Obstacle avoidance control method for robot and based on bÉzier curve fitting

Abstract

An obstacle avoidance control method for a robot and based on Bézier curve fitting, the obstacle avoidance control method for a robot comprising: step S1, a robot emitting a line laser in front of the robot; step S2, the robot collecting, in real time, obstacle contour points detected by the line laser, and acquiring the position coordinates of the obstacle contour points; step S3, setting a reference detection area (EFGH) on the side of the robot that is close to obstacles (#1, #2), and then using the position coordinates of obstacle contour points delineated by the reference detection area (EFGH) to perform Bézier curve fitting, so as to generate obstacle fitting curves (U1, U2); and step S4, the robot deviating from the obstacle fitting curves (U1, U2) by a preset obstacle avoidance margin (D) to obtain target obstacle avoidance curve trajectories (A2B2, A5B4); and then, the robot starting to move from the current position point to the target obstacle avoidance curve trajectories (A2B2, A5B4), and then moving according to the target obstacle avoidance curve trajectories (A2B2, A5B4).

Description

Robot obstacle avoidance control method based on Bezier curve fitting TECHNICAL FIELD

[0001] The present application relates to the field of robots, in particular to a robot obstacle avoidance control method based on Bezier curve fitting. BACKGROUND

[0002] At present, most of the sweeping robots in the process of executing navigation obstacle avoidance operation, the conventional is to plan an obstacle avoidance path in the marked obstacle grid in the grid map, wherein the obstacle contour is represented by the grid in the grid map.

[0003] Because of the resolution of the grid in the grid map, when a single grid is used to represent an obstacle position point, there is a distance error corresponding to the size of the resolution (for example, the grid resolution is 10mm, directly used to locate the obstacle will produce 10mm error), therefore, directly planning the path according to the grid map will bring the grid error, often will lose some of the obstacle contour points that should be collected, also makes the sweeping robot in the path moving according to the map planning will continue to jitter to appear positioning offset problem. SUMMARY

[0004] The present application discloses a robot obstacle avoidance control method based on Bezier curve fitting, and the specific technical solutions are as follows:

[0005] The robot obstacle avoidance control method based on Bezier curve fitting comprises the following steps: step S1, the robot emits a line laser to the front; step S2, the robot collects the obstacle contour points detected by the line laser in real time and obtains the position coordinates of the obstacle contour points; step S3, a reference detection area is set on the side of the robot close to the obstacle, and the position coordinates of the obstacle contour points framed by the reference detection area are used for Bezier curve fitting to generate an obstacle fitting curve; step S4, the robot offsets the obstacle fitting curve by a preset obstacle avoidance margin to obtain a target obstacle avoidance curve trajectory; then the robot moves from the current position point to the target obstacle avoidance curve trajectory, and moves according to the target obstacle avoidance curve trajectory.

[0006] In view of the grid positioning error caused by the grid map marking the obstacle contour points, the position coordinates of the obstacle contour points in the detection area on one side of the robot are extracted for Bezier curve fitting, so that, under the premise that each obstacle contour point is a control point, an obstacle fitting curve extending towards the front of the robot is fitted along the positive direction of the longitudinal axis of the robot body, the grid error existing in the position points marked by the grid map is eliminated, the profile sampling points that may be lost by using the grid map are filled, and the obstacle contour points are smoothed; then, based on the requirements of the robot body radius and the obstacle avoidance distance, the coordinates of each point of the obstacle fitting curve are offset to obtain the trajectory position points for the robot to avoid obstacles without collision, and a target obstacle avoidance curve trajectory for the robot to avoid obstacles is planned, so that the robot keeps a certain edge distance from the profile of the obstacle along the target obstacle avoidance curve trajectory, and a certain adjustable space is formed between the robot and the profile of the obstacle along, which is especially suitable for the robot to adjust the pose according to the speed difference between the left and right wheel groups to avoid obstacles.

[0007] On the target obstacle avoidance curve trajectory, the robot starts from the current position point, searches for a look-ahead point in the front of the robot along the extension direction of the target obstacle avoidance curve trajectory as the next target obstacle avoidance position point on the target obstacle avoidance curve trajectory, until the end point of the target obstacle avoidance curve trajectory is searched and walked to, so that the obstacle avoidance or collision avoidance is realized; on this basis, whenever a new obstacle contour point is framed by the detection area, a new obstacle fitting curve is fitted in the new ground area (a semi-closed area surrounded by the connecting line of the new control points) by using the Bezier curve fitting mode, so that the new target obstacle avoidance curve trajectory is connected and offset to form a new relatively smooth target obstacle avoidance curve trajectory, and the robot continuously walks along the profile of the obstacle on the side, and the profile of the obstacle on the side of the robot is tracked without collision. BRIEF DESCRIPTION OF DRAWINGS

[0008] FIG. 1 is a schematic diagram of an obstacle fitting curve U1 fitted in an embodiment, wherein the coordinate offset amount of the look-ahead point A2 and the starting point A1 of the obstacle fitting curve in the X-axis direction is equal to D; the obstacle fitting curve U1 is used to represent the right profile line of the obstacle #1 framed by the reference detection area EFGH, the starting point (equivalent to the starting point of the Bezier curve) of the obstacle fitting curve U1 and the lower endpoint of the right profile line framed by the reference detection area EFGH are both the position point A1, the terminal point (equivalent to the terminal point of the Bezier curve) of the obstacle fitting curve U1 and the upper endpoint of the right profile line framed by the reference detection area EFGH are both the position point B1; the target obstacle avoidance curve trajectory A2B2 is parallel to the obstacle fitting curve U1.

