A composite map construction method and system based on laser technology

Abstract

This invention relates to a method and system for constructing composite maps based on laser technology, belonging to the field of artificial intelligence technology. The method includes: installing positioning rods on various slopes in an application scenario; scanning the application scenario using 2D laser technology; determining the area where a mobile robot is located; when the mobile robot is on a plane, constructing a planar grid map of the plane, the planar grid map including height information; when the mobile robot is on a slope, constructing a slope point cloud map based on the positioning rods; and stitching and fusing all the planar grid maps and slope point cloud maps to form the composite map. This invention combines simple positioning rods with mature and low-cost 2D laser technology to achieve high-precision composite map construction for non-planar application scenarios, and has the advantages of system simplicity and low cost.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method and system for constructing composite maps based on laser technology. Background Technology

[0002] In recent years, with the development of mobile robot technology, mobile robots have been widely used in various scenarios, such as factory transportation, hotel delivery, and home cleaning. Currently, in scenarios with flat ground, mobile robots use 2D LiDAR technology for localization mapping. In scenarios with uneven ground, such as factories with ramps for covering wiring, buildings with elevation differences and ramps between interfaces on the same floor, and underground parking lots with entrances, exits, and speed bumps, mobile robots use 3D LiDAR technology for localization mapping.

[0003] However, in scenarios with uneven ground, mobile robots using 2D laser technology to build localization maps will scan high slopes as obstacles and treat them as impassable areas, resulting in a localization map that does not match the actual scene. While mobile robots can use 3D LiDAR technology to solve the above problems, it makes the system more complex and increases the cost of hardware and software. Therefore, for scenarios with uneven ground, it is necessary to develop a low-cost and high-precision composite map building method and system. Summary of the Invention

[0004] To address the challenge of building low-cost, high-precision maps in uneven terrain, this invention provides a composite map construction method and system based on laser technology.

[0005] To solve the above technical problems, the present invention adopts the following technical solution:

[0006] In a first aspect, embodiments of the present invention provide a method for constructing composite maps based on laser technology, comprising the following steps:

[0007] Install the positioning rods on various slopes in the application scenario;

[0008] Using 2D laser technology to scan the application scenario;

[0009] The location of the mobile robot is determined. When the mobile robot is on a plane, a planar grid map is constructed on the plane, which includes height information. When the mobile robot is on a slope, a slope point cloud map is constructed based on the positioning rod.

[0010] The composite map is created by stitching together all the planar raster maps and slope point cloud maps.

[0011] In some embodiments, the composite map construction method further includes the step of:

[0012] During the construction of the composite map, the laser scanning direction is adjusted to keep it parallel to the horizontal plane at all times.

[0013] In some embodiments, the composite map construction method further includes the step of:

[0014] When a mobile robot leaves one area and enters another, it saves the map information of the area it left, which includes the map of the area it left, the location of the boundary between the two areas, and the posture of the mobile robot at this time.

[0015] The composite map is created by stitching together all the planar raster maps and slope point cloud maps:

[0016] Based on the map information, all the planar raster maps and slope point cloud maps are stitched together and merged to form a composite map.

[0017] In some embodiments, constructing a planar raster map of the plane includes:

[0018] Use an encoder to calculate the odometry of a mobile robot;

[0019] Based on the odometer, a planar grid map is constructed for the plane in which it is located.

[0020] In some embodiments, determining the area where the mobile robot is located includes:

[0021] The acceleration, pitch angle, and roll angle of the mobile robot were measured using inertial measurement sensors.

[0022] The location of the mobile robot is determined based on data from inertial measurement sensors.

[0023] In some embodiments, installing the positioning rod on each slope in the application scenario is as follows:

[0024] The positioning rods are installed sequentially on the sides of each slope in the application scenario, starting from the bottom of the slope and arranged vertically to the horizontal plane, with the top of the previous positioning rod and the bottom of the next positioning rod on the same horizontal plane in each pair of adjacent positioning rods.

[0025] In some embodiments, the positioning rod is a tapered vertical rod with a regular hexagonal cross-section.

[0026] Secondly, embodiments of the present invention provide a composite map construction system based on laser technology, comprising: a positioning rod and a mobile robot;

[0027] The positioning rod is used for laser positioning and is installed on various slopes in the application scenario;

[0028] The mobile robot includes a laser scanning module, a region determination and construction module, and a fused map module;

[0029] The laser scanning module uses 2D laser technology to scan the application scene;

[0030] The region determination and construction module is used to determine the region where the mobile robot is located. When the mobile robot is on a plane, a planar grid map is constructed on the plane, and the planar grid map includes height information. When the mobile robot is on a slope, a slope point cloud map is constructed on the slope based on the positioning rod.

[0031] The fusion map module is used to stitch together and merge all planar raster maps and slope point cloud maps to form a composite map.

[0032] In some embodiments, the laser scanning module adjusts the laser scanning direction during the operation of the composite map building system so that it remains parallel to the horizontal plane at all times.

[0033] In some embodiments, the mobile robot further includes: a map information storage module;

[0034] When a mobile robot leaves one area and enters another, the map information storage module stores the map information of the area it leaves. The map information includes the map of the area it leaves, the location of the boundary between the two areas, and the posture of the mobile robot at this time.

[0035] The fusion map module, based on the map information, stitches and merges all the planar raster maps and slope point cloud maps to form a composite map.

[0036] In some embodiments, the mobile robot further includes: an encoder;

[0037] The encoder is used to calculate the mobile robot's odometer reading;

[0038] The mobile robot constructs a planar grid map of the plane it is on based on the odometer.

[0039] In some embodiments, the mobile robot further includes an inertial measurement sensor;

[0040] The inertial measurement sensor is used to measure the acceleration, pitch angle, and roll angle of the mobile robot.

