WO2022116657A1

WO2022116657A1 - Fused positioning method and mobile robot

Info

Publication number: WO2022116657A1
Application number: PCT/CN2021/120080
Authority: WO
Inventors: 赖钦伟; 肖刚军; 戴剑锋
Original assignee: 珠海一微半导体股份有限公司
Priority date: 2020-12-01
Filing date: 2021-09-24
Publication date: 2022-06-09
Also published as: CN112612037A; CN112612037B

Abstract

A fused positioning method and a mobile robot. The fused positioning method comprises: controlling a camera of a mobile robot to acquire image information for visual positioning; controlling a fan-shaped distance sensor (103) of the mobile robot to acquire distance information from the inside of a spherical pyramidal sensing area at a front end thereof (S501), and then supplementarily fusing a visual positioning result of the image information acquired by the camera and the distance information acquired by the fan-shaped distance sensor (103) to reduce positioning blind areas; wherein the fan-shaped distance sensor (103) is mounted at the front end of the mobile robot, and the camera is also mounted on the surface of the mobile robot, but the probe orientation of the fan-shaped distance sensor (103) is different from the orientation of the camera. The distance information covered by the non-rotary fan-shaped distance sensor (103) is used to be fused with conventional visual image positioning information, to compensate for positioning blind areas for each other on the same map, thereby describing the complete environment information.

Description

A fusion positioning method and mobile robot

technical field

The invention relates to the technical field of mobile robot navigation and positioning, in particular to a positioning method based on a sector scanning area and a mobile robot.

Background technique

Mobile robots with automatic actions have developed rapidly in recent years, such as common household cleaning sweepers. At present, the common SLAM technologies include visual navigation, laser navigation, inertial navigation and so on. Among them, the user experience of laser navigation is better, mainly because it can scan the outline of the room in advance, so that it can be displayed in the user's map interface for navigation and positioning, which is more intuitive, but the laser radar has assembly shortcomings, mainly requiring a rotating The laser mechanism is hollowed out or raised on the mold of the machine, thereby increasing the production cost; on the other hand, the visual navigation is affected by the installation position of the camera and is prone to blind spots of viewing angles, and is also greatly affected by the lighting environment, and the application scenarios are relatively limited. .

technical solutions

In order to overcome the above-mentioned problems of lidar opening cost and blind area of visual angle, the present invention adopts a non-rotating fan-shaped distance sensor and integrates traditional visual positioning technology to realize accurate positioning and navigation. The specific technical solution is as follows: control the camera of the mobile robot to collect image information for visual positioning, and control the fan-shaped distance sensor of the mobile robot to obtain distance information from the inside of the spherical pyramid-type sensing area at the front end; then use the image information collected by the camera The visual positioning results obtained from the sensor and the distance information obtained by the fan-shaped distance sensor are complemented and fused with each other, and then the complementary and fused positioning results are marked on the map; among them, the fan-shaped distance sensor is installed at the front end of the mobile robot, and the camera is also installed on the surface of the mobile robot. The probe orientation of the sector distance sensor is different from the orientation of the camera. The technical solution uses the distance information covered by the non-rotating sector-shaped distance sensor to fuse the traditional visual image positioning information to compensate for the positioning blind spot on the same map, so as to describe the complete environmental information.

Further, the method for controlling the fan-shaped distance sensor of the mobile robot to obtain distance information from the interior of the spherical pyramid-type sensing area at the front end thereof comprises: controlling the fan-shaped distance sensor to modulate and emit a spherical pyramid-shaped optical signal, wherein this The largest effective ranging area of the spherical pyramid optical signal is the spherical pyramid sensing area; when the sector distance sensor receives the feedback optical signal reflected from the obstacle in the spherical pyramid sensing area, according to the The time of flight recorded by the feedback optical signal is received to calculate and obtain the distance information of the position of the corresponding obstacle relative to the fan-shaped distance sensor. The technical solution can be regarded as controlling the fan-shaped distance sensor to transmit the spherical pyramid optical signal at the same position and receive the spherical pyramid optical signal (feedback optical signal) reflected by the obstacle, and then use the time of flight of the spherical pyramid optical signal to Calculate and determine the distance information of the obstacle relative to the position, and then determine the position of the obstacle.

Further, the distance of the three-dimensional point cloud inside the spherical pyramid sensing area relative to the fan-shaped distance sensor is all within the effective ranging range of the fan-shaped distance sensor; wherein, the effective range of the fan-shaped distance sensor is The maximum value of the ranging range is the maximum effective ranging distance; wherein, the projection of the spherical pyramid optical signal on the traveling plane of the mobile robot is a horizontal fan-shaped area, and this horizontal fan-shaped area is the installation position of the fan-shaped distance sensor is the vertex, and the maximum effective ranging distance is the sector of the radius. It is beneficial to detect obstacles within a certain angle range in front of the body.

Further, the spherical pyramid optical signal has a horizontal viewing angle on the traveling plane of the mobile robot, and the spherical pyramid optical signal has a vertical viewing angle in the vertical direction of the traveling plane of the mobile robot, wherein the horizontal viewing angle is greater than Vertical viewing angle. The technical solution can control the fan-shaped distance sensor to measure a fan-shaped angle on the traveling plane of the mobile robot, such as 120 degrees, and measure a relatively narrow viewing angle range, such as 10 degrees, in the vertical direction of the traveling plane of the mobile robot, so The travel plane of the mobile robot is a fan-shaped light with a certain height perpendicular to the travel plane of the mobile robot. Compared with the scanning method of the emitting surface light source, it is better to realize the uniformity of illumination coverage of the small area detection light signal. Thereby, the narrow 3D point cloud information in the area covered by the three-dimensional space similar to the spherical quadrangular pyramid is measured.