[0009] Figure 2 is a schematic diagram of a segment of an obstacle fitting curve U2 fitted in an embodiment, where the coordinate offset of a look-ahead point A5 from a position point A4 in the obstacle fitting curve in the X-axis direction is equal to D; the obstacle fitting curve U2 is used to represent the right contour line of the reference detection region EFGH-enclosed portion of obstacle #2, the starting point of the obstacle fitting curve U2 (equivalent to the starting point of the Bezier curve) and the lower end point of the right contour line enclosed by the reference detection region EFGH are both position point A3, the ending point of the obstacle fitting curve U2 (equivalent to the ending point of the Bezier curve) and the upper end point of the right contour line enclosed by the reference detection region EFGH are both position point B3; the target obstacle-avoiding curve trajectory A5B4 is parallel to the obstacle fitting curve U2.

[0010] Figure 3 is a flowchart of a robot obstacle-avoiding control method based on Bezier curve fitting disclosed in an embodiment. DETAILED DESCRIPTION

[0011] In order to more clearly illustrate the present application, specific embodiments are given below to further illustrate the present application. In the following description, specific details such as specific system structures, techniques, etc. are presented in order to thoroughly understand the embodiments of the present application. However, it should be clear to those skilled in the art that the present application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted in order not to obscure the description of the present application with unnecessary details.

[0012] As an embodiment, a robot obstacle-avoiding control method based on Bezier curve fitting is disclosed. The execution subject of the robot obstacle-avoiding control method is a robot with a circular body, which corresponds to the disc-shaped body shown in Figure 1 on the right side of obstacle #1, or which corresponds to the disc-shaped body shown in Figure 2 on the right side of obstacle #2; the robot is equipped with a line laser module at the front end, which is used to detect obstacles in the effective detection region. As shown in Figure 3, the robot obstacle-avoiding control method comprises:

[0013] Step S1, the robot emits line laser to its front direction, which can cover the advancing direction of the robot; then step S2 is executed. In order to obtain an effective detection area, the line laser module adopts two line laser emitters which are mutually arranged at a certain angle, the robot controls the line laser module to emit two intersecting line lasers to the front of the robot; in the moving plane of the robot, one of the line lasers is shot to the front left of the robot, and the other is shot to the front right of the robot, the two line lasers intersect at a perspective point, and a triangular area between the front end of the robot and the perspective point is a blind area generated by the intersecting line laser module, the triangular area formed by the line laser passing through the perspective point in the moving direction of the robot is an effective detection area, and further forms the working range of the line laser module in the moving direction of the robot. The included angle formed by the planes of the two line lasers is preferably 90 degrees, and the planes of the two line lasers are both perpendicular to the moving plane of the robot, and the planes of the two line lasers are preferably perpendicular to the ground.

[0014] Step S2, the robot collects the obstacle contour points detected by the line laser in real time and obtains the position coordinates of the obstacle contour points, and then step S3 is executed; in step S2, the obstacle contour points are reflection points formed by the line laser reflecting on the surface of the obstacle, each collected obstacle contour point has a three-dimensional coordinate information, including a position coordinate formed by the obstacle contour point relative to the current position point of the robot, height information and azimuth angle (the angle of the connecting line of the obstacle contour point and the current position point relative to the current moving direction of the robot), but the aforementioned position coordinate is not the grid coordinate in the coordinate system of the grid map (a map composed of a plurality of unit grids with a certain resolution).

[0015] It should be noted that the obstacle disclosed in the embodiment is obtained by real-time detection of the line laser, and the line laser hits the obstacle and forms a reflection point after reflecting on the surface of the obstacle; the height information of the front obstacle can be obtained through the reflection point, and whether the object detected by the line laser is an obstacle is determined according to the height information. For example, when the height of the reflection point is beyond the preset obstacle height range, it is determined that an obstacle is detected, and the obstacle needs to be avoided; wherein the height of the reflection point is the height difference of the reflection point relative to the ground area (based on the height of the ground area); the ground area is parallel to the moving plane of the robot, and the height uniformity of a partial area can be analyzed to determine whether the partial area is the ground.

[0016] Step S3, a reference detection area is set on the side of the robot close to the obstacle, and then the position coordinates of the obstacle profile points framed by the reference detection area are used to perform Bezier curve fitting to generate an obstacle fitting curve; and then step S4 is performed. The reference detection area is preferably the rectangular area EFGH set on the right side of the robot in FIGS. 1 and 2, and is displaced following the movement of the robot to frame the area newly traversed outside the side of the robot in real time. The rectangular area EFGH set in FIG. 1 can cover the inwardly concave profile of the right side of the obstacle #1 (represented by the discrete points shown in the figure, i.e., the discrete distributed obstacle profile points detected by the line laser module); and the rectangular area EFGH set in FIG. 2 can cover the outwardly convex profile of the right side of the obstacle #2 (represented by the discrete points shown in the figure, i.e., the discrete distributed obstacle profile points detected by the line laser module). Based on the obstacle profile points framed (actually covered) by the reference detection area, the obstacle fitting curve is used to represent the profile of the obstacle to which the obstacle profile points framed by the reference detection area belong, the obstacle profile points framed in the reference detection area are configured as the control points of the Bezier curve, and the obstacle profile points framed in the reference detection area include the start point of the Bezier curve and the end point of the Bezier curve, the relative positions between the start point of the Bezier curve and the end point of the Bezier curve control the curvature of the Bezier curve, and affect the smoothness of the obstacle fitting curve.

[0017] Step S4, the robot offsets the obstacle fitting curve by a preset obstacle avoidance margin to obtain a target obstacle circumventing curve trajectory; then the robot moves from the current position point to the target obstacle circumventing curve trajectory, specifically to the starting look-ahead point in the target obstacle circumventing curve trajectory; and then moves according to the target obstacle circumventing curve trajectory to achieve the purpose of circumventing the obstacle. In step S4, the manner of offsetting the obstacle fitting curve by the preset obstacle avoidance margin is to control the points in the obstacle fitting curve to offset in a specific direction (at least a direction related to the coordinate axis) so as to obtain look-ahead points outside the obstacle and located in the passable area, which are sequentially connected as the target obstacle circumventing curve trajectory, wherein the starting look-ahead point is the starting point of the robot in the target obstacle circumventing curve trajectory for circumventing the obstacle; in order to prevent the robot from colliding with the obstacle, the coordinate offset is set considering the requirements of the body radius and the circumventing distance, i.e., the preset obstacle avoidance margin is set. Then the robot moves from the current position point to the starting look-ahead point, and then starts to move on the target obstacle circumventing curve trajectory in the extension direction (which can be the direction of increasing longitudinal coordinates) of the target obstacle circumventing curve trajectory, point by point, to realize walking according to the target obstacle circumventing curve trajectory, and to realize walking along the extension direction of the obstacle fitting curve under the premise of not touching the obstacle considering the distance.

[0018] In step S4, the robot moves from the current position point to the starting lookahead point, generally in a straight line along a predetermined direction. Corresponding to FIG. 1, the center of the robot body moves from position point 0 (taken as the starting point of walking around the obstacle) to position point A2 (the starting lookahead point). Corresponding to FIG. 2, the center of the robot body moves from position point 0 (taken as the starting point of walking around the obstacle) to position point A5 (the starting lookahead point). Preferably, the positive direction of the longitudinal axis of the robot body is set as the current moving direction of the robot, and the real-time moving direction of the robot is marked by the coordinate axis direction of the robot coordinate system. Further, when the robot moves to the end point of the target obstacle-avoiding curve trajectory, a new obstacle fitting curve (the extension direction and the curvature change) is fitted again using step S3, a new target obstacle-avoiding curve trajectory is offset in step S4, and the robot walks along the new target obstacle-avoiding curve trajectory, thereby continuously tracking the side obstacle to effectively avoid the obstacle.

[0019] In summary, for the grid positioning error generated when the obstacle contour points are marked on the grid map, the application first extracts the position coordinates of the obstacle contour points in the detection area on one side of the robot for Bezier curve fitting, thereby determining the control points of each obstacle contour point, fitting the obstacle fitting curve extending towards the front of the robot along the positive direction of the longitudinal axis of the robot body, eliminating the grid error of the position points marked using the grid map, inserting target points between adjacent control points to connect to form each Bezier curve in the process of fitting the obstacle fitting curve, filling the contour sampling points that may be lost when the grid map is marked without fitting processing, and smoothing the obstacle contour points. Then, based on the requirements of the robot body radius and the obstacle-avoiding distance, the coordinates of each point of the obstacle fitting curve are offset to obtain the trajectory position points for the robot to walk around without collision, and a target obstacle-avoiding curve trajectory is planned for the robot to walk around, so that the robot maintains a certain edge distance from the contour of the obstacle walked along in the process of walking along the target obstacle-avoiding curve trajectory, and a certain adjustable space is formed between the robot and the contour of the obstacle walked along, which is especially suitable for robots that adjust the pose according to the speed difference between the left and right wheel groups to walk around the obstacle.

[0020] As an embodiment, in step S2, the obstacle profile points detected by the line laser are formed by the reflection of the line laser emitted by the at least one line laser emitter on the profile of the obstacle, and the obstacle profile points detected by the line laser are used to connect to form the profile of the obstacle detected by the line laser. The line laser module comprises a line laser emitter and a laser receiver, and the line laser emitter is installed at the front end of the side of the robot; if there are two line laser emitters, they are respectively installed on the left and right sides of the central axis of the robot body, and generally at the front end of the side of the robot; the line laser emitted by the two line laser emitters forms an effective detection area in the travel plane of the robot.