[0041] The region determination module is used to determine the region where the mobile robot is located based on the inertial measurement sensor data.

[0042] In some embodiments, the positioning rod is a tapered vertical rod with a regular hexagonal cross-section.

[0043] The present invention provides a method and system for constructing composite maps based on laser technology. Compared with the prior art, the technical effects achieved by the present invention include:

[0044] 1. This invention installs low-cost positioning rods on various slopes in the application scenario, uses mature and low-cost 2D laser technology to scan the application scenario, constructs a planar grid map for the plane, and constructs a slope point cloud map based on the positioning rods. All the planar grid maps and slope point cloud maps are stitched and merged to form the composite map. This effectively solves the problem that existing 2D laser technology cannot construct maps for non-planar application scenarios, and also solves the problems of low accuracy, high cost, and high system complexity when using 3D laser technology to construct maps. In other words, this invention combines simple positioning rods with mature and low-cost 2D laser technology to achieve high-precision composite map construction for non-planar application scenarios, and has the advantages of system simplicity and low cost.

[0045] 2. In the process of constructing a composite map, this invention adjusts the laser scanning direction to keep it parallel to the horizontal plane, thereby reducing the algorithm complexity for calculating the position of the positioning rod.

[0046] 3. When a mobile robot leaves one area and enters another, the present invention saves the map information of the area it leaves. The map information includes the map of the area it leaves, the location of the boundary between the two areas, and the posture of the mobile robot at this time. Based on the map information, the relationship between the stitching and fusion of the various maps can be found efficiently, making the map at the boundary between the various areas clearer. This makes it faster, more accurate, and more precise to stitch and fuse all the planar grid maps and slope point cloud maps into a composite map.

[0047] 4. This invention uses an encoder to calculate the odometer of a mobile robot, and based on the odometer, uses laser technology to construct a planar grid map of the plane, thereby making the constructed application scenario map more accurate.

[0048] 5. This invention uses an inertial measurement sensor to measure the acceleration, pitch angle, and roll angle of the mobile robot, and based on the data from the inertial measurement sensor, determines the area where the mobile robot is located, reducing the degree of human area judgment and human involvement in system operation, making the operation of the composite map construction system simpler and reducing the labor cost of composite map construction.

[0049] 6. The present invention installs positioning rods sequentially on the sides of each slope in the application scenario, starting from the bottom of the slope and arranged in a vertical and horizontal manner. In each pair of adjacent positioning rods, the top of the previous positioning rod and the bottom of the next positioning rod are located on the same horizontal plane, which reduces the computational complexity of the mobile robot's positioning and makes the construction of the slope point cloud map faster and more efficient.

[0050] 7. This invention uses a tapered vertical rod with a regular hexagonal cross-section for constructing the slope point cloud map, which enables the mobile robot to receive more and more regular laser reflection points from the positioning rod, making the positioning rod more accurate, thereby making the slope point cloud map construction more accurate and the calculation processing speed faster. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below only involve some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 This is a flowchart of the laser technology-based composite map construction method of the present invention;

[0053] Figure 2 This is a flowchart of step S16 of the present invention;

[0054] Figure 3a , 3b This is a simplified schematic diagram illustrating an application scenario of the present invention;

[0055] Figure 4 This is a schematic diagram of the positioning rod structure of the present invention;

[0056] Figure 5 This is a schematic diagram of the positioning rod installation of the present invention;

[0057] Figure 6 This is a schematic diagram of the positioning rod detection and positioning algorithm of the present invention;

[0058] Figure 7 This is a schematic diagram of the composite map construction system based on laser technology of the present invention;

[0059] Figure 8 This is a simplified model structural diagram of the mobile robot of the present invention;

[0060] Figure 9 This is a schematic diagram of the lidar gimbal structure of the present invention. Detailed Implementation

[0061] To enable those skilled in the art to better understand the solutions of the present invention, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, other solutions obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0062] To facilitate understanding of the embodiments of the present invention by those skilled in the art, the technical terms involved in the present invention are explained below.

[0063] Mobile robots using 2D laser technology are widely used in various industries due to their small size, low cost, and good performance. These mobile robots use 2D laser technology to scan and perceive the location information of obstacles in their application scene to build a planar grid map. However, the application scene in which the mobile robot works may have slopes, or even multiple planes connected by slopes, with the slope height higher than the height of the mobile robot's laser scan. When building a map in such an application scene, the mobile robot will inevitably scan the slope and the higher plane as obstacles and regard them as impassable areas. If the mobile robot is manually controlled to build a map of the slope using existing 2D laser technology, the map it builds will have huge errors and cannot be used for map navigation and positioning.

[0064] Example 1

[0065] Figure 1 This invention illustrates a flowchart of a composite map construction method based on laser technology, comprising... Figure 1 A method for constructing composite maps based on laser technology includes the following steps:

[0066] S12. Install the positioning rods on various slopes in the application scenario.

[0067] Application scenarios of the present invention are as follows Figure 3a As shown, it has two different height planes connected by a ramp, or has multiple ramps connecting two different height planes (not shown in the figure), or as... Figure 3b As shown, it has three planes at different heights connected by two ramps. This embodiment uses... Figure 3a The application scenarios shown illustrate the invention; other application scenarios are basically the same and will not be described in detail here.

[0068] The positioning rod of this invention is preferably a vertical rod of a specific shape, such as a conical vertical rod with a regular hexagonal cross-section, preferably 50cm in length, with a top hexagonal side length of 2cm and a bottom hexagonal side length of 22cm. Figure 4As shown. The positioning rod is made of laser-reflective material or its surface is covered with a layer of laser-reflective material to save costs. It is used to install on or near various slopes in the application scenario to clearly reflect the laser light for accurate distance measurement. Preferably, as... Figure 5 As shown, the application scenario consists of two planes 3 connected by a ramp 4. Positioning rods 2 are installed on the sides and front and back of the ramp 4 in a vertical and horizontal manner. Furthermore, the positioning rods 2 are sequentially installed on the sides of each ramp 4 in the application scenario, starting from the bottom of the ramp 4 and arranged vertically and horizontally, with the top of the preceding positioning rod and the bottom of the following positioning rod on the same horizontal plane in each pair of adjacent positioning rods. It is preferable that the distance between the positioning rods and surrounding obstacles is greater than 10cm.

[0069] Using a tapered vertical rod with a regular hexagonal cross-section for constructing the slope point cloud map allows the mobile robot to receive more and more regular laser reflection points from the positioning rod, making the positioning rod more accurate. This results in higher accuracy in constructing the slope point cloud map and faster computation speed.

[0070] Positioning rods are installed sequentially on the sides of each slope in the application scenario, starting from the bottom of the slope and arranged vertically and horizontally. In each pair of adjacent positioning rods, the top of the previous positioning rod and the bottom of the next positioning rod are located on the same horizontal plane. This reduces the computational complexity of the mobile robot's positioning and makes the construction of the slope point cloud map faster and more efficient.

[0071] S14. Use 2D laser technology to scan the application scene.

[0072] The mobile robot of this invention uses existing 2D laser technology to scan the application scene, such as... Figure 5 As shown, when the mobile robot 1 moves to the intersection of plane 3 and ramp 4, it will scan the positioning rod 2.

[0073] S16. Determine the area where the mobile robot is located. When the mobile robot is on a plane, construct a planar grid map of the plane, which includes height information. When the mobile robot is on a slope, construct a slope point cloud map based on the positioning rod. Figure 2 As shown, step S16 is implemented by the following sub-steps:

[0074] S162. Determine the area where the mobile robot is located.

[0075] The mobile robot uses 2D laser technology to scan the application scene and determine its location. The robot can be manually controlled to determine whether the area is flat or sloped. Preferably, inertial measurement sensors are used to measure the robot's acceleration, pitch angle, and roll angle. Based on this data, the robot's location is determined, reducing human intervention in area determination and system operation. This simplifies the composite map construction method and system operation, reducing the manual cost of composite map construction. Specific area condition determination is as follows:

[0076] Condition 1: The absolute value of the acceleration of the mobile robot in the Z-axis / height direction is less than a predetermined value, for example, 9 m / s². 2 And the duration is greater than the predetermined value, for example, 1 second; the absolute values ​​of the mobile robot's pitch angle and roll angle are greater than the predetermined value, for example, 10 degrees, and the duration is greater than the predetermined value, for example, 1 second;

[0077] Condition 2: The absolute value of the acceleration of the mobile robot in the Z-axis / height direction is greater than a predetermined value, for example, 9.3 m / s². 2 The sum of the absolute values ​​of the pitch and roll angles of the mobile robot is less than the predetermined value, for example, 10 degrees, and the duration is greater than the predetermined value, for example, 1 second.

[0078] When condition 1 is met, it is determined that the mobile robot has entered the ramp. After entering the ramp, the mobile robot meets condition 2, and it is determined that the mobile robot has left / exited the ramp.

[0079] S164. When the mobile robot is located on a plane, a planar grid map is constructed on the plane, and the planar grid map includes height information.

[0080] When the mobile robot is located on a plane, a planar grid map is constructed on the plane using 2D laser technology. For example, mature algorithms such as Cartographer, CoreSLAM, and Karto-SLAM are used to construct the planar grid map on the plane. The planar grid map constructed by this invention contains height information to distinguish planar areas at different heights.

[0081] Preferably, an encoder is used to calculate the mobile robot's odometer, and based on the odometer, 2D laser technology is used to construct a planar grid map of the plane, thereby making the constructed application scenario map more accurate.

[0082] S166. When the mobile robot is located on a slope, a point cloud map of the slope is constructed based on the positioning rod. The specific implementation method is as follows:

[0083] The mobile robot uses 2D laser technology to scan and acquire laser scanning data. Based on the scanning data, a positioning rod detection and positioning algorithm is used to obtain the position of the positioning rod and the radius of the positioning rod cross-section at the corresponding scanning height.

[0084] When the mobile robot moves to the bottom or top of the slope and enters the slope, it saves the current planar grid map and the initial position upon entering the slope. The mobile robot moves along the slope while simultaneously performing laser scanning, acquiring laser scan data for each frame. Preferably, while moving along the slope, the mobile robot adjusts its laser scanning direction to keep it parallel to the horizontal plane at all times. This reduces the complexity of the positioning rod detection and positioning algorithm, thereby making the application of the composite map construction method in this embodiment more efficient and simpler.

[0085] The positioning rod detection and positioning algorithm is as follows:

[0086] Step 1: Obtain the current frame laser scan data P = {p1, p2, ..., pn}, where pn represents the obstacle point scanned by the laser;

[0087] Step 2: Traverse P in order and calculate the distance between two adjacent obstacle points. When the distance is greater than the threshold, such as 5cm, segment it; otherwise, group it. That is, group the obstacle points in P to obtain a set of several obstacle points L, denoted as P = {L1, L2, ..., Lm}, where L = {pi, ..., pj}.

[0088] Step 3: For each set of obstacle points L, calculate the distance between each pair of points within it, and take the maximum distance dm. If the distance dm is greater than the threshold h1 (which can be 25cm) or less than the threshold h2 (which can be 15cm), then discard the set of obstacle points; otherwise, keep it.

[0089] Step 4: For each retained set of obstacle points L, use the Hough transform line detection method to extract the two lines with the largest response and calculate the angle α between them. If two lines cannot be extracted or the angle α does not meet the condition |α-120°|>10°, then discard the set of obstacle points L.