Further, the method for complementing and fusing the visual positioning result of the image information collected by the camera and the distance information obtained by the fan-shaped distance sensor to reduce the positioning blind area includes: when it is detected that the image information currently collected by the camera does not completely cover the local positioning area. , using the distance information of the three-dimensional point cloud inside a spherical pyramid-type sensing area formed by the current emission of the fan-shaped distance sensor to complete the supplementation of the pose information of the local positioning area on the same map; The current viewing angle range of the camera of the mobile robot and the overlapping area of a spherical pyramid-shaped sensing area formed by the current emission of the fan-shaped distance sensor. The technical solution uses the three-dimensional point cloud information in the spherical pyramid-type sensing area formed by the fan-shaped distance sensor to construct the regional position information of a small area at a close distance, and overcomes the camera's inability to locate the local area due to the influence of the lighting environment and the installation position. technical defects.

Further, the method for complementing and fusing the visual positioning result of the image information collected by the camera and the distance information obtained by the fan-shaped distance sensor to reduce the positioning blind spot includes: when detecting a spherical pyramid type formed by the current emission of the fan-shaped distance sensor. When the distance information of the 3D point cloud inside the sensing area does not completely cover the local positioning area, the image information currently collected by the camera is used to complete the supplementation of the pose information of the local positioning area on the same map; The current viewing angle range of the camera of the mobile robot and the overlapping area of a spherical pyramid-shaped sensing area formed by the current emission of the fan-shaped distance sensor. The technical solution solves the problem that the effective ranging distance of the fan-shaped distance sensor is insufficient to detect the position through visual positioning.

Further, the fan-shaped distance sensor is used to modulate and generate at least one of the spherical pyramid optical signals or other types of modulation signals, but only allows to control the emission of one of the spherical pyramid optical signals for distance measurement; wherein, The sector distance sensor is a 3d-tof sensor. Compared with the emission surface light source, the technical solution reduces the emission power of the sensor, and is suitable for uniformly detecting the distance information of obstacles in a small area.

Further, within the effective ranging range of the fan-shaped distance sensor, the current contour is scanned and marked using the distance information of a three-dimensional point cloud inside the spherical pyramid-type sensing area currently obtained by the fan-shaped distance sensor, Then judge whether the current contour and the historical contour in the same area of the pre-stored historical map library meet the preset coincidence degree; if the current contour and the historical contour do not meet the preset coincidence degree, then according to the pose Rotation-translation transformation is performed on the current contour, so as to correct the current contour after the rotation-translation transformation to conform to the preset coincidence degree with the historical contour. So as to achieve accurate matching of contour boundaries.

A mobile robot includes a camera, a fan-shaped distance sensor and a processing unit; the fan-shaped distance sensor is installed on the front end of the mobile robot; the camera is also installed on the surface of the mobile robot, but the orientation of the camera is different from that of the probe of the fan-shaped distance sensor, which makes the camera's angle of view The coverage range is not exactly the same as the effective ranging range of the fan-shaped distance sensor; the processing unit is used to perform the aforementioned fusion positioning method. The technical solution uses a non-rotating fan-shaped distance sensor combined with different cameras to complete fusion positioning, so as to achieve the technical effect of accurately marking the map boundary.

Further, the fan-shaped distance sensor is a 3d-tof sensor, and the projection of the spherical pyramid-shaped light signal emitted by the 3d-tof sensor on the traveling plane of the mobile robot is a horizontal fan-shaped area with a horizontal viewing angle of 120 degrees. The spherical pyramid optical signal emitted by the -tof sensor has a vertical viewing angle of 10 degrees in the vertical direction of the traveling plane of the mobile robot. In order to obtain the information of the effective ranging and positioning position in the three-dimensional space similar to the spherical quadrangular pyramid.

Further, the installation position of the fan-shaped distance sensor passes through the central axis of the body of the mobile robot to ensure that the fan-shaped distance sensor is directly in front of the body; or the installation position of the fan-shaped distance sensor is in a predetermined position with the body central axis of the mobile robot. Set the angle. Therefore, it is ensured that the central axis of the body of the mobile robot passes through the effective ranging area of the spherical pyramid optical signal emitted by the fan-shaped distance sensor.

Description of drawings

FIG. 1 is a schematic structural diagram of a mobile robot disclosed in an embodiment of the present invention.

FIG. 2 is a schematic diagram of a grid map of a horizontal fan-shaped area without obstacles scanned by a mobile robot according to another embodiment of the present invention.

FIG. 3 is a schematic diagram of a grid map of a horizontal sector area with obstacles scanned by a mobile robot according to another embodiment of the present invention.

4 is a schematic diagram of an optical path of a spherical pyramid optical signal emitted by a fan-shaped distance sensor disclosed in another embodiment of the present invention.

FIG. 5 is a flowchart of a fusion positioning method disclosed by another embodiment of the present invention.

Embodiments of the present invention

The technical solutions in the embodiments of the present invention will be described in detail below with reference to the accompanying drawings in the embodiments of the present invention. Taking the application of a monocular camera as an example, when a monocular camera is used for visual navigation and mapping in an unknown environment, the mobile robot lacks the shape and size information of the plane of the work area where it is located. The reliability of size information is low, because there are large areas of the same color in the collected images that are similar to the colors of individual landmarks, so it is not easy for the monocular camera to distinguish the boundary of obstacles and landmarks or obstacles when collecting the boundary of walls or obstacles in the environment. Other acquisition target objects lead to low mapping efficiency of the mobile robot. On the other hand, taking the mobile robot navigation application where lidar scans the outline of the room as an example, the lidar can scan the outline of the room in advance and present it in the user display interface, which is more intuitive. However, it is necessary to set a rotating laser mechanism on the top surface of the mobile robot, and the mold needs to be hollowed out or raised, which increases the design cost, and the point cloud data obtained by one scan is too large, the scanning angle of view is too large, and the amount of data processed is large. , the mapping efficiency is not high. The embodiment of the present invention does not use lidar to obtain the outline of the room, but uses a fan-shaped distance sensor that is fixed on the body and does not rotate to scan and mark the boundary of a certain viewing angle range, and gradually mark the mark on the map during the walking process of the mobile robot. Room outline and distribution of obstacles in the room.

An embodiment of the present invention discloses a fusion positioning method, as shown in FIG. 5 , which specifically includes: step S501 , controlling a camera of a mobile robot to collect image information for visual positioning, and controlling a fan-shaped distance sensor of the mobile robot from a spherical pyramid at the front end of the mobile robot at the same time. The distance information is obtained from the inside of the sensor area (the maximum effective ranging area formed by the construction of a single spherical pyramid optical signal), and then the process goes to step S502. In this step S501, the body position or the landmark position can be calculated according to the landmark feature information in the image information, but due to the aforementioned defects of visual positioning, there is still a visual positioning blind spot in the visual positioning result in the positioning of some scenes.

Step S502, the visual positioning result of the image information collected by the camera and the distance information obtained by the fan-shaped distance sensor are complemented and fused with each other to reduce the positioning blind area, and then the positioning result of the complementary fusion is marked on the map; wherein, the fan-shaped distance sensor is installed on the mobile robot. At the front of the mobile robot, the camera is also installed on the surface of the mobile robot, but the direction of the probe of the fan-shaped distance sensor is different from that of the camera. In this step, the result of this supplementary fusion positioning includes the pose information of the effective boundary within the single spherical pyramid sensing area sent by the fan-shaped distance sensor at a specific position of the mobile robot, and also includes the positioning result of the image information collected by the camera; As the mobile robot walks along the preset planned path, there are more and more results of complementary fusion positioning. Location information of the current work area. It should be noted that the distance information obtained by the fan-shaped distance sensor inside the spherical pyramid-type sensing area is converted into the marked coordinate position on the map, and the image information collected by the camera facing different from the probe of the fan-shaped distance sensor is also used. Converted to the marked coordinate positions on the map, even if they do not belong to the part of the area that needs to be supplemented with fusion positioning due to blind spots, thereby increasing the integrity of the environmental information of the map markers.

Therefore, in this embodiment, the distance information covered by the non-rotating sector distance sensor is used to fuse the traditional visual image positioning information, and the positioning blind areas are mutually compensated on the same map, thereby describing the complete environmental information.

It is worth noting that TOF is the time of flight (Time of Flight) The abbreviation of technology, that is, the sensor emits a modulated light signal, which is reflected after encountering an object. The sensor converts the distance of the photographed scene by calculating the time difference or phase difference between the emission and reflection of the modulated light signal to generate depth information, that is, spherical pyramid light signal. The distance between the obstacles in the maximum effective ranging area and the fan-shaped distance sensor.

It should be noted that the camera used for visual positioning in this embodiment includes, but is not limited to, a monocular camera.

The method for controlling the fan-shaped distance sensor of the mobile robot to obtain distance information from the interior of the spherical pyramid-type sensing area at the front end thereof includes: firstly, controlling the fan-shaped distance sensor to emit a spherical pyramid-shaped light signal, wherein the spherical pyramid-shaped light signal is The maximum effective ranging area of the optical signal is the spherical pyramid sensing area, which is equivalent to the maximum effective ranging area of the spherical pyramid optical signal on the premise that there is no obstruction and reflection. When the fan-shaped distance sensor receives the feedback light signal reflected by the obstacle in the spherical pyramid sensing area, it calculates and obtains the three-dimensional point cloud of the corresponding obstacle according to the flight time recorded by receiving the feedback light signal. With respect to the distance information of the fan-shaped distance sensor, the corresponding obstacle is an obstacle that reflects the spherical pyramid-shaped optical signal to the fan-shaped distance sensor after the fan-shaped distance sensor emits the spherical pyramid-shaped optical signal; The reflection signal of the spherical pyramid optical signal through the obstacle is the feedback optical signal. The distance calculation method of this embodiment can be regarded as controlling the fan-shaped distance sensor to transmit the spherical pyramid optical signal at the same position and receive the spherical pyramid optical signal (feedback optical signal) reflected by the obstacle, and then pass the spherical pyramid optical signal through the spherical pyramid optical signal. The flight time to calculate and determine the distance information of the obstacle relative to the position, and then determine the position of the obstacle.

It should be noted that the distances of the three-dimensional point cloud inside the spherical pyramid sensing area relative to the fan-shaped distance sensor are all within the effective ranging range of the fan-shaped distance sensor; wherein, the fan-shaped distance sensor The maximum value of the effective ranging range is the maximum effective ranging distance, that is, the radius d of the sector-shaped area 104 in FIG. 1 .

As shown in FIG. 4 , under the environmental conditions that the obstacles in the maximum effective ranging area block the reflection interference, the spherical pyramid-shaped sensing area formed by the spherical pyramid optical signal coverage is a spherical pyramid-shaped three-dimensional Space, a strip of light emitted from the fan-shaped distance sensor 103 to the distance. The projection of the spherical pyramid optical signal on the traveling plane of the mobile robot is a horizontal fan-shaped area, and this horizontal fan-shaped area takes the installation position of the fan-shaped distance sensor 103 as the vertex and the maximum effective ranging distance d as the radius. sector. In this embodiment, the fan-shaped distance sensor 103 emits a strip-shaped optical path, and the strip-shaped optical path is: the travel plane of the mobile robot is fan-shaped, and is perpendicular to the travel plane of the mobile robot and has a certain vertical height. Compared with the surface light source scanning method, it is better to realize the uniformity of illumination coverage of the detection light signal in a small area. Ensure the accuracy of ranging and positioning.