[0021] With the travel plane of the robot as the reference, the emission range of the line laser emitter forms a preset upward emission angle upward and a preset downward emission angle downward. Preferably, with the horizontal as the reference, the emission range of the line laser emitter is 20 degrees upward and 30 degrees downward. The laser receiver is installed on the side of the robot. With the travel plane of the robot as the reference, the receiving range of the laser receiver forms a preset upward receiving angle upward and a preset downward receiving angle downward. Preferably, with the horizontal as the reference, the receiving range of the laser receiver is 30 degrees upward and 30 degrees downward, so that the acquisition range of the laser receiver on the travel plane of the robot is wider than the emission range of the laser emitter. The laser receiver is used to acquire the reflection points of the line laser on the obstacle and configure the reflection points as the obstacle profile points detected by the line laser emitted by the line laser emitter. Each acquired obstacle profile point has position coordinates, azimuth angle information and height information, so as to determine the height, length and width of the detected obstacle in the moving direction of the robot and form the profile position of the detected obstacle; the laser receiver is used to acquire the reflection points of the line laser on the obstacle and configure the reflection points as the obstacle profile points detected by the line laser emitted by the line laser emitter. When the laser receiver adopts a camera, the intersection of the two line lasers emitted by the two line laser emitters respectively in the front direction of the robot is within the shooting range of the camera. In specific embodiments, the laser irradiation range and the shooting range of the camera can be adjusted by the installation height on the robot; the two line lasers respectively emitted by the two line laser emitters detect the height information of the obstacle profile points in the vertical plane area of the travel plane of the robot.

[0022] In the embodiment, the robot comprises a semi-circular head, a semi-circular body and two symmetrically arranged wheels connected by an axle S arranged at the boundary between the head and the body; the head is arranged in front of the axle S, the body is arranged behind the axle S, and the two symmetrically arranged wheels arranged on the axle S are respectively a wheel L on the left side of the axle S and a wheel R on the right side of the axle S; the semi-circular head and the semi-circular body are configured into a circular body, so that the semi-circular head and the semi-circular body are symmetrically arranged about the axle S; the center of the axle S is the body center 0 of the robot; the length of the axle S is less than the body diameter. The reference detection area is used to cover part or all of the profile of the side obstacle during the movement of the robot. In the embodiment, the length of the axle S is the distance between the wheel L and the wheel R in the positive direction of the body transverse axis, the distance from the wheel L or the wheel R to the body center 0 is less than the body radius, the length of the axle S is less than the body diameter, and the body diameter is equal to the body width. The longest distance from the body center 0 of the robot to the edge of the body of the robot is represented as the body radius of the robot, and the body radius is set to be equal to half of the body width or half of the head width. The side obstacle is an obstacle located outside the side of the body of the robot, and the side of the body of the robot includes the left side of the body or the right side of the body; then the side obstacle refers to an obstacle located on the left side or the right side of the robot and not in contact with the robot.

[0023] The moving direction of the robot refers to the front direction of the robot, and the moving direction of the robot is parallel to the positive direction of the body longitudinal axis, and the positive direction of the body longitudinal axis is vertically arranged with the axle S; the positive direction of the body transverse axis is in the direction towards the side obstacle of the robot and parallel to the axle. Illustratively, in FIG. 1, the positive direction of the body transverse axis is the positive direction of the X axis, the positive direction of the X axis shown in FIG. 1 is in the direction towards the side obstacle #1 of the robot and parallel to the axle S, the positive direction of the Y axis is the positive direction of the body longitudinal axis, and the positive direction of the body longitudinal axis is vertically arranged with the positive direction of the body transverse axis. Illustratively, in FIG. 2, the positive direction of the body transverse axis is the positive direction of the X axis, the positive direction of the X axis shown in FIG. 2 is in the direction towards the side obstacle #2 of the robot and parallel to the axle S, the positive direction of the Y axis is the positive direction of the body longitudinal axis, and the positive direction of the body longitudinal axis is vertically arranged with the positive direction of the body transverse axis.

[0024] As one embodiment, the position coordinates of each collected obstacle contour point are set as local coordinates relative to the robot's current position. These local coordinates are set within the robot's coordinate system, where the robot's current position is set as the origin. The positive direction of the robot's horizontal axis is the positive direction of the robot's X-axis, and the positive direction of the robot's vertical axis is the positive direction of the robot's Y-axis. It should be noted that the two-dimensional coordinates of the bright spot formed by the line laser striking the obstacle surface or the two-dimensional coordinates of the reflected point are also the position coordinates of the obstacle contour points, all belonging to local coordinates relative to the robot's current position. The positive direction of the robot's vertical axis is parallel to the positive direction of the robot's vertical axis, and the positive direction of the robot's horizontal axis is parallel to the positive direction of the robot's horizontal axis. To mark the collected obstacle contour points onto the grid map, this embodiment transforms the local coordinates to the global map coordinate system used to construct the grid map. This transformation involves converting the robot coordinate system to the global map coordinate system of the grid map. Specifically, the local coordinates are transformed into the global map coordinate system through rotation and translation to obtain the grid coordinates of the obstacle contour points. The rotation angle required for the local coordinate transformation is equal to the angle between the same attribute coordinate axes of the global map coordinate system and the robot coordinate system. Specifically, it is the angle between the positive Y-axis of the global map coordinate system and the positive Y-axis of the robot coordinate system, or the angle between the positive X-axis of the global map coordinate system and the positive X-axis of the robot coordinate system. The angle between the positive X-axis and the coordinate system; the coordinate offset required for local coordinate translation transformation is the coordinate of the robot's current position point in the global map coordinate system, that is, the coordinate offset relationship between the origin of the global map coordinate system and the origin of the robot coordinate system; the coordinate system transformation equation can be constructed using the rotation angle required for local coordinate rotation transformation and the coordinate offset required for local coordinate translation transformation, based on the trigonometric function transformation relationship that can be mastered by those skilled in the art; then, the position coordinates of the obstacle contour points are substituted into the coordinate system transformation equation to obtain the grid coordinates of the obstacle contour points, thereby marking the obstacle contour points in the grid of the grid map, and thus marking the position information of the obstacle contour in the grid map.

[0025] Specifically, the coordinate system transformation equations are:

[0026] x m =x p cosθ-y p sinθ+x0;

[0027] y m =x p sinθ+y p sinθ+y0;

[0028] Wherein, the position coordinates (coordinates in the robot coordinate system) of the obstacle contour points are (x...p , y p ), the coordinate of the current position point of the robot in the global map coordinate system is (x0, y0); the rotation angle required for the local coordinate rotation transformation is θ, which is the angle required for the rotation of the obstacle profile point in the coordinate system transformation; the grid coordinates of the obstacle profile point are (x m , y m ). The obstacle profile points collected by the line laser arranged at the front end of the head (belonging to the laser point cloud) are discrete points reflected to the side profile line of the obstacle close to the side of the robot, each discrete point can be converted to the global map coordinate system, then the straight line distance between the corresponding discrete point and the center 0 of the machine body is calculated by using the converted coordinates of each discrete point in the map coordinate system, which is reflected as the distance between the one side profile line of the obstacle and the side of the robot.

[0029] As an embodiment, in the step S3, a reference detection area for covering the contour of the obstacle is set on the side of the robot close to the obstacle, at this time, the distance between the robot moving to the center of the body 0 and the obstacle contour point is greater than or equal to the body radius; the reference detection area covers at least two obstacle contour points of the obstacle, facilitating subsequent Bezier curve fitting. In this embodiment, the reference detection area has an overlapping area with the body coverage area of the robot (covering the area near the edge of the body) or has a spacing (i.e., the reference detection area is completely located outside the body of the robot); the distance between the boundary point farthest from the boundary on the side of the body close to the obstacle in the reference detection area and the center of the body in the positive direction of the body transverse axis is greater than the body radius, preferably greater than a preset obstacle avoidance trigger distance, to cover the obstacle contour points outside the body, and the preset obstacle avoidance trigger distance is equal to the sum of the body radius and a first preset spacing, and the first preset spacing is selected from a distance value of 1 cm to 2 cm. Preferably, the length of the area boundary of the reference detection area EFGH in the positive direction of the body transverse axis is greater than or equal to a preset obstacle avoidance edge distance, and the preset obstacle avoidance edge distance is equal to the sum of the body radius and a second preset spacing, and the second preset spacing is 15 mm. In this embodiment, the reference detection area is set to cover the passable area between the body and the side obstacle, so that the robot can be controlled not to touch the obstacle; the straight-line distance covered by the reference detection area in the positive direction of the body longitudinal axis does not exceed the maximum detection distance of the line laser emitted by the line laser emitter, and can be equal to the body radius, so as to cover the relatively close obstacle contour points, without covering the obstacle contour points that have a far edge distance and thus cannot effectively walk along the edge, or the obstacle contour points that do not need to be close. The reference detection area is set to cover part or all of the passable area between the body and the side obstacle. The area covered by the reference detection area outside the body is greater than the area covered by the reference detection area inside the body. For example, the area covered by the reference detection area EFGH outside the body in FIGS. 1 and 2 accounts for a large proportion, the area covered by the reference detection area EFGH inside the body accounts for a small proportion, and the area covered by the reference detection area EFGH inside the body is a small arc-shaped area.

[0030] In the embodiment where the reference detection area is a polygonal area, if there is at least one area boundary of the reference detection area parallel or perpendicular to the wheel axis S, the angle conversion step can be reduced when calculating the side length of the reference detection area. When the reference detection area overlaps with the body coverage area of the robot, the coverage area of the reference detection area at the head portion is greater than that at the body portion, and within the area occupied by the reference detection area outside the body, the coverage area of the reference detection area at the head side is greater than that at the body side, so that the reference detection area is biased to cover the front area of the robot, adapting to the detection direction requirement of the line laser. Corresponding to the reference detection area EFGH in FIG. 1 and FIG. 2, the area of the reference detection area EFGH above the wheel axis S (above the X-axis) is greater than that below the wheel axis S (below the X-axis).

[0031] When the reference detection area is a regular geometric figure with a center point, the center of the reference detection area is set outside the body, and the distance between the center of the reference detection area and the center of the body in the positive direction of the body transverse axis is greater than or equal to the body radius. The distance between the center of the reference detection area and the center of the body in the positive direction of the body longitudinal axis can be less than the body radius, or greater than or equal to the body radius, but not exceeding the maximum detection distance of the line laser emitted by the line laser emitter. The polygonal detection area is preferably a rectangular area, and the longest side of the rectangular area is preferably greater than or equal to the body radius.

[0032] When a circular area is used to represent the reference detection area, only the center and the radius are determined to set the reference detection area, so as to include all the profile point information on one side of the obstacle with relatively small description, and then facilitate the limitation and real-time fitting processing of the profile points; wherein the reference detection area can intersect or externally tangent to the circular body, but the reference detection area does not completely coincide with the body; the radius of the reference detection area is greater than or equal to the body radius; but preferably less than the maximum detection distance of the line laser emitted by the line laser emitter. Further, the reference detection area can also be represented by an inscribed triangle or an externally tangent triangle of the circular area.

[0033] As a preferred example, as shown in FIG. 1 and FIG. 2, the reference detection area is a rectangular area EFGH. The setting method of the reference detection area includes:

[0034] A position point at a first preset longitudinal distance from the center of the machine body in the positive direction of the longitudinal axis of the machine body (the positive direction of the Y axis shown in the figure) and at a first preset lateral distance from the center of the machine body in the positive direction of the lateral axis of the machine body (the positive direction of the X axis shown in the figure) is selected as the upper left corner point H. If a coordinate system is constructed with the center of the machine body O as the origin, with the positive direction of the longitudinal axis of the machine body as the positive direction of the longitudinal coordinate axis, and with the positive direction of the lateral axis of the machine body as the positive direction of the lateral coordinate axis, the coordinates of the upper left corner point H are (the first preset lateral distance, the first preset longitudinal distance).

[0035] A position point at a second preset longitudinal distance from the center of the machine body in the opposite direction of the positive direction of the longitudinal axis of the machine body and at a first preset lateral distance from the center of the machine body in the positive direction of the lateral axis of the machine body is selected as the lower left corner point E. If a coordinate system is constructed with the center of the machine body O as the origin, with the positive direction of the longitudinal axis of the machine body as the positive direction of the longitudinal coordinate axis, and with the positive direction of the lateral axis of the machine body as the positive direction of the lateral coordinate axis, the coordinates of the lower left corner point E are (the first preset lateral distance, 0-the second preset longitudinal distance).

[0036] A position point at a first preset longitudinal distance from the center of the machine body in the positive direction of the longitudinal axis of the machine body and at a second preset lateral distance from the center of the machine body in the positive direction of the lateral axis of the machine body is selected as the upper right corner point G. If a coordinate system is constructed with the center of the machine body O as the origin, with the positive direction of the longitudinal axis of the machine body as the positive direction of the longitudinal coordinate axis, and with the positive direction of the lateral axis of the machine body as the positive direction of the lateral coordinate axis, the coordinates of the upper right corner point G are (the second preset lateral distance, the first preset longitudinal distance).

[0037] A position point at a second preset longitudinal distance from the center of the machine body in the opposite direction of the positive direction of the longitudinal axis of the machine body and at a second preset lateral distance from the center of the machine body in the positive direction of the lateral axis of the machine body is selected as the lower right corner point F. If a coordinate system is constructed with the center of the machine body O as the origin, with the positive direction of the longitudinal axis of the machine body as the positive direction of the longitudinal coordinate axis, and with the positive direction of the lateral axis of the machine body as the positive direction of the lateral coordinate axis, the coordinates of the lower right corner point F are (the second preset lateral distance, 0-the second preset longitudinal distance).

[0038] Then, the left lower corner point E and the left upper corner point H are connected, the left upper corner point H and the right upper corner point G are connected, the right upper corner point G and the right lower corner point F are connected, and the right lower corner point F and the left lower corner point E are connected to obtain the reference detection region EFGH. For the left lower corner point E, the left upper corner point H, the right upper corner point G and the right lower corner point F, the line segments are sequentially connected to form a rectangular region EFGH. If the body center 0 is arranged in the global map coordinate system, the body center 0 is not necessarily the origin of the global map coordinate system, and correspondingly, the positive direction of the body longitudinal axis is not necessarily the positive direction of the longitudinal coordinate axis of the global map coordinate system, and the positive direction of the body transverse axis is not necessarily the positive direction of the transverse coordinate axis of the global map coordinate system. Therefore, when the reference detection region EFGH is used to frame a map region in the grid map, the left lower corner point E, the left upper corner point H, the right upper corner point G and the right lower corner point F all need to be converted into the coordinates in the global map coordinate system according to the coordinate transformation equation disclosed in the foregoing embodiments.

[0039] Corresponding to FIG. 1 and FIG. 2, the vertical distance from the body center 0 to the dashed line segment EH is equal to the first preset transverse distance, the vertical distance from the body center 0 to the dashed line segment FG is equal to the second preset transverse distance, the distance from the wheel shaft to the dashed line segment HG is equal to the first preset longitudinal distance, and the distance from the wheel shaft to the dashed line segment EF is equal to the second preset longitudinal distance.

[0040] In the embodiment, the boundary length of the reference detection region in the positive direction of the body transverse axis is equal to a preset obstacle avoidance margin, and then in the positive direction of the body transverse axis, the sum of the first preset transverse distance and the preset obstacle avoidance margin is equal to the second preset transverse distance; or the second preset transverse distance is set to be greater than or equal to the preset obstacle avoidance triggering distance but less than the maximum detection distance of the linear laser emitted by the linear laser emitter; so that the line connecting the right upper corner point G and the right lower corner point F is located outside the body, and an obstacle avoidance space is reserved; the first preset longitudinal distance is set to be greater than or equal to the body radius but less than the maximum detection distance of the linear laser emitted by the linear laser emitter, so as to detect the region outside the front boundary of the body. At the same time, the first preset transverse distance is set to be greater than half the length of the wheel shaft and less than the body radius, so that the left lower corner point E is arranged inside the body, and the second preset longitudinal distance is set to be less than the body radius or half the length of the wheel shaft, so that the left lower corner point E is located on the side of the body close to the side boundary obstacle.

[0041] Generally, since the obstacle contour points are represented in advance using the grid map, and the robot directly avoids the obstacles according to the grid map, grid errors (determined by the resolution of the grid map, here the grid resolution is 10 mm, and directly used to avoid the obstacles, there is a positioning coordinate error of 10 mm) are caused, so that the differential robot such as the sweeping robot continuously shakes, and therefore fitting (equivalent to smoothing the grid, the smoothness of the curve obtained by fitting is less than the grid resolution) is needed to eliminate the errors and fill the missing sampling points due to the grid resolution. Specifically, the grid coordinates of the obstacle contour points are converted back to the robot coordinate system, wherein the robot coordinate system is a coordinate system with the center of the robot body as the origin, i.e., a coordinate system with the current position point of the robot as the origin.

[0042] In the embodiment, the reference detection area is set in the grid map to frame part of the map area, the robot extracts the grid coordinates of the obstacle contour points from the area covered by the reference detection area in the grid map, converts the extracted grid coordinates from the global map coordinate system back to the robot coordinate system to obtain the local coordinates relative to the current position point of the robot, and marks the position coordinates of the obstacle contour points framed by the reference detection area as the coordinates of the obstacle contour points in the robot coordinate system. It should be noted that the reference detection area is configured to cover the grid map in step S3 to frame the grid coordinates of the obstacle contour points in real time during the movement of the robot, and convert them into the position coordinates of the obstacle contour points in real time.

[0043] In the embodiment, the rotation transformation of the grid coordinates of the obstacle contour points between the global map coordinate system and the robot coordinate system is the inverse transformation of the rotation transformation of the local coordinates, and the translation transformation of the grid coordinates of the obstacle contour points is the inverse transformation of the translation transformation of the local coordinates; specifically, the coordinate system inverse transformation equation is:

[0044] x p =(x m -x0)cosθ+(y m -y0)sinθ;

[0045] y p =(y m -y0)cosθ-(x m -x0)sinθ;

[0046] wherein the position coordinates of the obstacle contour points (coordinates in the robot coordinate system) are (x p , y p), the coordinates of the current position point of the robot in the global map coordinate system are (x0, y0); the rotation angle required for the rotation transformation of the grid coordinates of the obstacle profile points is θ, which is the angle required for the rotation of the obstacle profile points in the coordinate system transformation; the grid coordinates of the obstacle profile points are (x m , y m ).

[0047] As an embodiment, in the step S3, the method of using the position coordinates of the obstacle profile points framed in the reference detection area to perform Bezier curve fitting includes:

[0048] The robot determines the start point of the obstacle fitting curve and the end point of the obstacle fitting curve among the obstacle profile points framed in the reference detection area along the positive direction of the body longitudinal axis (the discrete points near the left profile of the obstacle in the reference detection area EFGH in FIG. 1 and FIG. 2), and the Bezier curve to be fitted is extended along the positive direction of the body longitudinal axis. If the positive direction of the body longitudinal axis is the positive direction of the Y axis in the figure, then the obstacle profile points distributed between the start point and the end point along the positive direction of the body longitudinal axis in the reference detection area are sequentially marked as control points required for fitting the Bezier curve; the obstacle profile points distributed between the start point and the end point along the positive direction of the body longitudinal axis in the reference detection area are sequentially marked as control points required for fitting the Bezier curve, and the obstacle fitting curve belongs to the Bezier curve; the number of control points is equal to the sum of the order of the Bezier curve and the value 1; and the order of the Bezier curve is represented by n.

[0049] According to the start point of the obstacle fitting curve, the end point of the obstacle fitting curve, and the control points sequentially marked between the start point and the end point of the obstacle fitting curve, the obstacle fitting curve based on the n-order Bezier curve is generated in the robot coordinate system plane, and accordingly, the trajectory equation of the obstacle fitting curve based on the n-order Bezier curve is generated, that is, the n-order Bezier curve equation is obtained; so that the relatively discrete obstacle profile points (the discrete points near the left profile of the obstacle in the reference detection area EFGH in FIG. 1 and FIG. 2) are fitted into a relatively continuous profile line, wherein the start point of the obstacle fitting curve and the end point of the obstacle fitting curve both belong to the control points. As shown in the left concave dotted curve of the left profile of the obstacle #1 in FIG. 1, or the right convex dotted curve of the left profile of the obstacle #2 in FIG. 2. The grid error caused by directly avoiding the obstacle according to the grid map is overcome.