[0090] Step 5: For one of the retained obstacle point sets L, extract its two fitted lines, denoted as a and b respectively, as follows: Figure 6 As shown;

[0091] Calculate the intersection point of lines a and b, denoted as C; and let the endpoints of the two sides of the obstacle point set L be A and B respectively.

[0092] Take the larger of the lengths of line segments CA and CB, and denote it as R;

[0093] Assuming CA is longer, find another point B' on line b such that |CB'| = R;

[0094] Draw circles with radius R centered at points A and B'. The intersection of these two circles is the center point of the positioning rod, denoted as O. The actual calculation is as follows:

[0095] x O =x A +x B′ -x c

[0096] y O =y A +y B′ -y c

[0097] Among them, (x A y A ), (x B’ y B’ ), (x c y c ), (x O y O ) are the coordinates of points A, B', C, and center point O, respectively.

[0098] By processing all retained obstacle points in step five, the positions O of multiple positioning rods detected by the laser scan data P in the current frame can be obtained. n (Relative to the mobile robot's coordinate system), and the corresponding cross-sectional radius R can also be obtained. n .

[0099] Using the above-described positioning rod detection and positioning algorithm, the positions of all positioning rods in each frame of laser scanning data and the radius of the cross-section of the positioning rod at the corresponding scanning height can be obtained. That is, in the i-th frame of data, the positions of n positioning rods are obtained, denoted as the set of positioning rod position points. (Relative to the mobile robot's coordinate system), similarly, in the (i+1)th frame of data, the positions of k positioning rods are obtained and denoted as the set of positioning rod position points.

[0100] For the set of positioning rod position points in two adjacent frames of laser scanning data, such as the set of positioning rod position points in the i-th frame... and the set of positioning rod positions in frame i+1 A nearest-point iterative algorithm is employed to quickly align the same positions of the positioning rod sets. Simultaneously, the ICP matching algorithm or a general equation is used to solve for the two-dimensional planar position and attitude changes (Δx) of the mobile robot when measuring these two sets of positioning rod positions. i+1 ,Δy i+1 ,Δθ i+1 );

[0101] Based on the position of the mobile robot in the i-th frame and the changes in its two-dimensional planar position and posture, the position of the mobile robot in the (i+1)-th frame is calculated. Since the initial position upon entering the slope is known, the planar projection position and orientation angle of the mobile robot on the slope for each frame of laser scanning data can be obtained. That is, the planar projection position and orientation angle of the mobile robot in the i-th frame are denoted as (x...). i ,y i ,θ i );

[0102] Further, find the positioning rod closest to the mobile robot in the i-th frame of laser scan data, and obtain the corresponding cross-sectional radius of the positioning rod, denoted as Ri. Therefore, the current height of the laser on the mobile robot can be obtained as z. i The specific calculations are as follows:

[0103]

[0104] Where f represents the number of times the positioning rod is traversed, h represents the length of the positioning rod, the plus sign indicates that the mobile robot moves up the slope, and the minus sign indicates that it moves down the slope. The number of times f can be determined by detecting R. i The jump from the bottom to the top of the positioning rod is achieved, such as a jump from 2cm to 22cm.

[0105] Further, the 3D spatial position of the mobile robot on the slope in the i-th frame of laser scanning data is obtained, denoted as (x i ,y i ,z i ,θ i ).

[0106] During a single uphill or downhill climb by the mobile robot, the laser points of the positioning poles are removed from each frame of the robot's 3D spatial position obtained above. This data is then accumulated and projected onto the map coordinate system to obtain a slope point cloud map located in the same coordinate system as the planar grid map. It should be noted that due to the varying heights of consecutive frames and the horizontal orientation of the laser measurements, this slope point cloud map is derived from the accumulation of horizontally layered laser data, and therefore also consists of layered point clouds.

[0107] S18. All the planar raster maps and slope point cloud maps are stitched together and merged to form the composite map.

[0108] Using a mature map stitching and fusion algorithm, all the planar raster maps and slope point cloud maps obtained in the above steps are stitched and fused to form a composite map.

[0109] This embodiment installs low-cost positioning rods on various slopes in the application scenario. Mature and low-cost 2D laser technology is used to scan the application scenario, constructing a planar grid map for the plane. Based on the positioning rods, a slope point cloud map is constructed for the slopes. All the planar grid maps and slope point cloud maps are then stitched and merged to form the composite map. This effectively solves the problem that existing 2D laser technology cannot construct maps for non-planar application scenarios. It also addresses the issues of low accuracy, high cost, and high system complexity associated with using 3D laser technology. In short, this embodiment combines simple positioning rods with mature and low-cost 2D laser technology to achieve high-precision composite map construction for non-planar application scenarios, and offers the advantages of system simplicity and low cost.

[0110] Example 2

[0111] This embodiment is basically the same as embodiment 1, except that the composite map construction method based on laser technology also includes the following steps:

[0112] When a mobile robot leaves one area and enters another, it saves the map information of the area it left, which includes the map of the area it left, the location of the boundary between the two areas, and the posture of the mobile robot at this time.

[0113] At this point, step S18 stitches and merges all the planar raster maps and slope point cloud maps to form a composite map:

[0114] Based on the map information, all the planar raster maps and slope point cloud maps are stitched together and merged to form a composite map.

[0115] In this embodiment, when the mobile robot leaves one area and enters another, it saves the map information of the area it leaves. The map information includes the map of the area it leaves, the location of the boundary between the two areas, and the posture of the mobile robot at this time. Based on the map information, the relationship between the stitching and fusion of the various maps can be found efficiently, making the map at the boundary between the areas clearer. This makes it faster, more accurate, and more precise to stitch and fuse all the planar grid maps and slope point cloud maps into a composite map.

[0116] Example 3

[0117] Figure 7 This diagram illustrates the structure of the laser-based composite map construction system of the present invention. Figure 7 A composite map building system based on laser technology is known to include: a mobile robot 1 and a positioning rod 2; the mobile robot 1 includes a laser scanning module 11, a region judgment and construction module 12 and a fusion map module 13.

[0118] The positioning rod 2 is used for laser positioning and is installed on various slopes in the application scenario.

[0119] Application scenarios of the present invention are as follows Figure 3a As shown, it has two different height planes connected by a ramp, or has multiple ramps connecting two different height planes (not shown in the figure), or as... Figure 3b As shown, it has three planes at different heights connected by two ramps. This embodiment uses... Figure 3a The application scenarios shown illustrate the invention; other application scenarios are basically the same and will not be described in detail here.

[0120] The positioning rod 2 of this invention is preferably a vertical rod of a specific shape, such as a conical vertical rod with a regular hexagonal cross-section, preferably 50cm in length, with a top hexagonal side length of 2cm and a bottom hexagonal side length of 22cm. Figure 4 As shown. The positioning rod 2 is made of laser-reflective material or its surface is covered with a layer of laser-reflective material to save costs. It is used to install on or near various slopes in the application scenario to clearly reflect the laser and achieve accurate distance measurement. Preferably, as Figure 5 As shown, the application scenario consists of two planes 3 connected by a ramp 4. Positioning rods 2 are installed on the sides and front and back of the ramp 4 in a vertical and horizontal manner. Furthermore, the positioning rods 2 are sequentially installed on the sides of each ramp 4 in the application scenario, starting from the bottom of the ramp 4 and arranged vertically and horizontally, with the top of the preceding positioning rod and the bottom of the following positioning rod on the same horizontal plane in each pair of adjacent positioning rods. It is preferable that the distance between the positioning rods and surrounding obstacles is greater than 10cm.

[0121] Using a tapered vertical rod with a regular hexagonal cross-section for constructing the slope point cloud map allows the mobile robot to receive more and more regular laser reflection points from the positioning rod, making the positioning rod more accurate. This results in higher accuracy in constructing the slope point cloud map and faster computation speed.

[0122] Application scenarios of the present invention are as follows Figure 3a As shown, it has two different height planes connected by a ramp, or has multiple ramps connecting two different height planes (not shown in the figure), or as... Figure 3b As shown, it has three planes at different heights connected by two ramps. This embodiment uses... Figure 3a The application scenarios shown illustrate the invention; other application scenarios are basically the same and will not be described in detail here.

[0123] The positioning rod of this invention is preferably a vertical rod of a specific shape, such as a conical vertical rod with a regular hexagonal cross-section, preferably 50cm in length, with a top hexagonal side length of 2cm and a bottom hexagonal side length of 22cm. Figure 4 As shown. The positioning rod is made of laser-reflective material or its surface is covered with a layer of laser-reflective material to save costs. It is used to install on or near various slopes in the application scenario to clearly reflect the laser light for accurate distance measurement. Preferably, as... Figure 5 As shown, the application scenario consists of two planes 3 connected by a ramp 4. Positioning rods 2 are installed on the sides and front and back of the ramp 4 in a vertical and horizontal manner. Furthermore, the positioning rods 2 are sequentially installed on the sides of each ramp 4 in the application scenario, starting from the bottom of the ramp 4 and arranged vertically and horizontally, with the top of the preceding positioning rod and the bottom of the following positioning rod on the same horizontal plane in each pair of adjacent positioning rods. It is preferable that the distance between the positioning rods and surrounding obstacles is greater than 10cm.

[0124] Using a tapered vertical rod with a regular hexagonal cross-section for constructing the slope point cloud map allows the mobile robot to receive more and more regular laser reflection points from the positioning rod, making the positioning rod more accurate. This results in higher accuracy in constructing the slope point cloud map and faster computation speed.

[0125] Positioning rods are installed sequentially on the sides of each slope in the application scenario, starting from the bottom of the slope and arranged vertically and horizontally. In each pair of adjacent positioning rods, the top of the previous positioning rod and the bottom of the next positioning rod are located on the same horizontal plane. This reduces the computational complexity of the mobile robot's positioning and makes the construction of the slope point cloud map faster and more efficient.

[0126] The laser scanning module 11 uses 2D laser technology to scan the application scene.

[0127] In this embodiment, the laser scanning module 11 of the mobile robot 1 uses existing 2D laser technology to scan the application scene, such as... Figure 5 As shown, when the mobile robot 1 moves to the intersection of the plane 3 and the ramp 4, it will use the laser scanning module 11 to scan the positioning rod 2.

[0128] The region determination and construction module 12 is used to determine the region where the mobile robot 1 is located. When the mobile robot 1 is located on a plane, a planar grid map is constructed on the plane, and the planar grid map includes height information. When the mobile robot 1 is located on a slope 4, a slope point cloud map is constructed on the slope based on the positioning rod 2.

[0129] The laser scanning module 11 of the mobile robot 1 uses 2D laser technology to scan the application scene. The area judgment and construction module 12 receives instructions from human operators and determines whether the area where the mobile robot 1 is located is a plane or a slope. Preferably, the mobile robot 1 also includes an inertial measurement sensor, which measures the acceleration, pitch angle, and roll angle of the mobile robot 1. Based on the inertial measurement sensor data, the area judgment and construction module 12 determines the area where the mobile robot 1 is located, reducing the degree of human intervention in area judgment and system operation, making the composite map construction method and system operation simpler, and reducing the labor cost of composite map construction. The specific area condition judgment is as follows:

[0130] Condition 1: The absolute value of the acceleration of the mobile robot in the Z-axis / height direction is less than a predetermined value, for example, 9 m / s². 2 And the duration is greater than the predetermined value, for example, 1 second; the absolute values ​​of the mobile robot's pitch angle and roll angle are greater than the predetermined value, for example, 10 degrees, and the duration is greater than the predetermined value, for example, 1 second;

[0131] Condition 2: The absolute value of the acceleration of the mobile robot in the Z-axis / height direction is greater than a predetermined value, for example, 9.3 m / s². 2 The sum of the absolute values ​​of the pitch and roll angles of the mobile robot is less than the predetermined value, for example, 10 degrees, and the duration is greater than the predetermined value, for example, 1 second.

[0132] When condition 1 is met, it is determined that the mobile robot has entered the ramp. After entering the ramp, the mobile robot meets condition 2, and it is determined that the mobile robot has left / exited the ramp.

[0133] When the mobile robot 1 is located on a plane, a planar grid map is constructed on the plane using 2D laser technology. For example, mature algorithms such as Cartographer, CoreSLAM, and Karto-SLAM are used to construct the planar grid map on the plane. The planar grid map constructed by this invention contains height information to distinguish planar areas at different heights.

[0134] Preferably, the mobile robot 1 further includes an encoder, which is used to calculate the odometer of the mobile robot 1, and based on the odometer, uses 2D laser technology to construct a planar grid map of the plane, so that the constructed application scene map has higher accuracy.

[0135] When the mobile robot 1 is located on a slope, a point cloud map of the slope is constructed based on the positioning rod 2. The specific implementation method is as follows:

[0136] Mobile robot 1 uses 2D laser technology to scan and acquire laser scanning data. Based on the scanning data, a positioning rod detection and positioning algorithm is used to obtain the position of the positioning rod and the radius of the positioning rod cross-section at the corresponding scanning height.

[0137] When the mobile robot 1 moves to the bottom or top of the slope and enters the slope, it saves the current planar grid map and the initial position upon entering the slope. The mobile robot 1 moves along the slope while simultaneously performing laser scanning, acquiring laser scan data for each frame. Preferably, while moving along the slope, the laser scanning module 11 of the mobile robot 1 adjusts its laser scanning direction to keep it parallel to the horizontal plane at all times. This reduces the complexity of the positioning rod detection and positioning algorithm, thereby making the application of the composite map construction method in this embodiment more efficient and simpler.

[0138] The positioning rod detection and positioning algorithm is as follows:

[0139] Step 1: Obtain the current frame laser scan data P = {p1, p2, ..., pn}, where pn represents the obstacle point scanned by the laser;

[0140] Step 2: Traverse P in order and calculate the distance between two adjacent obstacle points. When the distance is greater than the threshold, such as 5cm, segment it; otherwise, group it. That is, group the obstacle points in P to obtain a set of several obstacle points L, denoted as P = {L1, L2, ..., Lm}, where L = {pi, ..., pj}.

[0141] Step 3: For each set of obstacle points L, calculate the distance between each pair of points within it, and take the maximum distance dm. If the distance dm is greater than the threshold h1 (which can be 25cm) or less than the threshold h2 (which can be 15cm), then discard the set of obstacle points; otherwise, keep it.

[0142] Step 4: For each retained set of obstacle points L, use the Hough transform line detection method to extract the two lines with the largest response and calculate the angle α between them. If two lines cannot be extracted or the angle α does not meet the condition |α-120°|>10°, then discard the set of obstacle points L.

[0143] Step 5: For one of the retained obstacle point sets L, extract its two fitted lines, denoted as a and b respectively, as follows: Figure 6 As shown;

[0144] Calculate the intersection point of lines a and b, denoted as C; and let the endpoints of the two sides of the obstacle point set L be A and B respectively.

[0145] Take the larger of the lengths of line segments CA and CB, and denote it as R;

[0146] Assuming CA is longer, find another point B' on line b such that |CB'| = R;

[0147] Draw circles with radius R centered at points A and B'. The intersection of these two circles is the center point of the positioning rod, denoted as O. The actual calculation is as follows:

[0148] x O =x A +x B′ -x c

[0149] y O =y A +y B′ -y c

[0150] Among them, (x A y A ), (x B’ y B’ ), (x c y c ), (x O y O Points A and B are respectively.

[0151] The coordinates of point B', point C, and center point O.

[0152] By processing all retained obstacle points in step five, the positions O of multiple positioning rods detected by the laser scan data P in the current frame can be obtained. n (Relative to the mobile robot's coordinate system), and the corresponding cross-sectional radius R can also be obtained. n .

[0153] Using the above-described positioning rod detection and positioning algorithm, the positions of all positioning rods in each frame of laser scanning data and the radius of the cross-section of the positioning rod at the corresponding scanning height can be obtained. That is, in the i-th frame of data, the positions of n positioning rods are obtained, denoted as the set of positioning rod position points. (Relative to the mobile robot's coordinate system), similarly, in the (i+1)th frame of data, the positions of k positioning rods are obtained and denoted as the set of positioning rod position points.

[0154] For the set of positioning rod position points in two adjacent frames of laser scanning data, such as the set of positioning rod position points in the i-th frame... and the set of positioning rod positions in frame i+1 A nearest-point iterative algorithm is employed to quickly align the same positions of the positioning rod sets. Simultaneously, the ICP matching algorithm or a general equation is used to solve for the two-dimensional planar position and attitude changes (Δx) of the mobile robot when measuring these two sets of positioning rod positions. i+1 ,Δy i+1,Δθ i+1 );

[0155] Based on the position of the mobile robot in the i-th frame and the changes in its two-dimensional planar position and posture, the position of the mobile robot in the (i+1)-th frame is calculated. Since the initial position upon entering the slope is known, the planar projection position and orientation angle of the mobile robot on the slope for each frame of laser scanning data can be obtained. That is, the planar projection position and orientation angle of the mobile robot in the i-th frame are denoted as (x...). i ,y i ,θ i );

[0156] Further, find the positioning rod closest to the mobile robot in the i-th frame of laser scan data, and obtain the corresponding cross-sectional radius of the positioning rod, denoted as Ri. Therefore, the current height of the laser on the mobile robot can be obtained as z. i The specific calculations are as follows:

[0157]

[0158] Where f represents the number of times the positioning rod is traversed, h represents the length of the positioning rod, the plus sign indicates that the mobile robot moves up the slope, and the minus sign indicates that it moves down the slope. The number of times f can be determined by detecting R. i The jump from the bottom to the top of the positioning rod is achieved, such as a jump from 2cm to 22cm.

[0159] Further, the 3D spatial position of the mobile robot on the slope in the i-th frame of laser scanning data is obtained, denoted as (x i ,y i ,z i ,θ i ).

[0160] During a single uphill or continuous downhill climb of mobile robot 1, the laser points of the positioning poles are removed from each frame of the 3D spatial position obtained from the robot's 3D position in each frame. This data is then accumulated and projected onto the map coordinate system to obtain a slope point cloud map that is in the same coordinate system as the planar grid map. It should be noted that due to the varying heights of consecutive frames and the horizontal orientation of the laser measurements, this slope point cloud map is derived from the accumulation of horizontally layered laser data, and therefore also consists of layered point clouds.

[0161] The fusion map module 13 is used to stitch together and merge all the planar raster maps and slope point cloud maps to form a composite map.

[0162] The fused map module 13 uses a mature map stitching and fusion algorithm to stitch and fuse all the planar raster maps and slope point cloud maps obtained above into a composite map.

[0163] This embodiment uses low-cost positioning rods installed on various slopes in the application scenario. Mature and low-cost 2D laser technology is used to scan the application scenario, constructing a planar grid map for the plane. Based on the positioning rods, a slope point cloud map is constructed for the slope. All the planar grid maps and slope point cloud maps are then stitched and merged to form the composite map. This effectively solves the problem that existing 2D laser technology cannot construct maps for non-planar application scenarios, and also addresses the issues of low accuracy, high cost, and high system complexity associated with using 3D laser technology. In short, this embodiment combines simple positioning rods with mature and low-cost 2D laser technology to achieve high-precision composite map construction for non-planar application scenarios, and has the advantages of system simplicity and low cost.

[0164] In another embodiment of the present invention, the mobile robot 1 further includes: a map information storage module;

[0165] When a mobile robot leaves one area and enters another, the map information storage module stores the map information of the area it leaves. The map information includes the map of the area it leaves, the location of the boundary between the two areas, and the posture of the mobile robot at this time.

[0166] The fusion map module 13, based on the map information, stitches and merges all the planar raster maps and slope point cloud maps to form a composite map.

[0167] In this embodiment, when the mobile robot leaves one area and enters another, the map information storage module stores the map information of the area it leaves. The map information includes the map of the area it leaves, the location of the boundary between the two areas, and the posture of the mobile robot at this time. Based on the map information, the relationship between the stitching and fusion of each map can be found efficiently, making the map at the boundary between each area clearer. This makes it faster, more accurate, and more precise to stitch and fuse all the planar grid maps and slope point cloud maps into a composite map.

[0168] Example 4

[0169] The laser-based composite map construction system in this embodiment is basically the same as that in Embodiment 3, consisting of... Figure 7 , 8As shown in section 9, the mobile robot 1 in this embodiment is a traditional mobile robot, supporting movement on flat surfaces and slopes, and possessing the ability to climb slopes. The laser scanning module 11 includes a lidar gimbal 111, which is fixed to the top of the mobile robot 1. The lidar gimbal 111 includes a lidar 112, a planar platform 113, a gimbal IMU and controller 114, a servo motor 115, and a fixed bracket 116. The gimbal IMU and controller 114 can acquire the pitch and roll angles of the planar platform it is fixedly connected to in real time, and based on the pitch and roll angles, adjust the planar platform 113 in real time through the servo motor 115 to ensure that it is parallel to the horizontal plane, thereby ensuring that the laser scanning line is always parallel to the horizontal plane.

[0170] In this embodiment, the mobile robot 1 moves on a slope, and its laser scanning line can still be kept horizontal, thereby ensuring the reliable operation of the system of the present invention. In order to ensure the reliability of the system in this embodiment, when the mobile robot 1 moves on a horizontal plane, the servo motor 115 of the lidar gimbal 111 is locked in the 0 position, the same as the conventional mobile robot; the servo motor 115 only starts to work when it detects that the mobile robot 1 is moving on a slope.

[0171] To ensure that the laser scanning line is not obstructed by the body of the mobile robot 1, the lidar gimbal 111 is mounted at a relatively high position on the top center of the mobile robot. Other obstacle sensors, such as ultrasonic, infrared, depth cameras, and lidar, can be installed on other lower positions of the mobile robot 1 to detect and avoid lower obstacles.

[0172] In this embodiment, during the real-time operation of the composite map construction system based on laser technology, the mobile robot 1 can achieve real-time positioning in the application scenario based on the constructed composite map.

[0173] When the mobile robot 1 is running on a plane, traditional laser map matching methods, such as AMCL positioning and Cartographer positioning, are directly used to achieve real-time tracking and positioning of the mobile robot in the corresponding plane area.

