1 (en) Three-dimensional processing apparatus and method between moving objects
CN109213154A (en) One kind being based on Slam localization method, device, electronic equipment and computer storage medium
CN115371599A (en) High-precision ground flatness measuring system and method
CN114646976A (en) Single-line laser radar road surface detection device based on holder and detection method thereof

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination