Line laser-based short barrier identification method

Abstract

The invention discloses a line laser-based low barrier identification method, which comprises the following steps: step 1, in the process that a robot walks along a navigation direction, a line laser sensor is controlled to emit line laser along the inclined lower part of the navigation direction, and two frames of line laser data are sequentially collected; the mounting direction of the line laser sensor relative to the center of the body is parallel to the navigation direction; 2, in the two frames of line laser data collected in sequence, the number of repeated angles and the total number of occupied angles are obtained by counting the situation of repeated occupied angles of laser points in the two frames of line laser data collected in sequence, and when the size relation between the number of repeated angles and the total number of occupied angles meets a preset screening condition, screening is carried out on the two frames of line laser data; and recognizing the surface of the to-be-detected object as a short barrier in combination with the trend characteristics of the laser slope in each frame of line laser data. A short ridge which the robot should cross originally is prevented from being recognized as a curve type obstacle, and the robot can cross directly after recognizing the short ridge type obstacle.

Description

Technical Field

[0001] The present application relates to the field of line laser technology, and in particular to a method for identifying low-rise obstacles based on line laser. Background Art

[0002] At present, sweeping robots use line lasers to identify low obstacles (generally no higher than 20mm), such low obstacles include ropes, etc. In some home scenes, low barriers include guide rails, sliding door rails, and door steps. The regularity of the contour surface of the rope that is zigzag and flat on the ground is different from that of the low barriers. The contour height distribution characteristics of the rope that is zigzag and flat on the ground at a specific detection angle recognized by the sweeping robot are similar to the contour height distribution characteristics reflected from the surface of the low barriers when the sweeping robot walks along a specific oblique direction relative to the low barriers. This makes it easy for the sweeping robot to identify the low barriers that should have been crossed as ropes. Therefore, in order to prevent the sweeping robot from being entangled by ropes and unable to walk normally, the sweeping robot will turn around when it identifies the low barriers that should have been crossed as ropes, which will cause the sweeping robot to miss sweeping and misclassify low obstacles. Summary of the invention

[0003] A method for identifying low-hill obstacles based on line laser, wherein a line laser sensor is installed on the front side of a robot body; the method for identifying low-hill obstacles comprises: step 1, when the robot walks along a navigation direction, controlling the line laser sensor to emit line laser along an oblique downward direction of the navigation direction, and sequentially collecting two frames of line laser data; the installation direction of the line laser sensor relative to the center of the body is parallel to the navigation direction; step 2, in the two frames of line laser data sequentially collected, obtaining the number of repeated angles and the total number of occupied angles by counting the repeated occupied angles of obstacle points in the two frames of line laser data sequentially collected, and when the size relationship between the number of repeated angles and the total number of occupied angles meets a preset screening condition, the surface of the object to be measured is identified as having low-hill obstacles in combination with the trend characteristics of the laser slope in each frame of line laser data.

[0004] By executing the method for identifying low-ridge obstacles, the robot of the present application utilizes the laser slope in each frame of line laser data, and the size relationship between the number of repeated angles and the total number of occupied angles, to successively eliminate the interference of curve obstacles and wall obstacles, and identify the existence of the low-ridge obstacles on the surface of the object to be measured; prevent the low ridge that the robot should have crossed from being identified as a curve obstacle, and overcome the problem of misidentification of the low-ridge obstacles pointed out in the background technology because the contour height distribution characteristics of the rope and the contour height distribution characteristics of the low-ridge surface are similar. The robot can directly cross the low-ridge obstacle after identifying it, reducing the problem of missing the robot's navigation traversal area. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Figure 1 A flowchart of a method for identifying low-hill obstacles is disclosed in an embodiment of the present application.

[0006] Figure 2 An embodiment of the present application discloses a schematic diagram of a detection area 103 of a line laser sensor 102 in front of a robot 101 and a preset detection angle range (angle AOB with the body center O as the vertex) of the line laser sensor 102; within the detection range 103, a line segment 104 is a light strip AB formed in front of the robot by the line laser emitted by the line laser sensor; the preset detection angle range (angle AOB with the body center O as the vertex) is an angle of 60 degrees formed by the line segment 104 relative to the body center O.

[0007] Figure 3 An embodiment of the present application discloses a schematic diagram of a line laser projected onto a low-ridge obstacle 105 when the robot walks to a first position along a navigation direction OP in a scenario where the robot walks toward the low-ridge obstacle 105; wherein, when the light bar 104 passes through the low-ridge obstacle 105, a laser slope 1041 appears (defined as an ascending laser slope in a clockwise direction and a descending laser slope in a counterclockwise direction), and an extension direction of the low-ridge obstacle 105 forms a certain angle with the navigation direction OP but is not perpendicular to each other.

[0008] Figure 4 The present application discloses a schematic diagram of a scenario in which a robot walks toward a low-step obstacle 105 and a line laser is projected onto the low-step obstacle 105 when the robot walks along the navigation direction OP to a second position. The second position is relative to Figure 3 The first position is closer to the short-hill obstacle 105; wherein, when the light bar 104 passes the short-hill obstacle 105, a laser slope 1042 appears (defined as an ascending laser slope in a clockwise direction and a descending laser slope in a counterclockwise direction), and the extension direction of the short-hill obstacle 105 forms a certain angle with the navigation direction OP but is not perpendicular to each other.

[0009] Figure 5 An embodiment of the present application discloses a scenario in which a robot walks toward a curved obstacle 106. When the robot walks to a first position along a navigation direction OP, a line laser is projected onto the curved obstacle 106. The curved obstacle 106 is placed horizontally on the walking surface of the robot in a curved and uneven shape.

[0010] Figure 6An embodiment of the present application discloses a scenario in which a robot walks toward a curved obstacle 106. When the robot walks to a second position along a navigation direction OP, a line laser is projected onto the curved obstacle 106. The curved obstacle 106 is placed horizontally on the walking surface of the robot in a curved and uneven shape. The second position is relative to Figure 5 The first position is closer to the curved obstacle 106 .

[0011] Figure 7 An embodiment of the present application discloses a robot in Figure 5 A frame of line laser data collected in Figure 6 Schematic diagram of the logic processing of a frame of line laser data collected in the image according to the observation angle range divided in sequence within the preset detection angle range, where frame1 represents the robot in Figure 5 A frame of line laser data collected in , frame2 represents the robot Figure 6 A frame of line laser data collected in , each square represents the obstacle marking logic information corresponding to the laser points occupying the corresponding observation angle range. The white squares indicate that the laser points occupying the corresponding observation angle range are not marked as obstacle points, and the black squares indicate that the laser points occupying the corresponding observation angle range are marked as obstacle points.

[0012] Figure 8 yes Figure 1 A flowchart of the execution method of step 2 in the disclosed method for identifying low-hill obstacles. Implementation

[0013] The technical solutions in the embodiments of the present invention will be described in detail below in conjunction with the drawings in the embodiments of the present invention. To further illustrate the embodiments, the present invention provides drawings. These drawings are part of the disclosure of the present invention, which are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. In the description of this application, the directions or positional relationships of "left", "right", "up", "down", "clockwise", and "counterclockwise" are based on the attached drawings. Figure 2 To Attachment Figure 7 It should be understood that the orientation or positional relationships shown are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.

[0014] As an embodiment, a method for identifying low-rise obstacles is disclosed. The executor of the method for identifying low-rise obstacles is a robot. A line laser sensor is installed on the front side of the robot body. The robot can be a sweeping robot and can have any suitable shape, such as a more traditional round body, a D-shaped body, or a rectangular body. The robot is used to control the line laser sensor to emit a line laser to the surface of the object to be measured during walking. The line laser forms a light strip on the surface of the object to be measured. The light strip at the same position is reflected to form a frame of line laser data. A frame of line laser data is line laser data at each angle within a preset detection angle range. The line laser sensor is used to collect line laser data within the preset detection angle range.

[0015] like Figure 1 As shown, the method for identifying low-hill obstacles includes: Step 1: When the robot walks along the navigation direction, the line laser sensor is controlled to emit a line laser obliquely downward along the navigation direction, and two frames of line laser data formed by the line laser reflected from the surface of the object to be measured are collected in sequence, and then step 2 is performed. Preferably, the installation direction of the line laser sensor relative to the center of the body is parallel to the navigation direction. When the navigation direction is set to be parallel to the central axis of the body, the line laser sensor is installed on the front side of the robot body along the central axis of the body, and the line laser sensor emits a line laser to the surface of the object to be measured, and the surface of the object to be measured reflects to form a light strip, which forms a corresponding bright section, which can be regarded as composed of projected light spots (i.e., laser spots) of multiple laser beams. The extension direction of the light strip formed on the horizontal ground is set perpendicular to the navigation direction; in some embodiments, the light strip formed by the line laser sensor emitted to the surface of the object to be measured is regarded as being separated on the left and right sides of the central axis of the body (equivalent to being parallel to the navigation direction OP), which can be intersecting, parallel or partially overlapping, and the specific distribution form varies depending on the contour shape of the obstacle protruding from the ground.

[0016] When obstacle detection is not started, the line laser sensor is activated. Figure 2 This is a schematic diagram of an embodiment of the present application that discloses a detection area 103 of a line laser sensor 102 in front of a robot 101 and a preset detection angle range (angle AOB with the center O of the body as the vertex) of the line laser sensor 102; within the detection range 103, a line segment 104 is a light bar AB formed in front of the robot by the line laser emitted by the line laser sensor. The installation direction of the line laser sensor 102 relative to the center O of the body is parallel to the navigation direction OP, and the navigation direction OP is perpendicular to the light bar AB.