When the fan-shaped distance sensor is a 3d-tof sensor, in the prior art, multiple beams of the spherical pyramid optical signal can be simultaneously emitted during the ranging process, but only one beam of the spherical pyramid optical signal is selected to be emitted in this embodiment. , as shown in FIG. 4 , the spherical pyramid optical signal has a horizontal viewing angle on the traveling plane of the mobile robot, and it can also be considered that the spherical pyramid sensing area has a horizontal sector area on the traveling plane of the mobile robot. A horizontal viewing angle, so as to control the fan-shaped distance sensor to measure a fan-shaped angle on the traveling plane of the mobile robot, such as 120 degrees; the spherical pyramid optical signal has a vertical viewing angle in the vertical direction of the traveling plane of the mobile robot, It can also be considered that the spherical pyramid sensing area has a certain height of viewing angle in the vertical direction of the traveling plane of the mobile robot, so that a relatively narrow viewing angle range, such as 10 degrees, is measured in the vertical direction of the traveling plane of the mobile robot. , you need to ensure that the horizontal viewing angle is greater than the vertical viewing angle. Thereby, the narrow 3D point cloud information in the area covered by the three-dimensional space similar to the spherical quadrangular pyramid is measured.

Preferably, the method for complementing the visual positioning result of the image information collected by the camera and the distance information obtained by the fan-shaped distance sensor to reduce the positioning blind area includes: when it is detected that the image information currently collected by the camera does not completely cover the local positioning area, That is to say, the actual position that matches the landmark features converted from the currently collected image information cannot cover the complete local positioning area (there is an idle area, but this idle area belongs to the field of view of the currently collected image information, but due to the camera's perspective Blind area and can not be measured), use the distance information of the three-dimensional point cloud inside the single spherical pyramid sensing area formed by the current emission of the fan-shaped distance sensor to complete the pose information supplement of the boundary of the local positioning area on the same map , to achieve a fusion positioning effect of vision sensor and distance sensor. Specifically, the three-dimensional point cloud information in the spherical pyramid-type sensing area currently formed in front of the fan-shaped distance sensor is used to construct a short-range small-area boundary to supplement the part of the local positioning area that is not covered by the image information of the camera. Area position, so that the visual positioning result of the image information and the distance information obtained by the fan-shaped distance sensor complement each other to reduce the blind area of the scanning position, for example, it is a too far failure position for one of the sensors, or for one of the sensors. It is too close to the position where the distance measurement is missed, which is easy to generate a positioning blind spot. Overcome the technical defect that the camera cannot locate local areas due to the influence of the lighting environment and installation location. It is worth noting that, it is worth noting that the local positioning area is the overlapping area of the current viewing angle range of the camera of the mobile robot and a spherical pyramid-shaped sensing area currently emitted by the fan-shaped distance sensor. At the same time, if the point cloud information of the area outside the local positioning area is also acquired by the fan-shaped distance sensor, the corresponding distance information is also converted and marked on the map.

Preferably, the method for complementing the visual positioning result of the image information collected by the camera and the distance information obtained by the fan-shaped distance sensor to reduce the positioning blind spot includes: when detecting that a spherical pyramid-shaped transmission formed by the current emission of the fan-shaped distance sensor is detected When the distance information of the three-dimensional point cloud inside the sensing area does not completely cover the local positioning area, that is, the actual position converted from the distance information of the three-dimensional point cloud inside the formed spherical pyramid sensing area cannot completely cover the local positioning area. (There is a free area, but this free area is originally a part of a spherical pyramid-type sensing area, but it cannot be measured due to the measurement error of the effective ranging distance of the fan-shaped distance sensor or the insufficient distance), use the camera current The collected image information completes the supplement of the pose information of the boundary of the local positioning area on the same map, so as to realize another fusion positioning effect of the vision sensor and the distance sensor, that is, the boundary position information or visual image matched by the fusion of visual features. The located boundary position information supplements the boundary positions beyond the maximum effective ranging distance of the fan-shaped distance sensor, so as to obtain a relatively complete obstacle contour. It is worth noting that the local positioning area is the overlapping area between the current viewing angle range of the camera of the mobile robot and a spherical pyramid-shaped sensing area formed by the current emission of the sector-shaped distance sensor, so that the sector-shaped distance sensor currently emits A spherical pyramid-shaped sensing area formed actually includes this local positioning area, but the position information of some areas appears in the effective ranging of the fan-shaped distance sensor due to errors or insufficient effective ranging distance of the fan-shaped distance sensor. Outside the range, the distance blind area of the scanning boundary is generated. At the same time, if the image of the area outside the local positioning area is captured by the camera, the corresponding image information is converted and marked on the same map.

It is worth noting that the spherical pyramid optical signal is one of all modulation signals of the sector distance sensor, and the rest of the modulation signals are not configured for ranging; wherein the sector distance sensor is a 3d- tof sensor. Compared with the ranging scanning scheme of the emitted surface light source, this embodiment does not require high power, and is suitable for uniformly detecting distance information in a small area.

It should be noted that in general scenarios, the fan-shaped distance sensor can simultaneously modulate to generate at least one of the spherical pyramid optical signals or other types of modulation signals, and these modulated signals and a plurality of the spherical pyramid optical signals are emitted synchronously to become A surface light source that covers a large area. However, in this embodiment, it is only allowed to control the emission of one of the spherical pyramid-type optical signals for distance measurement. Compared with the emission surface light source, the emission power of the 3d-tof sensor is reduced, which is suitable for uniform detection in a small area. The obstacle distance information; wherein, the sector distance sensor is a 3d-tof sensor. The spherical pyramid optical signal is a strip-shaped optical path emitted by a 3d-tof sensor. Multiple strip-shaped optical paths form the aforementioned surface light source. The horizontal plane of the strip-shaped optical path is fan-shaped, and the strip-shaped optical path is a There are strips of light with a certain height in the vertical direction of the horizontal plane.

Before the mobile robot performs the supplementary fusion positioning of the boundary pose information in the preceding preferred example, or after performing the supplementary fusion positioning of the boundary pose information, this embodiment also requires: using one of the spherical surfaces currently acquired by the fan-shaped distance sensor The distance information of the three-dimensional point cloud inside the pyramid-type sensing area is scanned to mark the current contour, and then it is judged whether the current contour and the historical contour in the same area of the pre-stored historical map library meet the preset coincidence degree; If the contour does not conform to the preset coincidence degree, the current contour is rotated and translated according to the pose relationship between the current contour and the historical contour, and then the current contour is rotated and translated, so that the current contour after the rotation and translation is consistent with the historical contour. Preset the coincidence degree to complete the correction of the boundary scanned by the fan-shaped distance sensor. In this embodiment, in the spherical pyramid sensing area, the contour scanned by the fan-shaped distance sensor is used to perform pose matching, so as to compensate the misalignment data caused by the bumping or slipping of the mobile robot.

As shown in FIG. 4 , under the environmental conditions that the obstacles in the maximum effective ranging area block the reflection interference, the spherical pyramid-shaped sensing area formed by the spherical pyramid optical signal coverage is a spherical pyramid-shaped three-dimensional Space, a strip of light emitted from the fan-shaped distance sensor 103 to the distance. The projection of the spherical pyramid optical signal on the traveling plane of the mobile robot is a horizontal fan-shaped area, and this horizontal fan-shaped area takes the installation position of the fan-shaped distance sensor 103 as the vertex and the maximum effective ranging distance d as the radius. sector. In this embodiment, the fan-shaped distance sensor 103 emits a strip-shaped optical path, and the strip-shaped optical path is: the traveling plane of the mobile robot is fan-shaped, and is perpendicular to the traveling plane of the mobile robot and has a certain height. Compared with the surface light source scanning method, it is better to realize the uniformity of illumination coverage of the small area detection light signal. Guarantee the accuracy of ranging.

As an embodiment, the present invention also discloses a fusion positioning method for marking map boundary information through positioning, which specifically includes: during the process of the mobile robot walking along a preset planned path, receiving feedback light according to a fan-shaped distance sensor In the case of the signal, mark the boundary of the effective ranging area of a spherical pyramid optical signal emitted by the fan-shaped distance sensor on the map, which can include the contour position of the obstacle within the maximum effective ranging area of the spherical pyramid optical signal; Among them, the fan-shaped distance sensor is installed at the front end of the mobile robot. As shown in FIG. 1 , a fan-shaped distance sensor 103 is installed at the front end of the body 101 of the mobile robot; the maximum effective ranging area of the spherical pyramid optical signal is the spherical pyramid in front of the mobile robot. As shown in the fan-shaped distance sensor 103 in FIG. 1, the probe direction of the fan-shaped distance sensor 103 also faces the front of the body 101, and emits a fan-shaped area 104 with the fan-shaped distance sensor 103 as the vertex. The fan-shaped area 104 is The projection plane of the spherical pyramid optical signal on the traveling plane of the mobile robot; wherein, the feedback optical signal is obtained by the spherical pyramid optical signal emitted by the fan-shaped distance sensor and reflected from the obstacles in the maximum effective ranging area of the spherical pyramid optical signal . In this embodiment, whenever the mobile robot moves to a position along the preset planned path, the fan-shaped distance sensor is kept controlled to emit a spherical pyramid optical signal to detect the environmental distance information of the corresponding coverage area, and the fan-shaped distance sensor receives the information in real time. The relative distance information of the obstacle reflecting the feedback signal is obtained after the distance calculation of the feedback optical signal, and it is also the shape and outline of the detected obstacle that defines the boundary position range of the effective ranging area of the spherical pyramid optical signal. It is understandable It is: the existence of this obstacle reduces the occupation area of the maximum effective ranging area of a currently transmitted spherical pyramid optical signal on the traveling plane of the mobile robot. Compared with the prior art, the present embodiment emits a spherical pyramid optical signal to detect the three-dimensional point cloud in the environmental outline in the maximum effective ranging area, and uses these three-dimensional point clouds to mark the boundaries on the map and describe the environment. , Compared with the lidar scanning profile, the fan-shaped distance sensor does not need to rotate and scan for emission, and only emits a spherical pyramid-shaped optical signal, with good output uniformity, ensuring the accuracy of ranging, and the cost of mold opening is moderate, suitable for large-scale industries. change.

As an example, when the mobile robot walks along the preset planned path, if the fan-shaped distance sensor does not receive the feedback light signal during the process of the mobile robot walking along the preset planned path, it is determined that the mobile robot does not Detect the obstacles on both sides of the current walking direction and the obstacles in front of the current position, and then mark the boundary of the spherical pyramid sensing area formed by covering a spherical pyramid optical signal emitted by the fan-shaped distance sensor as the mobile robot at the current location. The map boundary of the effective ranging area of the location; it should be noted that the feedback optical signal is the reflection signal of a spherical pyramid optical signal emitted by the fan-shaped distance sensor, and the nature of the signal does not change. Specifically, as shown in FIG. 2 , when the mobile robot goes straight to the first preset position P1 along the second preset planning direction of the preset planning path 201 (the arrow in FIG. 2 points to), if the fan-shaped distance sensor does not receive If the light signal is fed back, it is determined that the mobile robot has not detected the obstacles on both sides of the second preset planning direction and the obstacles in front of the first preset position P1, that is, the first preset position P1 is in the front fan-shaped grid pointed by the arrow. No obstacle is detected on the grid area 204, and then a spherical pyramid-shaped optical signal (a strip-shaped optical path mentioned in the previous embodiment) emitted by the fan-shaped distance sensor is covered to form the boundary of the spherical pyramid-shaped sensing area and marked as moving The map boundary of the effective ranging area of the robot at the first preset position P1, this spherical pyramid sensing area corresponds to the fan-shaped grid area 204 covering the travel plane of the mobile robot in FIG. 2, and the fan-shaped grid area 204 is the spherical surface The pyramid-shaped sensing area is in the projection area of the traveling plane of the mobile robot; as shown in FIG. 2 , the mobile robot walks along the second preset planning direction of the preset planning path 201 (the arrow in FIG. 2 points to) but does not reach the first preset path. Before setting the position P1, a spherical pyramid optical signal emitted by the fan-shaped distance sensor scans the grid areas on the left and right sides of the preset planning path 201 in FIG. A spherical pyramid optical signal has been scanned to cover the parallelogram grid area 203 shown in FIG. 2 (the grid area on the right side of the preset planned path 201, and the scanning method of the grid area on the left side of the preset planned path 201 is similar), At this time, the mobile robot walks along the preset planned path 201 and occupies the diagonally filled grid area 202 in FIG. 2 . Among them, the spherical pyramid sensing area is a spherical pyramid-shaped three-dimensional space. If no obstacle is detected, the point cloud information in the spherical pyramid sensing area is directly marked on the map to describe a spherical pyramid light. The boundary features of the effective ranging area of the signal. In this embodiment, the spherical pyramid-shaped optical signal emitted by the fan-shaped distance sensor constructs a three-dimensional effective ranging space in the shape of a spherical quadrangular pyramid on the premise of no obstruction and reflection, and it is not necessary to construct a three-dimensional point that is too large and too wide. Cloud space boundary. Let the user obtain the environmental contour of the scanned map in advance on the display interface.

As another embodiment, when the mobile robot walks along the preset planned path, if the sector distance sensor receives the feedback light signal, the time of flight corresponding to the feedback light signal is calculated to reflect the feedback light signal. The distance information between the obstacle and the fan-shaped distance sensor, and then mark the contour position of the obstacle reflecting the feedback light signal on the map based on this distance information, so that the contour position of the obstacle becomes a spherical surface emitted by the fan-shaped distance sensor The boundary of the effective ranging area of the pyramidal optical signal, or the effective boundary of the spherical pyramidal optical signal on the traveling plane of the mobile robot. It should be noted that this feedback optical signal is reflected by the obstacle, and is obtained by reflecting a spherical pyramid optical signal emitted by the fan-shaped distance sensor by the obstacle, and the reflected spherical pyramid optical signal is the feedback optical signal.

Specifically, as shown in FIG. 3 , when the robot goes straight to the second preset position P2 along the first preset planning direction (the arrow in FIG. 3 points to) of the preset planning path 301 , if the fan-shaped distance sensor receives the feedback light signal , the distance information between the boundary 3041 of the linear obstacle reflecting the feedback optical signal and the fan-shaped distance sensor is calculated according to the flight time recorded by the feedback optical signal, and the right angle reflecting the feedback optical signal is calculated at the same time. The distance information between the boundary 30421 of the obstacle and the fan-shaped distance sensor at the same position, and the distance information between the boundary 30422 of the right-angle obstacle reflecting the feedback light signal and the fan-shaped distance sensor at the same position, and then Based on the three distance information, the contour positions of the straight-line obstacles and the right-angle obstacles that reflect the feedback light signal are marked on the map. Then, the detection result of the mark on the map is shown in Figure 3. The linear obstacle occupies the left part of the grid area of the spherical pyramid sensing area of a beam of spherical pyramid optical signals, so that the boundary 3041 of the linear obstacle becomes the position P2 The boundary of the effective ranging area on the left side of the spherical pyramid optical signal emitted by the fan-shaped distance sensor.

As shown in Figure 3, the right-angled obstacle occupies the right part of the grid area of the spherical pyramid-shaped sensing area of a beam of spherical pyramid-shaped optical signals, so that the boundary 30421 of the right-angled obstacle becomes the spherical surface emitted by the fan-shaped distance sensor at position P2 The obstacle boundary of the effective ranging area on the right side of the pyramid optical signal; the right angle obstacle occupies the front part of the grid area of the spherical pyramid sensing area of a beam of spherical pyramid optical signal, so that the boundary 30422 of the right angle obstacle becomes The obstacle boundary of the effective ranging area on the front side of the spherical pyramid optical signal emitted by the fan-shaped distance sensor at position P2. Of course, in the spherical pyramid sensing area of the spherical pyramid optical signal, the boundary of the grid area that is not occupied by obstacles remains the same spherical pyramid sensing area projected on the robot traveling plane in Figure 3. The boundary of the fan-shaped grid area is also the arc-shaped boundary of the part of the fan-shaped grid area that is not occupied by obstacles.

As shown in FIG. 3 , before the mobile robot walks along the first preset planning direction of the preset planning path 301 (pointed by the arrow in FIG. 3 ) but does not reach the second preset position P2 , the fan-shaped distance sensor emits a spherical pyramid-shaped The optical signal scans out the map grid areas on the left and right sides of the preset planning path 301 in FIG. 3 . When walking to the second preset position P2, a spherical pyramid optical signal emitted by the fan-shaped distance sensor has been scanned and covered as shown in FIG. 3 . Grid area 303 (the grid area on the right side of the preset planning path 301, the scanning method of the grid area on the left side of the preset planning path 301 is similar). Regular graphic boundaries in the example. At this time, the mobile robot walks along the preset planned path 301 and occupies the grid area 302 filled with oblique lines in FIG. 3 . Among them, the spherical pyramid sensing area is a spherical pyramid-shaped three-dimensional space. If no obstacle is detected, the point cloud information in the spherical pyramid sensing area is directly marked on the map to describe a spherical pyramid light. The boundary features of the effective ranging area of the signal. This embodiment constructs a three-dimensional effective ranging space inside a spherical quadrangular pyramid, allowing the user to obtain the contour features of obstacles in advance, forming the farthest distance that the spherical pyramid optical signal can measure on the traveling plane of the mobile robot effective frontier.

On the basis of the foregoing embodiment, the walking direction of the mobile robot is adjusted by changing the path direction of the preset planned path, so that when the walking direction of the mobile robot changes each time, the coverage area of the spherical pyramid optical signal emitted in real time is increased. There are different effective ranging areas. As the walking direction of the mobile robot changes, the coverage of the effective ranging area of the spherical pyramid optical signal emitted by the mobile robot becomes wider. After a period of planning time and traversing a planned path, the effective ranging area of the same spherical pyramid optical signal emitted by the fan-shaped distance sensor covers the boundary of the current working area, or the same spherical pyramid formed by the fan-shaped distance sensor emission The actual physical boundary of the current working area is covered by the sensor area. In this way, the surrounding environment can be scanned and marked on the map.

As an example, the camera of the mobile robot is controlled to collect image information for visual positioning; the fan-shaped distance sensor of the mobile robot is controlled from a single spherical pyramid sensing area at the front end (the maximum effective measurement area formed by the single spherical pyramid optical signal construction). The distance information is obtained from the inside of the distance area), and then the visual positioning results of the image information collected by the camera and the distance information obtained by the fan-shaped distance sensor are complemented and fused with each other to reduce the blind area of the scanned boundary, and then the complementary fusion positioning results are marked on the map. ; Among them, the result of the supplementary fusion positioning is the pose information of the effective boundary within the single spherical pyramid-shaped sensing area emitted by the fan-shaped distance sensor of the mobile robot at a specific position. As the mobile robot walks along the preset planned path, the supplementary fusion is performed. There are more and more localization results, and the spherical pyramid-shaped sensing area sent by the fan-shaped distance sensor and the image information collected by the camera have an increasing range, which is enough to cover the current working area.

The method for controlling the fan-shaped distance sensor of the mobile robot to obtain distance information from the interior of the spherical pyramid-type sensing area at the front end thereof includes: firstly, controlling the fan-shaped distance sensor to emit a spherical pyramid-shaped light signal, wherein the spherical pyramid-shaped light signal is The maximum effective ranging area of the optical signal is the spherical pyramid sensing area, which is equivalent to the maximum effective ranging area of the spherical pyramid optical signal on the premise that there is no obstruction and reflection. When the fan-shaped distance sensor receives the feedback light signal reflected by the obstacle in the spherical pyramid sensing area, it calculates and obtains the three-dimensional point cloud of the corresponding obstacle according to the flight time recorded by receiving the feedback light signal. Relative to the distance information of the fan-shaped distance sensor, the corresponding obstacle is the obstacle that reflects the spherical pyramid-shaped optical signal to the fan-shaped distance sensor after the fan-shaped distance sensor emits the spherical pyramid-shaped optical signal; wherein, The reflection signal of the spherical pyramid optical signal through the obstacle is the feedback optical signal. The distance calculation method of this embodiment can be regarded as controlling the fan-shaped distance sensor to transmit the spherical pyramid optical signal at the same position and receive the spherical pyramid optical signal (feedback optical signal) reflected by the obstacle, and then pass the spherical pyramid optical signal through the spherical pyramid optical signal. The time of flight is calculated to determine the distance information of the obstacle relative to the position, and then the boundary contour of the obstacle is determined.

As shown in FIG. 1 , an embodiment of the present invention discloses a mobile robot, which includes a camera, a fan-shaped distance sensor 103 and a processing unit; The sides are respectively equipped with driving wheels 102. Preferably, the center of the body of the mobile robot is the center point of the left driving wheel and the right driving wheel. Right-hand drive wheel drive. The camera is also installed on the surface of the mobile robot, but the orientation of the camera is different from the orientation of the probe of the fan-shaped distance sensor 103, so that the viewing angle coverage of the camera is not exactly the same as the effective distance measurement range of the fan-shaped distance sensor 103, and there may be some overlapping collection viewing angles. ; The processing unit is configured to execute the boundary marking method based on the spherical pyramid sensing area of the foregoing embodiments. The mobile robot uses a non-rotating fan-shaped distance sensor to replace the lidar that needs to be rotated to obtain the outline information of the room area, so as to achieve the same technical effect of accurate map boundary marking.

Preferably, the fan-shaped distance sensor is a 3d-tof sensor, and the projection of the spherical pyramid-shaped light signal emitted by the 3d-tof sensor on the traveling plane of the mobile robot is a horizontal fan-shaped area with a horizontal viewing angle of 120 degrees. The spherical pyramid optical signal emitted by the -tof sensor has a vertical viewing angle of 10 degrees in the vertical direction of the traveling plane of the mobile robot. Figuratively speaking, the maximum effective ranging area of the spherical pyramid optical signal is like a cake cut into a fan shape. This sensor can measure the distance information of the area covered by the cake, which is a narrow 3D point cloud information. In this embodiment, the fan-shaped distance sensor does not need to be controlled to collect information in a wide range, and the 3d-tof sensor only needs to emit a beam of spherical pyramid-shaped optical signals to meet the requirements of distance measurement and positioning, and the transmission power is also reduced, which is easy to achieve small The uniformity of light received by the area. It can also be understood as: the effective ranging area of the spherical pyramid optical signal emitted by the fan-shaped distance sensor is a type of fan-shaped cake, so as to measure the distance information of the area covered by the cake, and form a narrow three-dimensional point cloud information.

It should be noted that the fan-shaped distance sensor is a long-distance sensor, and the effective distance measurement range of the fan-shaped distance sensor is 4 meters to 50 meters. It is placed on the surface of the mobile robot body 101 and irradiated outward. It is suggested to place the fan-shaped distance sensor directly in front of the mobile robot body 101 . But the sector-shaped distance sensor is not a single-point distance sensor.

Preferably, the installation position of the fan-shaped distance sensor passes through the central axis of the body of the mobile robot to ensure that the fan-shaped distance sensor is directly in front of the body; or the installation position of the fan-shaped distance sensor and the body central axis of the mobile robot form a predetermined Set the angle. Therefore, it is ensured that the central axis of the body of the mobile robot passes through the effective ranging area of the spherical pyramid optical signal emitted by the fan-shaped distance sensor.