[0174] When the mobile robot 1 is detected to have moved onto a slope, the servo motor 115 of the LiDAR gimbal 111 of the mobile robot 1 starts to ensure that the laser scanning of the mobile robot 1 is horizontal, and at the same time, it enters the slope positioning mode. Its operation is as follows:

[0175] The planar positioning is paused, and the initial position (X,Y,Z,A) of the current mobile robot 1 on the planar map is recorded.

[0176] Starting from the initial position (X,Y,Z,A) on the plane, set it as the initial position of the mobile robot on the slope, and start the slope positioning mode;

[0177] Considering the scenario where the mobile robot continuously ascends and descends a slope, the current laser scan data is used as input, and the previous moment's position of the mobile robot is used as the starting point. In the slope point cloud map, the nearest matching pose of the mobile robot is searched for, which serves as the real-time localization of the mobile robot.

[0178] In this search and matching process, only small changes are considered. Within a small range, a particle filtering method is used to sample and update the position and pose of the mobile robot. The matching error is calculated to achieve localization. Specifically, the method for calculating the matching error of the particle at the i-th position is as follows:

[0179]

[0180] Where N represents the number of matching points, s t Let T(s) represent the t-th laser scanning point. t |(x i ,y i ,z i ,θ i )) indicates that s t According to (x) i ,y i ,z i ,θ i Project the position and attitude transformation points, p t This represents the point in the slope point cloud map that matches the projected point.

[0181] The larger the matching error, the smaller the weight of the corresponding robot positioning particle will be. After multiple iterations, the particle swarm can converge to the accurate position of the mobile robot.

[0182] When the mobile robot finishes traversing the ramp, it exits the ramp localization model. At the same time, the robot's current position is used as the initial position for the next segment of planar localization, and planar localization continues.

[0183] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A method for constructing composite maps based on laser technology, characterized in that, Includes the following steps: The positioning rods are installed on various slopes in the application scenario; the positioning rods are installed on various slopes in the application scenario in a vertical and horizontal manner, starting from the bottom of the slope, and the top of the previous positioning rod and the bottom of the next positioning rod are located on the same horizontal plane in each pair of adjacent positioning rods. Using 2D laser technology to scan the application scenario; The location of the mobile robot is determined. When the mobile robot is on a plane, a planar grid map is constructed on the plane, which includes height information. When the mobile robot is on a slope, a slope point cloud map is constructed based on the positioning rod. The composite map is formed by stitching together and merging all the planar raster maps and slope point cloud maps; During the construction of the composite map, the laser scanning direction is adjusted to keep it parallel to the horizontal plane at all times.

2. The method for constructing composite maps based on laser technology according to claim 1, characterized in that, It also includes the following steps: When a mobile robot leaves one area and enters another, it saves the map information of the area it left, which includes the map of the area it left, the location of the boundary between the two areas, and the posture of the mobile robot at this time. The composite map is created by stitching together all the planar raster maps and slope point cloud maps: Based on the map information, all the planar raster maps and slope point cloud maps are stitched together and merged to form a composite map.

3. The method for constructing composite maps based on laser technology according to claim 1, characterized in that, The construction of a planar raster map of the plane includes: Use an encoder to calculate the odometry of a mobile robot; Based on the odometer, a planar grid map is constructed for the plane in which it is located.

4. The method for constructing composite maps based on laser technology according to claim 1, characterized in that, The determination of the area where the mobile robot is located includes: The acceleration, pitch angle, and roll angle of the mobile robot were measured using inertial measurement sensors. The location of the mobile robot is determined based on data from inertial measurement sensors.

5. The method for constructing composite maps based on laser technology according to claim 1, characterized in that, The positioning rod is a tapered vertical rod with a regular hexagonal cross-section.

6. A composite map construction system based on laser technology, characterized in that, include: Positioning rods and mobile robots; The positioning rod is used for reflective laser positioning and is installed on various slopes in the application scenario. The installation of the positioning rod on various slopes in the application scenario is as follows: the positioning rod is installed sequentially on the side of each slope in the application scenario from the bottom of the slope in a vertical and horizontal manner, and in each pair of adjacent positioning rods, the top of the previous positioning rod and the bottom of the next positioning rod are located on the same horizontal plane. The mobile robot includes a laser scanning module, a region determination and construction module, and a fused map module; The laser scanning module uses 2D laser technology to scan the application scene; The region determination and construction module is used to determine the region where the mobile robot is located. When the mobile robot is on a plane, a planar grid map is constructed on the plane, and the planar grid map includes height information. When the mobile robot is on a slope, a slope point cloud map is constructed on the slope based on the positioning rod. The fusion map module is used to stitch together and merge all planar raster maps and slope point cloud maps to form a composite map. During the operation of the composite map construction system, the laser scanning module adjusts the laser scanning direction to keep it parallel to the horizontal plane at all times.

7. The composite map construction system based on laser technology according to claim 6, characterized in that, The mobile robot also includes: a map information storage module; When a mobile robot leaves one area and enters another, the map information storage module stores the map information of the area it leaves. The map information includes the map of the area it leaves, the location of the boundary between the two areas, and the posture of the mobile robot at this time. The fusion map module, based on the map information, stitches and merges all the planar raster maps and slope point cloud maps to form a composite map.

8. The composite map construction system based on laser technology according to claim 6, characterized in that, The mobile robot also includes: an encoder; The encoder is used to calculate the mobile robot's odometer reading; The mobile robot constructs a planar grid map of the plane it is on based on the odometer.

9. The composite map construction system based on laser technology according to claim 6, characterized in that, The mobile robot also includes inertial measurement sensors; The inertial measurement sensor is used to measure the acceleration, pitch angle, and roll angle of the mobile robot. The region determination module is used to determine the region where the mobile robot is located based on the inertial measurement sensor data.

10. The composite map construction system based on laser technology according to claim 6, characterized in that, The positioning rod is a tapered vertical rod with a regular hexagonal cross-section.

Images