[0017] For the line laser sensor, the emission angle of the line laser can be 15 degrees, 30 degrees, 45 degrees, 60 degrees, etc. from the horizontal direction. This embodiment does not limit the angle obliquely downward from the walking direction or the value of the emission angle of the line laser. The line laser transmitter is configured to illuminate a width area larger than the width of the robot, forming a bright section or a light strip on the surface of the object to be measured. The light strip may not be a straight line or a continuous curve because some obstacle points protrude from the ground, and there may be disconnected positions, resulting in discontinuity of the light strip.

[0018] In the robot, the line laser sensor is also used to collect laser points formed by the reflection of the line laser on the surface of the object to be measured. In essence, the infrared camera collects images of laser points formed by the reflection of the line laser on the surface of the object to be measured, covering all or part of the image of the environment illuminated by the line laser; the infrared camera captures and records multiple frames of line laser data per second, preferably 33 frames per second, and each frame of line laser data includes approximately 120 laser points, and each laser point includes three types of posture data: angle, distance, and height.

[0019] When the robot walks along the navigation direction, the position where the robot is located when collecting the first frame of line laser data is marked as the first position, and the position where the robot is located when collecting the second frame of line laser data is marked as the second position. When the robot walks from the first position to the second position along the navigation direction, the distance between the robot and the surface of the object to be measured becomes smaller. The prior art may use multiple frames (more than two frames) of line laser data to jointly detect low obstacles. Whenever one more frame of data is saved for comparison and matching, the amount of calculation and storage increases, and the distance the robot needs to walk will also increase a lot, causing the robot to avoid the low obstacle before it has time to identify the low obstacle. In this embodiment, two frames of line laser data can be continuously collected.

[0020] Step 2. In two frames of line laser data collected sequentially, the number of repeated angles and the total number of occupied angles are obtained by counting the repeated occupied angles of obstacle points (laser points marked as obstacle points) in the two frames of line laser data collected sequentially. When the size relationship between the number of repeated angles and the total number of occupied angles meets the preset screening conditions, the surface of the object to be measured is identified as having low-step obstacles in combination with the direction characteristics of the laser slope in each frame of line laser data. The low-step obstacles include round stick-type low steps, square low steps and other obstacles. The round stick-type low steps can be guide rails or sliding door rails, and the square low steps can be door steps. Preferably, the height of the low-step obstacles from the ground does not exceed 20 mm, which can be regarded as a low obstacle, supporting the robot to cross.

[0021] In this embodiment, in the required identification of the surface of the object to be measured, it is necessary to distinguish between regular crossable obstacles and winding and continuous ropes (belonging to curve obstacles). In order to reduce the misjudgment problem caused by the single comparison and analysis of the height information of the laser point, this embodiment can distinguish between curve obstacles and wall obstacles based on the number of obstacle points that appear repeatedly at the same angle in two frames of line laser data. Moreover, this is performed during the walking process of the robot. The line laser detects two different areas of the surface of the object to be measured at the same position as the walking position of the robot changes. For example, in two frames of line laser data collected sequentially, the first frame of line laser data is reflected from the contour of a part of the angle interval of the curve obstacle, and the second frame of line laser data is reflected from the contour of another part of the angle interval of the curve obstacle. There are repeatedly detected contours in the two parts of the angle interval.

[0022] The obstacle points that appear repeatedly at the same angle are located in the overlapping area of ​​the two different areas. In this embodiment, the number of repeated angles can be expressed by using the situation that the obstacle points in the two frames of line laser data collected sequentially occupy the angle repeatedly; preferably, the more the number of obstacle points that appear repeatedly at the same angle, the larger the size of the obstacles on the surface of the object to be measured, including the larger the contour surface and the larger the height. For example, large obstacles such as wall obstacles need to be excluded first. Therefore, after obtaining the number of repeated angles and the total number of occupied angles, it is determined whether the size relationship between the number of repeated angles and the total number of occupied angles meets the preset screening conditions. When the preset screening conditions are met, the number of repeated angles obtained is reasonable, at least it is reasonable and not too large relative to the total number of occupied angles. It can be determined that the height of the laser points in the two frames of line laser data collected sequentially is the obstacle height that the robot is allowed to cross. In this case, there may be curved obstacles such as zigzag ropes or other types of low obstacles on the surface of the object to be measured.

[0023] On this basis, in a robot working scenario where the placement direction of the low-step obstacles or other types of obstacles on the surface of the object to be measured that the robot needs to identify is not perpendicular to the navigation direction relied on by the robot to walk, the obstacle contour line that the robot needs to identify on the surface of the object to be measured may not be parallel to the extension direction of the light strip formed by the projection of the line laser sensor; when the low-step obstacles or other types of obstacles on the surface of the object to be measured that the robot can identify are inclined at a certain angle relative to the navigation direction relied on by the robot to walk, the size relationship between the number of repeated angles and the total number of occupied angles can easily meet the preset screening conditions, and misidentification is prone to occur, including misidentification of low-step obstacles or curved obstacles (for example, a rope is misidentified as a low-step obstacle, or a slide rail threshold is misidentified as a curved obstacle); therefore, it is still necessary to combine the direction characteristics of the laser slope in each frame of line laser data to identify the surface of the object to be measured as having the said low-step obstacles. Therefore, when the preset screening conditions are met, it is judged whether there are the low step-like obstacles on the surface of the object to be measured based on the direction characteristics of the laser slope in the two-frame line laser data, including the direction characteristics of the laser slope in the same frame line laser data and the direction change characteristics of the laser slope in the two-frame line laser data collected sequentially. Specifically, it is judged whether there are regular and traversable step-like obstacles to eliminate interference from other types of low obstacles, such as curved obstacles.

[0024] In summary, in this embodiment, the robot utilizes the laser slope in each frame of line laser data, and the size relationship between the number of repeated angles and the total number of occupied angles, to successively eliminate the interference of curve obstacles and wall obstacles, and identify the existence of the short-hill obstacles on the surface of the object to be measured; prevent the short hill that the robot should have crossed from being identified as a curve obstacle, and overcome the problem of misidentification of the short hill obstacles pointed out in the background technology because the contour height distribution characteristics of the rope and the contour height distribution characteristics of the short hill surface are similar. The robot can directly cross the short hill obstacle after identifying it, reducing the problem of missing the robot's navigation traversal area.

[0025] As an embodiment, in step 2, the method of identifying the surface of the object to be measured as having a low-hill obstacle by combining the direction characteristics of the laser slope in each frame of line laser data includes: the robot determines whether two frames of line laser data collected successively have and only have laser slopes with the same direction.

[0026] When it is determined that two frames of line laser data collected successively have and only have laser slopes in the same direction, that is, the laser slope can be connected in each frame of line laser data and the direction of the slope is the same (either rising or falling, and rising and falling are not allowed to exist at the same time), the robot will identify the surface of the object to be measured as having a low-hill obstacle, and determine that the placement direction of the low-hill obstacle is at a certain inclination angle to the navigation direction, and it can be determined that the maximum height of the laser point in each frame of line laser data is less than the height threshold of the surmountable obstacle, and the height threshold of the surmountable obstacle is preferably 20 mm.

[0027] In this embodiment, it is determined that the two frames of line laser data collected sequentially have and only have laser slopes in the same direction when the size relationship between the number of repeated angles and the total number of occupied angles meets the preset screening condition. The two frames of line laser data collected sequentially can be understood as two frames of line laser data collected continuously, or two consecutive frames of line laser data, or two adjacent frames of line laser data. There is a collection time sequence between different frames of line laser data, and it is also associated with the position of the robot, that is, the robot collects two frames of line laser data at different positions at different times through the line laser sensor.

[0028] In this embodiment, the navigation direction is perpendicular to the extension direction of the light bar formed by the line laser sensor so that the light bar is not parallel to the placement direction of the low-ridge obstacle; the placement direction of the low-ridge obstacle is preferably the extension direction of the longest contour of the low-ridge obstacle.

[0029] In some embodiments, the low-hill obstacle has a contour extending at the same height along a preset direction, the preset direction is not parallel to the navigation direction, and the preset direction is not perpendicular to the navigation direction.

[0030] Low obstacles such as Figure 3 As shown, Figure 3 An embodiment of the present application discloses a schematic diagram of a line laser projected onto a low-ridge obstacle 105 when the robot walks to a first position along a navigation direction OP in a scenario where the robot walks toward a low-ridge obstacle 105; wherein, when the light bar 104 passes through the low-ridge obstacle 105, a laser slope 1041 appears (defined as an ascending laser slope in the clockwise direction (monotonically increasing at the height of the laser point), and defined as a descending laser slope in the counterclockwise direction (monotonically decreasing at the height of the laser point)), the placement direction of the low-ridge obstacle 105 forms a certain angle with the navigation direction OP but is not perpendicular to each other, and the placement direction of the low-ridge obstacle 105 is regarded as the extension direction of the longest contour of the low-ridge obstacle 105.

[0031] When it is determined that the two frames of line laser data collected successively do not all have and only have laser slopes in the same direction, it is determined that the robot detects two laser slopes in different directions in at least one frame of line laser data, and the robot recognizes the surface of the object to be measured as having a curved obstacle.

[0032] Figure 5 An embodiment of the present application discloses a scenario in which a robot walks toward a curved obstacle 106. When the robot walks to a position along the navigation direction OP, a line laser is projected onto the curved obstacle 106. The curved obstacle 106 is placed horizontally on the robot's walking surface in a tortuous and uneven shape. Then, the robot can identify the curved obstacle 106 on the surface of the object to be measured. Specifically, the light bar 104 is projected onto the surface of the curved obstacle 106. It can be regarded that the contour line of the curved obstacle 106 passes through the light bar 104 in a curved manner and a light spot offset occurs at the intersection, such as Figure 5 As shown in the black bold rectangle, multiple convex points with different directions are generated on the straight line where the light bar 104 is located; the intersection position of the light bar 104 and the curved obstacle 106 can be regarded as the position where the curved obstacle 106 is detected by the line laser, and the light bar 104 changes at the intersection position and reflects back to the line laser sensor to form a laser point reflected from the surface of the curved obstacle, and its height reflects the height characteristics of the curved obstacle surface.

[0033] For light strips, combined Figure 5 It can be seen that in the same extension direction, there are rising and falling directions of the light spot. Figure 5 On the line segment 104, the second intersection from left to right (black bold rectangle) has a rising trend of the light spot, and the fourth intersection from left to right (black bold rectangle) has a falling trend of the light spot. It can be understood that the light strip 104 has laser slopes with different directions, for example, there is at least one rising laser slope and at least one falling laser slope at the same time.

[0034] It is worth noting that after the robot collects two consecutive frames of line laser data from the curved obstacles, the size relationship between the number of repeated angles and the total number of occupied angles meets the preset screening conditions, which is consistent with the size relationship between the number of repeated angles and the total number of occupied angles extracted from the aforementioned low-ridge obstacles. Specifically, it is judged in the identification scene where the placement direction of the low-ridge obstacle is at a certain inclination angle to the navigation direction, that is, in the scene where the low-ridge obstacle is tilted relative to the walking direction of the robot, it is not easy to use the size relationship between the number of repeated angles and the total number of occupied angles to distinguish between the curved obstacles and the low-ridge obstacles. It can be understood that the laser points reflected back by low obstacles with different degrees of regularity of contour surfaces at the same angle have similar change characteristics in multiple frames of line laser data. Specifically, although the regularity of the contour surface of the rope zigzagging and laying on the ground is different from that of the low step, the contour height distribution characteristics of the rope zigzagging and laying on the ground at a specific angle collected by the robot are similar to the contour height distribution characteristics reflected from the surface of the low step collected when the robot walks along a specific oblique direction relative to the low step. However, at least it can be determined that the obstacles on the surface of the object to be measured are low obstacles, excluding the influence of high obstacles such as wall obstacles, and then the curve obstacles and the low step obstacles can be distinguished within the category of low obstacles.

[0035] Therefore, after determining that the size relationship between the number of repeated angles and the total number of occupied angles meets the preset screening condition, this embodiment determines whether there is a low-hill obstacle on the surface of the object to be measured by judging whether the two frames of line laser data collected in sequence have and only have laser slopes in the same direction. This overcomes the problem of misjudgment caused by the similarity between the contour height distribution characteristics of the rope recognized by the robot and the contour height distribution characteristics of the low-hill surface pointed out in the background technology.

[0036] As an example, Figure 8 As shown, the step 2 includes: Step 21, while the robot keeps walking along the navigation direction, the number of observation angle ranges repeatedly occupied by obstacle points marked in two frames of line laser data collected sequentially is counted to obtain the number of repeated angles, and between the two frames of line laser data collected sequentially, only one observation angle range that is repeatedly occupied is counted, so that the laser points marked as obstacle points are screened in the angle repetition area in the two frames of line laser data; and the number of observation angle ranges that are not repeatedly occupied by obstacle points marked in the two frames of line laser data collected sequentially is counted to obtain the total number of occupied angles, so that the laser points marked as obstacle points in different observation angle ranges are extracted non-repeatedly in the two frames of line laser data; then step 22 is executed; and the number of repeated angles and the total number of occupied angles are obtained by counting the situation that the laser points in the two frames of line laser data collected sequentially repeatedly occupy angles.

[0037] It should be noted that in step 1, the robot marks the laser points in each frame of line laser data whose height is greater than the reference height as obstacle points, which can be identified as obstacles by the robot, and it is necessary to decide whether to cross them to prevent collision or entanglement. In step 1, the robot controls the laser points in each frame of line laser data whose height is less than or equal to the reference height not to be marked as obstacle points, which can be identified as ground point clouds by the robot, allowing the robot to directly cross them and regard them as horizontal ground. The reference height is preferably 2mm or 5mm.

[0038] The observation angle range is an angle feature of a laser point, which represents the azimuth angle range formed relative to the center of the robot body and covering the corresponding obstacle contour, located Figure 2 The angle AOB is within the range and can be used Figure 7 A row of squares in represents. Figure 7 As shown, the squares in the first row represent the situation where the laser points occupying various observation angle ranges in the first frame line laser data are marked as obstacle points, and the squares in the second row represent the situation where the laser points occupying various observation angle ranges in the second frame line laser data are marked as obstacle points. The squares filled with black in the first row and the second row are aligned, and the squares filled with black in the third row represent the observation angle ranges occupied by the laser points marked as obstacle points that are extracted non-repeatedly from the first frame line laser data and the second frame line laser data, corresponding to Figure 7 The total number of occupied angles obtained in is equal to the value 20; the black filled squares in the fourth row represent the observation angle range occupied by the laser points marked as obstacle points in the angle repetition area in the first frame line laser data and the second frame line laser data, corresponding to Figure 7 The number of repetition angles obtained is equal to a value of 5.

[0039] Step 22, determine whether the size relationship between the number of repeated angles and the total number of occupied angles meets the preset screening condition, if yes, execute step 23, otherwise stop executing the low-hill obstacle identification method. This embodiment uses the angle number information in the two-frame line laser data to obtain the number of repeated angles and the total number of occupied angles. Based on this, it can preferentially identify walls or other large obstacles when the number of repeated angles and / or the total number of occupied angles is too large, because such large obstacles will occupy a larger angle range in the two-frame line laser data (not limited to whether they are repeatedly occupied).

[0040] When the rope (a curved obstacle) is placed on the ground, it is very low, and the laser point connection line formed by the projection is tortuous, which is not very obvious, and is not enough to be directly judged using the laser point connection feature. Otherwise, the recognition result is consistent with the low obstacle, rather than being identified as a curved obstacle. In addition, curved obstacles such as wires, network cables, and charging cables can be twisted and stacked to a height higher than 20mm, and are not identified as low obstacles, that is, they are not identified as low obstacles or curved obstacles, so the height or the laser slope formed by the laser point connection line is not directly used for the identification of low obstacles. Therefore, step 22 is required for preliminary screening before formally identifying low obstacles.

[0041] Step 23. In two frames of line laser data collected sequentially, determine whether each frame of line laser data has and only has laser slopes in the same direction. If so, the robot identifies the surface of the object to be measured as having a low-step obstacle and determines that the placement direction of the low-step obstacle is inclined at a certain angle to the navigation direction. Otherwise, the robot identifies the surface of the object to be measured as having a curved obstacle. In summary, based on the determination that the relationship between the number of repeated angles and the total number of occupied angles meets the preset screening conditions, the direction characteristics of the laser point connection lines of the same frame of line laser data and the consistency between the directions of the laser point connection lines of the two frames of line laser data are used to distinguish between low-step obstacles and curved obstacles, which is equivalent to identifying low-step obstacles after excluding the characteristics of curved obstacles.

[0042] On the basis of the above embodiment, in step 23, the method for determining whether two frames of line laser data collected sequentially have and only have laser slopes in the same direction includes: When the robot keeps walking along the navigation direction, in the two frames of line laser data collected sequentially, it is determined whether each frame of line laser data shows the presence of an ascending laser slope and no descending laser slope (equivalent to determining whether each laser point connection line in each frame of line laser data shows the presence of an ascending laser slope and no descending laser slope), or whether each frame of line laser data shows the presence of a descending laser slope and no ascending laser slope (equivalent to determining whether each laser point connection line in each frame of line laser data shows the presence of a descending laser slope and no ascending laser slope). When each frame of line laser data shows the presence of an ascending laser slope and no descending laser slope, or each frame of line laser data shows the presence of a descending laser slope and no ascending laser slope, the robot identifies the surface of the object to be measured as having the short-hill obstacle, and determines that the placement direction of the short-hill obstacle is at a certain inclination angle with the navigation direction. The sweeping robot uses the monotonous change characteristics of the lines connecting the laser points in each frame of line laser data to accurately identify the low-hill obstacles, so as not to avoid such obstacles and reduce the missed sweeping areas.

[0043] It should be noted that the laser slopes in the same direction are rising laser slopes or falling laser slopes. Figure 3 When the light strip 104 is projected onto the low-hill obstacle 105, a laser slope 1041 appears, and the line connecting the laser points belonging to the line segment 1041 is formed. Figure 4 When the light strip 104 is projected onto the low-hill obstacle 105, a laser slope 1042 appears, and a line connecting the laser points belonging to the line segment 1042 is formed. Figure 3 The laser ramp 1041 and Figure 4 The laser slope 1042 in the figure is a laser slope in the same direction. The clockwise direction is defined as an ascending laser slope, and the counterclockwise direction is defined as a descending laser slope. The connecting line of each laser point in the laser slope in the same direction shows a monotonically changing characteristic. The laser slope 1041 and the laser slope 1042 are both detected by the robot within a continuous period of time, and the position of the robot changes so that the laser slope 1041 and the laser slope 1042 both span multiple angles.

[0044] For identifying low-rise obstacles, the inclination direction of the laser slope is stable. When the robot approaches an obstacle, because the robot's walking direction does not change, if the laser slope shows an upward trend, the laser slope detected by the robot in the process of walking along the navigation direction will remain upward; if the laser slope shows a downward trend, the laser slope detected by the robot in the process of walking along the navigation direction will remain downward; anyway, there will not be both an upward and downward trend in the same frame of laser data. Therefore, the height feature of the laser point of the low-rise obstacle on the ground is linear, to distinguish it from the curved obstacle.

[0045] In some embodiments, when the magnitude relationship between the number of repeated angles and the total number of occupied angles satisfies a preset screening condition, there exists: If the robot collects one frame of line laser data each time it walks a distance step along the navigation direction, and after collecting two frames of line laser data in sequence, if the lines connecting the laser points in each frame of line laser data collected corresponding to the height less than or equal to the obstacle height threshold all have rising laser slopes and no falling laser slopes, then the robot identifies the surface of the object to be measured in front of it as a low-ridge obstacle, and determines that the placement direction of the low-ridge obstacle is at a certain inclination angle to the navigation direction, that is, the walking direction of the robot is oblique to the low-ridge obstacle, wherein the placement direction of the low-ridge obstacle is the extension direction of the longest contour of the low-ridge obstacle.

[0046] If the robot collects one frame of line laser data each time it walks a distance step along the navigation direction, and after collecting two frames of line laser data in sequence, if the lines connecting the laser points in each frame of line laser data collected corresponding to the height less than or equal to the obstacle height threshold all have descending laser slopes and no ascending laser slopes, then the robot identifies the surface of the object to be measured in front of it as having the low-ridge obstacle, and determines that the placement direction of the low-ridge obstacle is at a certain inclination angle to the navigation direction, that is, the walking direction of the robot is oblique to the low-ridge obstacle, wherein the placement direction of the low-ridge obstacle is the extension direction of the longest contour of the low-ridge obstacle.

[0047] Wherein, the reference height is less than the surmountable obstacle height threshold, and the surmountable obstacle height threshold is less than the installation height of the line laser sensor in the robot (preferably 45 mm). Each laser point whose height is less than or equal to the surmountable obstacle height threshold (preferably 20 mm) is marked as an obstacle point when the height is greater than the reference height (preferably 2 mm), and is connected to form a surmountable obstacle outline, which meets the setting requirements for the outline of the short-bridge obstacle. The short-bridge obstacle extends at the same height along the extension direction of its longest outline to form a surmountable and regular obstacle. The extension direction of the longest outline of the short-bridge obstacle is not parallel to the navigation direction, nor is it perpendicular to the navigation direction. The robot is diagonally facing the short-bridge obstacle during walking, and the extension direction of the longest outline of the short-bridge obstacle is inclined relative to the central axis of the robot body. In this scenario, the size relationship between the number of repeated angles and the total number of occupied angles meets the preset screening conditions, and the rising laser slope or the descending laser slope is detected.

[0048] Indicatively, Figure 3 This is a schematic diagram of a robot walking toward a low-ridge obstacle 105 in an embodiment of the present application. When the robot walks to the first position along the navigation direction OP, a line laser is projected onto the low-ridge obstacle 105; wherein, when the light bar 104 passes the low-ridge obstacle 105, a laser slope 1041 appears (defined as an ascending laser slope in the clockwise direction and a descending laser slope in the counterclockwise direction), and the extension direction of the low-ridge obstacle 105 forms a certain angle with the navigation direction OP but is not perpendicular to each other. The left and right sides of the laser slope 1041 are both regarded as straight laser lines perpendicular to the navigation direction OP, and the height of the laser point on each side remains unchanged, and is not configured as a laser slope.

[0049] Figure 4 The present application discloses a schematic diagram of a scenario in which a robot walks toward a low-step obstacle 105 and a line laser is projected onto the low-step obstacle 105 when the robot walks along the navigation direction OP to a second position. The second position is relative to Figure 3The first position is closer to the low-ridge obstacle 105, which can be understood as the robot walking a distance step from the first position along the navigation direction OP to the second position; wherein, when the light bar 104 passes the low-ridge obstacle 105, a laser slope 1042 appears (defined as an ascending laser slope in a clockwise direction and a descending laser slope in a counterclockwise direction), and the extension direction of the low-ridge obstacle 105 is at a certain angle to the navigation direction OP but not perpendicular to each other. The extension direction of the longest contour of the low-ridge obstacle 105 is at a certain angle to the navigation direction OP but not perpendicular to each other. The left and right sides of the laser slope 1042 are both regarded as straight laser lines perpendicular to the navigation direction OP, and the height of the laser point on each side remains unchanged and is not configured as a laser slope. The length of the straight laser line distributed on the left side of the laser slope 1042 is greater than Figure 3 The length of the straight laser line distributed on the left side of the laser slope 1041 is short, and the number of corresponding distributed laser points is reduced. The length of the straight laser line distributed on the right side of the laser slope 1042 is shorter than Figure 3 The length of the straight laser line distributed on the right side of the laser slope 1041 is long, and the number of correspondingly distributed laser points increases, indicating that the light bar 104 moves obliquely along the left end of the low-hill obstacle 105 .

[0050] In summary, the robot walks at least one distance step along the navigation direction, continuously collects two frames of line laser data, and then detects based on the two frames of line laser data that all the laser points whose height is less than or equal to the obstacle height threshold that can be crossed have an ascending laser slope and no descending laser slope, or detects that all the laser points whose height is less than or equal to the obstacle height threshold that can be crossed have an ascending laser slope and no descending laser slope. Therefore, when the embodiment judges the short-hill obstacle based on two consecutive frames of line laser data, there must be an ascending laser slope in two consecutive frames and no descending laser slope in the two frames, or there must be a descending laser slope in two consecutive frames and no ascending laser slope in the two frames. In this way, the height change characteristics of each laser point within a specific area are used to reduce the probability of misidentification, and the recognition accuracy of the short-hill obstacle is improved without simply using the height of a single laser point for comparison and judgment.

[0051] In some embodiments, when the size relationship between the number of repeated angles and the total number of occupied angles satisfies a preset screening condition, while the robot keeps walking along the navigation direction, in the two frames of line laser data collected sequentially, it is judged that the line connecting the laser points in at least one frame of line laser data presents an ascending laser slope and a descending laser slope, then it is determined that the robot has detected two laser slopes with different directions in at least one frame of line laser data, and that the robot has identified the surface of the object to be measured as having the curved obstacle, and that the heights of the laser points in the two frames of line laser data collected sequentially are all less than or equal to the traversable obstacle height threshold, the reference height is less than the traversable obstacle height threshold, and the traversable obstacle height threshold is less than the installation height of the line laser sensor in the robot.

[0052] Indicatively, Figure 5 An embodiment of the present application discloses a scenario in which a robot walks toward a curved obstacle 106. When the robot walks to a first position along a navigation direction OP, a line laser is projected onto the curved obstacle 106. The curved obstacle 106 is placed horizontally on the walking surface of the robot in a curved and uneven shape. Figure 6 An embodiment of the present application discloses a scenario in which a robot walks toward a curved obstacle 106. When the robot walks to a second position along a navigation direction OP, a line laser is projected onto the curved obstacle 106. The curved obstacle 106 is placed horizontally on the walking surface of the robot in a curved and uneven shape. The second position is relative to Figure 5 The first position of the robot is closer to the curved obstacle 106, which can be understood as the robot walking a distance step from the first position along the navigation direction OP to the second position; Figure 6 The black bold rectangle generated at the intersection of the middle light bar 104 and the curved obstacle 106 is Figure 5 There are many black bold rectangles at the intersection of the middle light bar 104 and the curved obstacle 106, especially on the right side of the light bar 104. Figure 5 and Figure 6 On the straight line where the light bar 104 is located, there are convex points with an ascending trend and convex points with a descending trend. The convex points with an ascending trend indicate the direction in which the light spot deviates upward after being projected on the curved obstacle 106, and the convex points with a descending trend indicate the direction in which the light spot deviates downward after being projected on the curved obstacle 106; thereby reflecting the morphological distribution characteristics of the curved obstacle 106, such as the upper and lower sides of the zigzag line segment 104.

[0053] The curved obstacles identified in this embodiment are mainly lines lying across the floor, and need to be judged by combining the previous and next two frames of line laser data. Specifically, the curved obstacles are continuous and curved and laid flat. For example, when the robot walks along the navigation direction at a certain distance step, it collects the first frame of line laser data when it moves to the first position, and extracts a part of the rope. After moving a distance step, it reaches the second position and collects the second frame of line laser data, and extracts other parts of the rope. The two parts extracted successively by the robot may have overlapping areas. It is judged that the size relationship between the number of repeated angles and the total number of occupied angles meets the preset screening condition, and then it is judged that at least one frame When the lines connecting the laser points in the line laser data show an ascending laser slope and a descending laser slope, the curved obstacle is identified. The main characteristics of the curved obstacle are that it is short (generally not higher than 20 mm), continuous, and curved to the horizontal ground. If the height feature of the laser point is directly used for identification, since the height is small and the light strip formed by the line laser emission is a straight line, it is not enough to form an obvious recognition effect at one time. Therefore, this embodiment only uses two frames of line laser data collected in sequence to detect the curved obstacle based on the consideration of the calculation amount and the walking distance of the robot, which saves the calculation amount and prevents the robot from crossing the obstacle or getting stuck before the type of the obstacle is identified due to the excessive walking distance.

[0054] As an embodiment, the robot divides the preset detection angle range of the line laser sensor into the number of observation angle ranges for target detection in sequence according to the predetermined clock direction, and each observation angle range can be an angle interval of equal arc size evenly distributed from the preset detection angle range; between two observation angle ranges divided successively, the maximum angle value of the first observation angle range is the minimum angle value of the second observation angle range; the preset detection angle range of the line laser sensor forms a fixed angle range on both sides of the central axis of the body. In this embodiment, for the same frame of line laser data, each observation angle range remains unchanged while the robot walks along the navigation direction, and the observation angle range has nothing to do with the angle formed by the laser point actually reflected back from the surface of the object to be measured relative to the body center of the robot, because in the process of the robot approaching the surface of the object to be measured, the angle formed by a position on the surface of the object to be measured relative to the gradually approaching body center is changing, and the observation angle range is divided in advance.

[0055] Figure 2An embodiment of the present application discloses a schematic diagram of a detection area 103 of a line laser sensor 102 in front of a robot 101 and a preset detection angle range of the line laser sensor 102. The preset detection angle range is an angle AOB with the body center O as the vertex, and may be symmetrical about the navigation direction OP, that is, the angle of the angle AOP is equal to the angle of the angle BOP; within the detection range 103, the line segment 104 is a light strip AB formed in front of the robot by the line laser emitted by the line laser sensor; the preset detection angle range (angle AOB with the body center O as the vertex) is preferably 60 degrees.

[0056] If a frame of line laser data collected by the line laser sensor is divided into two parts, the left and right parts have approximately 60 laser points respectively, and the index values ​​from left to right are 0 to 60 respectively. The inclusion here does not represent the occupied angle range.

[0057] If the angle where the ray OP is located is set to 0 degrees and measured in a counterclockwise direction, starting from the middle, the angle range of the left half of the angle AOB is from 0 to 30 degrees, and the angle range on the right is from 330 degrees to 360 degrees and then back to 0 degrees (360 degrees). The preset detection angle range is the line laser coverage angle formed relative to the body center O of the robot to cover the laser point at the corresponding angle, not the angle formed relative to the line laser sensor 102. However, there is a large amount of angle conversion in the calculation and detection of the two angle ranges from 0 to 30 degrees and from 330 to 360 degrees. Therefore, in this embodiment, the preset detection angle range is divided into target detection number observation angle ranges according to the established clockwise direction. For example, the established clockwise direction is preferably counterclockwise, the preset detection angle range is 60 degrees, the target detection number is preferably 60, and the angle span of each observation angle range is 1 degree. According to the counterclockwise direction, the 60-degree angle AOB is divided into 60 observation angle ranges from right to left, and the observation angle ranges of 0 to 1 degree, 1 to 2 degrees, 2 to 3 degrees, ..., 59 to 60 degrees are measured in turn. It can be understood that the angle step required for measurement is 1 degree; then there are 120 laser points in a frame of line laser data collected by the line laser sensor, and the 120 laser points can be distributed in 60 observation angle ranges, and there is at least 1 laser point in one observation angle range.

[0058] Then an array can be created, each element of which indicates whether a corresponding observation angle range is occupied by an obstacle point. After the laser points falling into the corresponding observation angle range are selected as the laser points occupying the corresponding observation angle range, their angle values ​​are rounded.

[0059] In a frame of line laser data, whenever there is a laser point with a height greater than 2 mm, the robot configures the obstacle mark logic information corresponding to the laser point to logic 1 and marks the laser point as an obstacle point, so that the robot knows the number of obstacle points in a frame of line laser data.

[0060] Corresponding to Figure 7 In the embodiment, an observation angle range is represented by a square in a row where frame1 or frame2 is located, and the laser points occupying the corresponding observation angle range can also be understood as the laser points occupying the square. In some embodiments, the observation angle range of 0 degree to 1 degree is represented by a square in a row where frame1 or frame2 is located. Figure 7 The obstacle point is represented by the leftmost square in the row where frame1 or frame2 is located; the obstacle point is represented by a black-filled square in the row where frame1 or frame2 is located. A square can be understood as a bit in a frame of line laser data, so as to facilitate logical calculation using obstacle marking logic information. Figure 7 The squares in each column represent the observation angle range of the same angle value; there are 60 squares in each row, representing 60 observation angle ranges; the marking of the squares in the row where frame1 is located reflects Figure 5 The obstacle marking situation generated by the line laser sensor detecting the curved obstacle 106, and the marking situation of the squares in the row where frame2 is located reflects Figure 6 The obstacle marking situation generated by the line laser sensor detecting the curved obstacle 106, obviously, the squares in the row where frame2 is located are filled with more black squares than the squares in the row where frame1 is located, especially on the right side of a row of squares, which is consistent with Figure 6 The longer black bold rectangle on the right side of the middle light strip 104 is adapted to the light spot, that is, a longer light spot deviation phenomenon is generated.

[0061] It should be noted that the multiple laser points collected by the line laser sensor generally fall into an observation angle range. Note that the prior art uses "fall into" to indicate the angular distribution characteristics of the laser points, while the present application uses "occupy" to indicate that a laser point occupies an observation angle range, or is understood as a laser point occupying / falling into an observation angle. Specifically, the present embodiment selects a laser point with the smallest angle value from the multiple laser points that fall into the current observation angle range and configures it as the laser point occupying the current observation angle range. In a scenario where the 120 laser points in a frame of line laser data are distributed in a range of 60 observation angles, the points are rotated counterclockwise. Figure 2The angle AOB shown is configured to be in the range of 0 to 60 degrees and the angle step of each observation angle range is 1 degree. Assuming that there are three laser points "falling into" the observation angle range of 1 to 2 degrees, they "fall into" 1.2 degrees, 1.4 degrees, and 1.8 degrees respectively. When this embodiment traverses to the laser point at 1.2 degrees in a counterclockwise direction, the laser point at 1.2 degrees is regarded as a laser point occupying the observation angle range of 1 to 2 degrees, and can also be regarded as a representative point at the observation angle of 1 degree, while the laser points at angles 1.4 and 1.8 with larger angle values ​​are ignored. On this basis, the obstacle points marked in the same frame line laser data are configured not to repeatedly occupy the same observation angle range, so that only one obstacle point is marked in one observation angle range.

[0062] On the basis of step 2 or step 23 disclosed in the above embodiment, the method for determining whether a frame of line laser data has an ascending laser slope or a descending laser slope includes: In the process of the robot collecting the one frame of line laser data, the robot will record the divided observation angle ranges and the height of the laser points occupying the observation angle ranges in sequence according to the predetermined clock direction. Then, in the one frame of line laser data, the robot will detect the height changes of the laser points occupying the corresponding observation angle ranges in sequence according to the predetermined clock direction, mainly to capture the height change characteristics of the laser points in the section where the laser slope exists.

[0063] In the frame of line laser data collected by the robot, it is detected whether there is an increase in height of laser points occupying at least three observation angle ranges that are continuously distributed. If so, it is determined that there is a rising laser slope in the frame of line laser data; otherwise, it is determined that there is no rising laser slope in the frame of line laser data; specifically, if an increase in height of laser points occupying at least three observation angle ranges that are continuously distributed is detected along the predetermined time needle direction, it is determined that there is an ascending laser slope on the line connecting each laser point in the frame of line laser data collected by the robot, which can be understood as the laser points or obstacle points occupying at least three observation angle ranges that are continuously distributed in the frame of line laser data are connected in sequence to form an ascending laser slope; if no increase in height of laser points occupying at least three observation angle ranges that are continuously distributed is detected along the predetermined time needle direction, it is determined that there is no ascending laser slope on the line connecting each laser point in the frame of line laser data collected by the robot.

[0064] It should be noted that the rising laser slope is represented by connecting the laser points rising along the predetermined clock direction in sequence. In this embodiment, the heights of the laser points occupying at least three observation angle ranges of continuous distribution increase gradually, corresponding to Figure 7In the example, there are at least three consecutive squares. In the predetermined clock direction, except for the first square, the height of the laser point corresponding to each square (or the laser point occupying the square) is greater than the height of the laser point corresponding to the previous square. Then, the adjacent laser points occupying the corresponding squares are connected in sequence to form an ascending laser slope, and the slope rises along the predetermined clock direction. If the predetermined clock direction is clockwise, a laser slope can be formed. Figure 5 Middle slope 1041 or Figure 6 The slope in 1042.

[0065] In the frame of line laser data collected by the robot, it is detected whether there is a decrease in height of laser points occupying at least three observation angle ranges that are continuously distributed. If so, it is determined that there is a descending laser slope in the frame of line laser data; otherwise, it is determined that there is no descending laser slope in the frame of line laser data. Specifically, if a decrease in height of laser points occupying at least three observation angle ranges that are continuously distributed is detected along the predetermined clock direction, it is determined that there is a descending laser slope on the line connecting each laser point in the frame of line laser data collected by the robot, which can be understood as the laser points or obstacle points occupying at least three observation angle ranges that are continuously distributed in the frame of line laser data are connected in sequence to form a descending laser slope. If no decrease in height of laser points occupying at least three observation angle ranges that are continuously distributed is detected along the predetermined clock direction, it is determined that there is no descending laser slope on the line connecting each laser point in the frame of line laser data collected by the robot.

[0066] It should be noted that the descending laser slope is represented by connecting the laser points descending along the predetermined clock direction in sequence. In this embodiment, the heights of the laser points occupying at least three observation angle ranges of continuous distribution decrease, corresponding to Figure 7 In the given clockwise direction, there are at least three consecutive squares. In the given clockwise direction, except for the first square, the height of the laser point corresponding to each square (or the laser point occupying the square) is smaller than the height of the laser point corresponding to the previous square. Then, the adjacent laser points occupying the corresponding squares are connected in sequence to form a descending laser slope, and the slope descends along the given clockwise direction. If the given clockwise direction is counterclockwise, a descending laser slope can be formed. Figure 5 Middle slope 1041 or Figure 6 The slope in 1042.

[0067] To summarize, in the two frames of line laser data collected sequentially, for each frame of line laser data, the robot detects that there is an increase in the height of the laser points occupying at least three observation angle ranges of continuous distribution, and detects that there is no decrease in the height of the laser points occupying at least three observation angle ranges of continuous distribution, then it is determined that the lines connecting the laser points in this frame of line laser data all show that there is an ascending laser slope and no descending laser slope. At this time, the robot identifies the surface of the object to be measured as having the low-step obstacle, and determines that the placement direction of the low-step obstacle is at a certain inclination angle with the navigation direction.

[0068] In the two frames of line laser data collected sequentially, for each frame of line laser data, the robot detects that there is no increase in the height of the laser points occupying at least three observation angle ranges of continuous distribution, and detects that there is a decrease in the height of the laser points occupying at least three observation angle ranges of continuous distribution, then it is determined that the lines connecting the laser points in the frame of line laser data all show that there is a descending laser slope and there is no ascending laser slope. At this time, the robot recognizes the surface of the object to be measured as having the short-hill obstacle, and determines that the placement direction of the short-hill obstacle is at a certain inclination angle with the navigation direction. Therefore, the robot uses the monotonous change characteristics of the lines connecting the laser points in each frame of line laser data to accurately identify the short-hill obstacles, so as not to avoid such obstacles and reduce the missed traversal areas.

[0069] As an embodiment, in step 22, the method for judging whether the size relationship between the number of repeated angles and the total number of occupied angles meets the preset screening condition includes: when the total number of occupied angles is greater than or equal to the first preset observation angle number threshold, and the number of repeated angles is less than the ratio of the total number of occupied angles to the preset reduction coefficient, determining that the size relationship between the number of repeated angles and the total number of occupied angles meets the preset screening condition, and then executing step 23; otherwise, determining that the size relationship between the number of repeated angles and the total number of occupied angles does not meet the preset screening condition. Thus, some obstacles such as walls are excluded.

[0070] In this embodiment, in the required identification of the surface of the object to be measured, it is necessary to distinguish between regular crossable obstacles and winding and continuous ropes (belonging to curve obstacles). In order to reduce the misjudgment problem caused by the single comparison and analysis of the height information of the laser point, this embodiment can eliminate the influence of wall obstacles by combining the number of obstacle points that repeatedly occupy the same observation angle range in the two frames of line laser data. Moreover, this is performed during the walking process of the robot. The line laser detects two different areas of the surface of the object to be measured at the same position as the walking position of the robot changes. For example, in the two frames of line laser data collected sequentially, the first frame of line laser data is reflected from the contour of a part of the angle interval of the curve obstacle, and the second frame of line laser data is reflected from the contour of another part of the angle interval of the curve obstacle. There are repeatedly detected contours in the two parts of the angle interval. Corresponding to step 22, the number of repeated angles and the total number of occupied angles are the results of logical analysis and processing of the number of obstacle points in the two-frame line laser data, including setting a first preset observation angle number threshold and a preset reduction coefficient, and requiring the number of repeated angles and the total number of occupied angles to be limited to corresponding numerical ranges, respectively, so as to exclude large contour surface obstacles such as wall obstacles (obviously not rope obstacles) in advance for identifying curved obstacles.

[0071] Specifically, the obstacle points that appear repeatedly at the same angle are located in the overlapping area of ​​the two different areas. In this embodiment, the number of repeated angles can be expressed by the situation that the laser points in the two frames of line laser data collected sequentially repeatedly occupy angles; preferably, the more the number of obstacle points that appear repeatedly at the same angle, the larger the size of the obstacles on the surface of the object to be measured, including the larger the contour surface and the larger the height. For example, large obstacles such as wall obstacles need to be excluded first. Therefore, after obtaining the number of repeated angles and the total number of occupied angles, it is determined whether the total number of occupied angles is greater than or equal to the first preset observation angle number threshold, and it is also determined whether the number of repeated angles is less than the ratio of the total number of occupied angles to the preset reduction coefficient, so as to determine whether the preset screening conditions are met, determine the change characteristics of the laser points in the two frames of line laser data reflected from the surface of the object to be measured as the robot walks, and then determine the height distribution characteristics of the laser points in the two frames of line laser data collected sequentially and the distribution characteristics of the obstacle contours.

[0072] In the above embodiment, it should also be noted that before executing the step 21, it also includes counting the number of observation angle ranges occupied by laser points marked as obstacle points in a frame of line laser data collected in step 1; thus, before formally counting the number of relevant observation angle ranges in step 21, it is necessary to confirm whether the two frames of line laser data collected in sequence in step 1 are legal through the information on the number of angles occupied by the obstacle points.

[0073] Therefore, in a frame of line laser data collected in step 1, if the number of observation angle ranges occupied by the detected obstacle points is greater than or equal to the second preset observation angle number threshold, and the number of observation angle ranges occupied by the obstacle points is less than or equal to the third preset observation angle number threshold, then the frame of line laser data collected in step 1 is marked as a legal data frame; further, if two frames of legal data frames are detected successively, the two frames of legal data frames are marked as two frames of line laser data required to be collected for counting the number of repeated angles and the total number of occupied angles; wherein the legal data frames detected successively can be two frames of line laser data collected sequentially in step 1. Then, the two frames of legal data frames are updated to the two frames of line laser data collected sequentially in step 21, and step 21 is executed again. Illegal data frames are eliminated in advance, the processing amount of line laser data is reduced, and the recognition accuracy of the curve obstacles and the low-hill obstacles is improved.

[0074] Among them, the second preset observation angle number threshold is less than the first preset observation angle number threshold, which means that the number of observation angle ranges occupied by obstacle points in a frame of line laser data is too small and obviously not the curve obstacle or the low-hill obstacle; the first preset observation angle number threshold is less than the third preset observation angle number threshold. Of course, if the number of observation angle ranges occupied by obstacle points in a frame of line laser data is too large, it is obviously not the curve obstacle. For example, a rope laid flat on the ground generally does not occupy a large angle range in a frame.

[0075] Preferably, the number of target detections is 60, that is, the number of observation angle ranges that can be divided in a frame of line laser data can reach 60; the first preset observation angle number threshold is 20, and the preset reduction factor is 1.5; the second preset observation angle number threshold is 8, and the third preset observation angle number threshold is 30. Therefore, in order to eliminate the interference of obstacles other than the curve obstacles and the low-hill obstacles, corresponding to Figure 7 In the first frame of line laser data frame1 (represented by each square in the row where frame1 is located) and the second frame of line laser data frame2 (represented by each square in the row where frame2 is located), the laser point in each frame of line laser data needs to occupy more than or equal to 8 black-filled squares and less than 30 black-filled squares to form a legal data frame. At the same time, in order to meet the preset screening conditions, in the two frames of legal data collected in sequence, the total number of occupied angles needs to be greater than or equal to 20, and the number of repeated angles needs to be less than the ratio of the total number of occupied angles to 1.5. Therefore, this preferred example improves the recognition accuracy of the curve obstacles and the low-hill obstacles to a reasonable level.

[0076] Figure 7 An embodiment of the present application discloses a robot in Figure 5 A frame of line laser data collected in Figure 6 Schematic diagram of the logic processing of a frame of line laser data collected in the image according to the detection angles divided in sequence within the preset detection angle range, where the row of squares where frame1 is located represents the robot Figure 5 The first row of squares in the figure represents the observation angle range occupied by the corresponding laser point. The squares in the first row represent the situation where each laser point in the first frame of line laser data collected is marked as an obstacle point; the squares in the row where frame2 is located represent the situation where the robot is in Figure 6 The one-frame line laser data collected in the figure is regarded as the observation angle range occupied by the corresponding laser point represented by each square in the second row of the diagram, and the squares in the second row represent the obstacle marking logic information corresponding to each laser point in the one-frame line laser data currently collected; the laser points occupying the white squares are not marked as obstacle points, and the laser points occupying the black-filled squares are marked as obstacle points. There are 8 black-filled squares in the row of squares where frame1 is located, that is, there are 8 squares occupied by laser points with the obstacle marking logic information set to logic 1, and the one-frame line laser data collected previously is a legal data frame. There are 17 black-filled squares in the row of squares where frame2 is located, that is, there are 17 squares occupied by laser points with the obstacle marking logic information set to logic 1, and the one-frame line laser data currently collected is a legal data frame.

[0077] The row of squares where frame1 is located indicates that the robot is Figure 5 A frame of line laser data collected in the diagram is considered to be represented by the first row of squares representing the observation angle range occupied by the corresponding laser point. The squares in the first row that are arranged continuously and filled with black represent Figure 5 The observation angle range occupied by the black bold rectangle (which can be regarded as a black convex dot) generated at the intersection of the middle light bar 104 and the curved obstacle 106 can be represented by at least one square. The longer the length of the black bold rectangle in the vertical direction of the navigation direction OP is, the more black-filled squares are represented and the more they are arranged continuously.

[0078] The row of squares where frame2 is located indicates that the robot is Figure 6 A frame of line laser data collected in the figure is regarded as each square in the second row of the diagram representing the observation angle range occupied by the corresponding laser point, and the squares in the first row of the grid that are arranged continuously and filled with black represent Figure 6The observation angle range occupied by the black bold rectangle (which can be regarded as a black convex dot) generated at the intersection of the middle light bar 104 and the curved obstacle 106 can be represented by at least one square. The longer the length of the black bold rectangle in the vertical direction of the navigation direction OP is, the more black-filled squares are represented and the more they are arranged continuously. Figure 6 The black bold rectangle generated at the intersection of the middle light bar 104 and the curved obstacle 106 is Figure 5 There are many black bold rectangles generated at the intersection of the light bar 104 and the curved obstacle 106, especially on the right side of the light bar 104. Figure 7 In the example, the number of the rightmost consecutive squares filled with black in the second row of squares is at least 13, and the number of the rightmost consecutive squares filled with black in the first row of squares is 2.

[0079] As an embodiment, in step 21, the method for counting the number of observation angle ranges repeatedly occupied by obstacle points marked in two frames of linear laser data collected sequentially includes: The robot divides the two frames of line laser data collected in sequence into a frame of line laser data collected previously and a frame of line laser data collected currently; for each frame of line laser data, the robot divides the preset detection angle range of the line laser sensor into the number of observation angle ranges for target detection in advance according to the established clock direction, and the angle step involved in each observation angle range is equal; in the same frame of line laser data, each observation angle range remains unchanged while the robot walks along the navigation direction; a laser point with the smallest angle value is selected from multiple laser points falling into the current observation angle range and configured as the laser point occupying the current observation angle range. Among them, the robot sets the obstacle marking logic information corresponding to the laser point marked as an obstacle point in each frame of line laser data to logic 1; the robot sets the obstacle marking logic information corresponding to the laser point not marked as an obstacle point in each frame of line laser data to logic 0.

[0080] The robot sequentially controls the obstacle marking logic information corresponding to each laser point in a frame of line laser data collected previously and the obstacle marking logic information corresponding to the laser points occupying the same observation angle range in a frame of line laser data collected currently to perform a logic AND operation, and obtains the obstacle marking logic operation results corresponding to the laser points occupying each observation angle range, and then counts the number of observation angle ranges occupied by the laser points whose obstacle marking logic operation results are logic 1, and then sets the counted number of observation angle ranges as the repeated angle number, so that between two frames of line laser data collected sequentially, the repeatedly occupied observation angle range is counted only once.

[0081] exist Figure 7In the figure, the black-filled squares in the first row and the second row are aligned as the obstacle mark logic information with a logic value of 1 corresponding to the laser points occupying the same observation angle range participating in the logic AND operation in the two frame line laser data; after the obstacle mark logic information corresponding to the squares in the first row and the second row is logically ANDed, the obtained logic AND operation result is represented by the squares in the row where frame1&frame2 are located, and in the squares in the fourth row of the figure, the obstacle mark logic operation result corresponding to the laser points at each observation angle range is obtained, and the black-filled squares in the fourth row of squares represent the observation angle range occupied by the laser points marked as obstacle points in the repeated area of ​​the observation angle range in the first frame line laser data (the black-filled squares in the first row) and the second frame line laser data (the black-filled squares in the second row); since the number of black-filled squares in the fourth row of squares is equal to 5, the corresponding number of repeated angles is equal to the value 5, which is equal to the number of observation angle ranges occupied by the laser points whose obstacle mark logic operation result is logic 1. The bitwise "AND" logic operation is used to count the number of obstacle points in the angular overlap area of ​​two adjacent frames of laser data.

[0082] As an embodiment, in step 21, the number of observation angle ranges that are not repeatedly occupied by obstacle points marked in two frames of linear laser data collected sequentially is counted, and the method for obtaining the total number of occupied angles includes: The robot divides the two frames of line laser data collected in sequence into a frame of line laser data collected previously and a frame of line laser data collected currently; for each frame of line laser data, the robot divides the preset detection angle range of the line laser sensor into the number of observation angle ranges for target detection in advance according to the established clock direction, and the angle value of each observation angle range is equal; the preset detection angle range of the line laser sensor forms a fixed angle range on both sides of the robot's body axis; in the same frame of line laser data, each observation angle range remains unchanged while the robot walks along the navigation direction; a laser point with the smallest angle value is selected from multiple laser points falling into the current observation angle range and configured as the laser point occupying the current observation angle range. Among them, the robot sets the obstacle marking logic information corresponding to the laser point marked as an obstacle point in each frame of line laser data to logic 1; the robot sets the obstacle marking logic information corresponding to the laser point not marked as an obstacle point in each frame of line laser data to logic 0.

[0083] The robot sequentially controls the obstacle marking logic information corresponding to each laser point in the first frame of line laser data collected previously and the obstacle marking logic information corresponding to the laser points occupying the same observation angle range in the currently collected frame of line laser data to perform a logical OR operation, and obtains the obstacle marking logic operation results corresponding to the laser points occupying each observation angle range, and then counts the number of observation angle ranges occupied by the laser points whose obstacle marking logic operation results are logic 1, and then sets the counted number of observation angle ranges as the total number of occupied angles, forming the number of observation angle ranges that are not repeatedly occupied by obstacle points in the two frames of line laser data collected sequentially, which is equivalent to the result obtained by subtracting the aforementioned repeated angle number (the number of observation angle ranges that have been repeatedly occupied by laser points whose obstacle marking logic information is set to logic 1) from the total number of observation angle ranges occupied by all laser points in the two frames of line laser data collected sequentially.

[0084] exist Figure 7 In the figure, the black-filled squares in the first and second rows are aligned as the obstacle marking logic information with a logic value of 1 corresponding to the laser points occupying the same observation angle range participating in the logic OR operation in the two frames of line laser data; after the obstacle marking logic information corresponding to the squares in the first and second rows are logically ORed, the obtained logic AND operation result is represented by the squares in the row where frame1|frame2 is located, and in the squares in the third row of the figure, the obstacle marking logic operation result corresponding to the laser points at each observation angle range is obtained, and the black-filled squares in the third row of squares represent the observation angle range occupied by all the laser points marked as obstacle points in the first frame of line laser data (black-filled squares in the first row) and the second frame of line laser data (black-filled squares in the second row); since the number of black-filled squares in the third row of squares is equal to 20, the corresponding total number of occupied angles is equal to the value 20. The bitwise "OR" logic operation is used to count the number of observation angle ranges that are not repeatedly occupied by the obstacle points marked in two adjacent frames of laser data.

[0085] It is worth noting that after the line laser sensor detects the wall, the number of obstacle points counted will account for the majority after the logical AND operation between the two frames of line laser data collected in sequence and the corresponding obstacle marking logic information is required. Therefore, combined with the number of obstacle points in the angle overlap area of ​​the two adjacent frames of laser data obtained by the bitwise "AND" logic operation in the aforementioned embodiment, the influence of walls or other large obstacles (such as rope obstacles zigzagging and flattened on the ground) is excluded.

[0086] Based on the above embodiments, Figure 7In the two frames of line laser data collected sequentially, the total number of occupied angles is greater than or equal to 20, and the number of repeated angles (equal to 5) needs to be less than the ratio (greater than 10) of the total number of occupied angles (equal to 20) to a preset reduction factor (preferably 1.5), and the preset screening conditions are met. Then, the curve obstacle and the low-hill obstacle can be distinguished by executing step 23, or the low-hill obstacle can be identified in step 2 in combination with the trend characteristics of the laser slope in each frame of line laser data.

[0087] The above is a detailed introduction to an embodiment of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope.

Claims

1. A method for identifying low-rise obstacles based on line laser, wherein a line laser sensor is installed on the front side of the robot body; characterized in that: The method for identifying low-hill obstacles comprises: Step 1: When the robot is walking along the navigation direction, the line laser sensor is controlled to emit a line laser obliquely downward along the navigation direction, and two frames of line laser data formed by the line laser being reflected by the surface of the object to be measured are collected in sequence; Step 2: In two frames of line laser data collected sequentially, the number of repeated angles and the total number of occupied angles are obtained by counting the repeated angles occupied by obstacle points in the two frames of line laser data collected sequentially. When the relationship between the number of repeated angles and the total number of occupied angles meets the preset screening conditions, the surface of the object to be measured is identified as having a low-hill obstacle based on the direction characteristics of the laser slope in each frame of line laser data.

2. The method for identifying low-rise obstacles according to claim 1, characterized in that: In step 2, the method of identifying the surface of the object to be measured as having a low-hill obstacle by combining the trend characteristics of the laser slope in each frame of line laser data includes: Determine whether the two frames of line laser data collected sequentially have and only have laser slopes in the same direction. If so, the robot identifies the surface of the object to be measured as having a low-step obstacle and determines that the placement direction of the low-step obstacle is at a certain inclination angle with the navigation direction. Otherwise, the robot identifies the surface of the object to be measured as having a curved obstacle. The navigation direction is perpendicular to the extension direction of the light bar formed by the projection of the line laser sensor, so that the light bar is not parallel to the placement direction of the low-hill obstacle.

3. The method for identifying low-rise obstacles according to claim 2, characterized in that: The step 2 comprises: Step 21, while the robot keeps walking along the navigation direction, the number of observation angle ranges repeatedly occupied by obstacle points marked in two frames of linear laser data collected sequentially is counted to obtain the number of repeated angles; and the number of observation angle ranges not repeatedly occupied by obstacle points marked in two frames of linear laser data collected sequentially is counted to obtain the total number of occupied angles; then step 22 is executed; and the number of repeated angles and the total number of occupied angles are obtained by counting the number of angles repeatedly occupied by obstacle points in two frames of linear laser data collected sequentially; Step 22, determining whether the magnitude relationship between the number of repeated angles and the total number of occupied angles meets a preset screening condition, if yes, executing step 23; Step 23, in the two frames of line laser data collected sequentially, determine whether each frame of line laser data has and only has a laser slope in the same direction, if so, the robot recognizes the surface of the object to be measured as having the short hill obstacle, otherwise, the robot recognizes the surface of the object to be measured as having the curved obstacle; Among them, in step 1, the robot marks the laser points in each frame of line laser data whose height is greater than the reference height as obstacle points, and in step 1, the robot controls the laser points in each frame of line laser data whose height is less than or equal to the reference height not to be marked as obstacle points.

4. The method for identifying low-rise obstacles according to claim 3, characterized in that: In step 23, the method for determining whether two frames of line laser data collected sequentially have and only have laser slopes in the same direction includes: While the robot keeps walking along the navigation direction, in the two frames of line laser data collected sequentially, it is determined whether each frame of line laser data shows an ascending laser slope and no descending laser slope, or whether each frame of line laser data shows a descending laser slope and no ascending laser slope; The laser slopes in the same direction are rising laser slopes or falling laser slopes.

5. The method for identifying low-rise obstacles according to claim 4, characterized in that: When the magnitude relationship between the number of repeated angles and the total number of occupied angles satisfies a preset screening condition, there exists: If the robot collects one frame of line laser data each time it walks a distance step along the navigation direction, and after collecting two frames of line laser data in sequence, if the lines connecting the laser points whose heights are less than or equal to the obstacle height threshold that can be crossed in each frame of line laser data collected have rising laser slopes and no falling laser slopes, then the robot recognizes the surface of the object to be measured in front of it as having the short-hill obstacle; If the robot collects one frame of line laser data each time it walks a distance step along the navigation direction, and after collecting two frames of line laser data in sequence, if the lines connecting the laser points whose heights are less than or equal to the obstacle height threshold that can be crossed in each frame of line laser data collected have descending laser slopes and no ascending laser slopes, then the robot recognizes the surface of the object to be measured in front of it as having the short-hill obstacle; The short-ridge obstacles are extended at the same height along the extension direction of their longest contours to form a crossable and regular obstacle; the placement direction of the short-ridge obstacles is the extension direction of the longest contour of the short-ridge obstacles; The reference height is less than a traversable obstacle height threshold, and the traversable obstacle height threshold is less than an installation height of the line laser sensor in the robot.

6. The method for identifying low-rise obstacles according to claim 4, characterized in that: The robot divides the preset detection angle range of the line laser sensor into target detection number observation angle ranges in sequence according to the predetermined clock direction; wherein the preset detection angle range of the line laser sensor forms a fixed angle range on both sides of the robot's body central axis; In the same frame of line laser data, each observation angle range remains unchanged while the robot walks along the navigation direction; a laser point with the smallest angle value is selected from multiple laser points falling within the current observation angle range and configured as the laser point occupying the current observation angle range.

7. The method for identifying low-rise obstacles according to claim 6, characterized in that: In step 23, the method for determining whether a rising laser slope or a falling laser slope exists in a frame of line laser data includes: In the one frame of line laser data, the robot sequentially detects the height changes of the laser points occupying the corresponding observation angle range according to the predetermined clock direction; Detecting whether there is a laser point with a height increase of at least three observation angle ranges that are continuously distributed, and if so, determining that there is a rising laser slope in the one frame of line laser data, otherwise determining that there is no rising laser slope in the one frame of line laser data; the rising laser slope is represented by connecting the laser points that rise along the predetermined clock direction in sequence; Detect whether there is a decreasing height of laser points occupying at least three observation angle ranges of continuous distribution, and if so, determine that there is a descending laser slope in the one frame line laser data, otherwise determine that there is no descending laser slope in the one frame line laser data; the descending laser slope is represented by connecting the laser points descending along the predetermined clock direction in sequence.

8. The method for identifying low-rise obstacles according to claim 6, characterized in that: In step 22, the method for determining whether the magnitude relationship between the number of repeated angles and the total number of occupied angles satisfies a preset screening condition includes: When the total number of occupied angles is greater than or equal to a first preset observation angle number threshold, and the number of repeated angles is less than the ratio of the total number of occupied angles to a preset reduction coefficient, it is determined that the size relationship between the number of repeated angles and the total number of occupied angles meets a preset screening condition.

9. The method for identifying low-rise obstacles according to claim 8, characterized in that: Before executing step 21, the method further includes counting the number of observation angle ranges occupied by laser points marked as obstacle points in a frame of line laser data collected in step 1; In a frame of line laser data collected in step 1, if it is detected that the number of observation angle ranges occupied by obstacle points is greater than or equal to the second preset observation angle number threshold, and the number of observation angle ranges occupied by obstacle points is less than or equal to the third preset observation angle number threshold, then the frame of line laser data collected in step 1 is marked as a legal data frame; If two frames of legal data frames are detected successively, the two frames of legal data frames are marked as two frames of line laser data required to be collected for counting the number of repeated angles and the total number of occupied angles; Among them, the second preset observation angle quantity threshold is smaller than the first preset observation angle quantity threshold, and the first preset observation angle quantity threshold is smaller than the third preset observation angle quantity threshold.

10. The method for identifying low-rise obstacles according to claim 9, characterized in that: The target detection quantity is 60, the first preset observation angle quantity threshold is 20; the preset reduction coefficient is 1.5; The second preset observation angle quantity threshold is 8, and the third preset observation angle quantity threshold is 30.

11. The method for identifying low-rise obstacles according to claim 6, characterized in that: In step 21, the method for counting the number of observation angle ranges repeatedly occupied by obstacle points marked in two frames of linear laser data collected sequentially includes: The robot divides the two frames of line laser data collected sequentially into a frame of line laser data collected previously and a frame of line laser data collected currently; The robot sequentially controls the obstacle marking logic information corresponding to each laser point in a frame of line laser data previously collected and the obstacle marking logic information corresponding to the laser points occupying the same observation angle range in a frame of line laser data currently collected to perform a logic AND operation, and obtains the obstacle marking logic operation results corresponding to the laser points occupying each observation angle range, and then counts the number of observation angle ranges occupied by the laser points whose obstacle marking logic operation results are logic 1, and then sets the number of observation angle ranges counted as the number of repeated angles, so that between two frames of line laser data collected sequentially, only one observation angle range that is repeatedly occupied is counted; Among them, the robot sets the obstacle marking logic information corresponding to the laser points marked as obstacle points in each frame of line laser data to logic 1; the robot sets the obstacle marking logic information corresponding to the laser points not marked as obstacle points in each frame of line laser data to logic 0.

12. The method for identifying low-rise obstacles according to claim 6, characterized in that: In step 21, the number of observation angle ranges that are not repeatedly occupied by obstacle points marked in two frames of linear laser data collected sequentially is counted, and the method for obtaining the total number of occupied angles includes: The robot divides the two frames of line laser data collected sequentially into a frame of line laser data collected previously and a frame of line laser data collected currently; The robot sequentially controls the obstacle marking logic information corresponding to each laser point in the previously collected frame of line laser data and the obstacle marking logic information corresponding to the laser points occupying the same observation angle range in the currently collected frame of line laser data to perform a logical OR operation, and obtains the obstacle marking logic operation results corresponding to the laser points occupying each observation angle range, and then counts the number of observation angle ranges occupied by the laser points whose obstacle marking logic operation results are logic 1, and then sets the counted number of observation angle ranges as the total number of occupied angles, thereby forming the number of observation angle ranges that are not repeatedly occupied by obstacle points in the two frames of line laser data collected sequentially; Among them, the robot sets the obstacle marking logic information corresponding to the laser points marked as obstacle points in each frame of line laser data to logic 1; the robot sets the obstacle marking logic information corresponding to the laser points not marked as obstacle points in each frame of line laser data to logic 0.

Images