Obviously, the above-mentioned embodiments are only examples for clear description, and are not intended to limit the implementation manner. For those of ordinary skill in the art, changes or modifications in other different forms can also be made on the basis of the above description. There is no need and cannot be exhaustive of all implementations here. And the obvious changes or changes derived from this are still within the protection scope of the present invention.

Claims

A fusion positioning method, comprising:

Control the camera of the mobile robot to collect image information for visual positioning;

It is characterized in that it also includes:

The fan-shaped distance sensor that controls the mobile robot obtains distance information from the inside of the spherical pyramid-shaped sensing area at its front end, and then fuses the visual positioning results of the image information collected by the camera with the distance information obtained by the fan-shaped distance sensor, and then fuses the complementary The positioning result is marked on the map;

Among them, the fan-shaped distance sensor is installed on the front end of the mobile robot, and the camera is also installed on the surface of the mobile robot, but the direction of the probe of the fan-shaped distance sensor is different from that of the camera.
The fusion positioning method according to claim 1, wherein the method for controlling the fan-shaped distance sensor of the mobile robot to obtain distance information from the interior of the spherical pyramid-type sensing area at the front end of the mobile robot comprises:

Controlling the sector distance sensor to modulate and emit a spherical pyramid optical signal, wherein the largest effective ranging area of the spherical pyramid optical signal is the spherical pyramid sensing area;

When the fan-shaped distance sensor receives the feedback light signal reflected from the obstacle in the spherical pyramid sensing area, it calculates and obtains the position of the corresponding obstacle relative to the time of flight recorded by receiving the feedback light signal. The distance information of the fan-shaped distance sensor.
The fusion positioning method according to claim 2, wherein the distance of the three-dimensional point cloud inside the spherical pyramid sensing area relative to the fan-shaped distance sensor is all within the effective ranging range of the fan-shaped distance sensor Inside;

Wherein, the maximum value of the effective ranging range of the fan-shaped distance sensor is the maximum effective ranging distance;

Wherein, the projection of the spherical pyramid optical signal on the traveling plane of the mobile robot is a horizontal fan-shaped area, and this horizontal fan-shaped area takes the installation position of the fan-shaped distance sensor as the vertex and the maximum effective ranging distance as the radius. sector.
The fusion positioning method according to claim 3, wherein the spherical pyramid optical signal has a horizontal viewing angle on the traveling plane of the mobile robot, and the spherical pyramid optical signal is in the vertical direction of the traveling plane of the mobile robot. There is a vertical viewing angle, where the horizontal viewing angle is greater than the vertical viewing angle.
The fusion positioning method according to claim 4, wherein the method for complementing and fusing the visual positioning result of the image information collected by the camera and the distance information obtained by the fan-shaped distance sensor comprises:

When it is detected that the image information currently collected by the camera does not completely cover the local positioning area, use the distance information of the three-dimensional point cloud inside a spherical pyramid sensing area formed by the current emission of the sector distance sensor to complete the local positioning on the same map Supplementing the pose information of the area; wherein, the local positioning area is the overlapping area of the current viewing angle range of the camera of the mobile robot and a spherical pyramid-shaped sensing area formed by the current emission of the fan-shaped distance sensor.
The fusion positioning method according to claim 4, wherein the method for complementing and fusing the visual positioning result of the image information collected by the camera and the distance information obtained by the fan-shaped distance sensor comprises:

When it is detected that the distance information of the three-dimensional point cloud inside a spherical pyramid sensing area formed by the current emission of the fan-shaped distance sensor does not completely cover the local positioning area, use the image information currently collected by the camera to complete the local positioning on the same map Supplementing the pose information of the area; wherein, the local positioning area is the overlapping area of the current viewing angle range of the camera of the mobile robot and a spherical pyramid-shaped sensing area formed by the current emission of the fan-shaped distance sensor.
The fusion positioning method according to any one of claims 2 to 6, wherein the fan-shaped distance sensor is used to modulate and generate at least one of the spherical pyramid optical signals or other types of modulation signals, but only allows control of emission one of the spherical pyramid optical signal transmission for ranging;

Wherein, the fan-shaped distance sensor is a 3d-tof sensor.
The fusion positioning method according to claim 5 or 6, characterized in that, within the effective ranging range of the fan-shaped distance sensor, the interior of one of the spherical pyramid-type sensing areas currently acquired by the fan-shaped distance sensor is used. The distance information of the three-dimensional point cloud is scanned to mark the current contour, and then it is judged whether the current contour and the historical contour in the same area of the pre-stored historical map library meet the preset coincidence degree;

If the current contour and the historical contour do not meet the preset coincidence degree, the current contour is rotated and translated according to the pose relationship between the current contour and the historical contour, so as to correct the current contour after the rotation and translation transformation to be consistent with the historical contour. Set coincidence.
A mobile robot is characterized in that it includes a camera, a fan-shaped distance sensor and a processing unit; the fan-shaped distance sensor is installed on the front end of the mobile robot; the camera is also installed on the surface of the mobile robot, but the orientation of the camera is different from that of the probe of the fan-shaped distance sensor. , so that the viewing angle coverage of the camera is not exactly the same as the effective ranging range of the fan-shaped distance sensor;

The processing unit is configured to execute the fusion positioning method according to any one of claims 1 to 8.
The mobile robot according to claim 9, wherein the fan-shaped distance sensor is a 3d-tof sensor, and the projection of the spherical pyramid optical signal emitted by the 3d-tof sensor on the traveling plane of the mobile robot is a horizontal viewing angle It is a horizontal sector of 120 degrees, and the spherical pyramid optical signal emitted by this 3d-tof sensor has a vertical viewing angle of 10 degrees in the vertical direction of the traveling plane of the mobile robot.
The mobile robot according to claim 9 or 10, wherein the installation position of the fan-shaped distance sensor passes through the center axis of the body of the mobile robot, or the installation position of the fan-shaped distance sensor and the body center axis of the mobile robot form a preset angle.