[0050] Specifically, the trajectory equation of the obstacle fitting curve is:

[0051] wherein B x(t) is the horizontal coordinate of the point in the obstacle fitting curve, B y (t) is the vertical coordinate of the point in the obstacle fitting curve, i represents the serial number of the control point required by the obstacle fitting curve, the serial number i represents the serial number of the control point required by the obstacle fitting curve along the positive direction of the body longitudinal axis, that is, the arrangement order of the obstacle profile points framed in the reference detection area along the positive direction of the body longitudinal axis, and is configured to start from the value 0 and increase along the positive direction of the body longitudinal axis. t is equal to the ratio between i and (n+1), which is used to represent the ratio of the serial number of the newly fitted point along the positive direction of the body longitudinal axis to the total number of points required to be fitted. The absolute value of the difference between the vertical coordinate of the point with the serial number i and the vertical coordinate of the point with the serial number (i+1) can be fixed; P 0x represents the horizontal coordinate of the starting point position of the obstacle fitting curve in the reference detection area, P 0y represents the vertical coordinate of the starting point position of the obstacle fitting curve in the reference detection area; P nx represents the horizontal coordinate of the ending point position of the obstacle fitting curve in the reference detection area, P ny represents the vertical coordinate of the ending point position of the obstacle fitting curve in the reference detection area. represents the combination number of (P ix , P iy ) at a specific t value, which is similar to the binomial coefficient (combination number) in the binomial expansion formula, and follows the Yang Hui triangle distribution. P ix represents the horizontal coordinate of the obstacle profile point in the reference detection area for setting as the control point P i represents the vertical coordinate of the obstacle profile point in the reference detection area for setting as the control point P iy represents the vertical coordinate of the obstacle profile point in the reference detection area for setting as the control point P i represents the vertical coordinate of the obstacle profile point in the reference detection area for setting as the control point P

[0052] It should be noted that the trajectory equation of the obstacle fitting curve belongs to the Bezier curve formula. The Bezier curve is a mathematical curve applied to two-dimensional graphics applications; the curve definition: starting point (also called starting point), ending point (also called ending point), control point; by adjusting the control point, the shape of the Bezier curve changes.

[0053] In the process of Bezier curve fitting, the number of types of t values is equal to the number of points fitted by the obstacle fitting curve, so that the points with low continuity are interpolated into relatively continuous points; when the relative positions of two adjacent control points change in the positive direction of the body longitudinal axis, the curvature (degree of bending) of the obstacle fitting curve changes. The two adjacent control points use P i and P i+1wherein, i∈[0, n-1], the start point of the obstacle fitting curve is denoted as P0, and the end point of the obstacle fitting curve is denoted as Pn. n wherein, i gradually increases from the start point to the end point in the process of the Bezier curve fitting, and t equals to 1 when the end point is reached; so that the relatively discrete obstacle profile points are fitted into a relatively continuous profile line, and the Bezier curve fitting can fill the missing profile points in the collection process using the fitting points and fit a curve closer to the profile of the obstacle, so as to realize the smoothing processing of the obstacle grid using the Bezier curve.

[0054] As known by those skilled in the art, when the Bezier curve is locked at the start point and the end point, the middle point on the line connecting the adjacent two control points is moved to change t, and the Bezier curve does not pass through all the middle points, but it is ensured that it will pass through the start point and the end point.

[0055] As an embodiment, the start point of the obstacle fitting curve is the obstacle profile point closest to the boundary of the reference detection area in the opposite direction of the positive direction of the body longitudinal axis, or the boundary point of the reference detection area through which the profile of the obstacle represented by the obstacle fitting curve passes in the opposite direction of the positive direction of the body longitudinal axis, so that the start point of the obstacle fitting curve is the obstacle profile point farthest from the obstacle-avoiding walking start point in the opposite direction of the positive direction of the body longitudinal axis; and the end point of the obstacle fitting curve is the obstacle profile point closest to the boundary of the reference detection area in the positive direction of the body longitudinal axis, or the boundary point of the reference detection area through which the profile of the obstacle represented by the obstacle fitting curve passes in the positive direction of the body longitudinal axis, so that the start point of the obstacle fitting curve is the obstacle profile point farthest from the obstacle-avoiding walking start point in the positive direction of the body longitudinal axis. Accordingly, the start point of the obstacle fitting curve in the reference detection area in FIG. 1 is the position point A1, and the end point of the obstacle fitting curve in the reference detection area in FIG. 1 is the position point B1. The start point of the obstacle fitting curve in the reference detection area in FIG. 2 is the position point A3, and the end point of the obstacle fitting curve in the reference detection area in FIG. 2 is the position point B3.

[0056] Preferably, the profile of the obstacle characterized by the obstacle fitting curve is located in the positive direction of the body lateral axis of the robot, and the negative direction of the body lateral axis is configured as the target offset direction, wherein the preset obstacle avoidance margin is set to be equal to the sum of the body radius and a preset obstacle circumventing distance, and the preset obstacle circumventing distance is preferably 15 mm; the preset obstacle circumventing distance is the second preset distance disclosed in the foregoing embodiment. The target offset direction herein is used to define the offset direction of the target point in the obstacle fitting curve in step S4. The obstacle fitting curve is translated along the target offset direction by the preset obstacle avoidance margin to obtain a target obstacle circumventing curve trajectory, so that the position point actually moved by the robot can be controlled to maintain a certain distance from the obstacle fitting curve, that is, in the obstacle fitting curve and the target obstacle circumventing curve trajectory, the coordinate offset amount of the point with the same vertical coordinate as the starting forward-looking vertical coordinate in the horizontal direction of the body lateral axis is equal to the preset obstacle avoidance margin; the forward-looking point for the actual movement of the robot is obtained from the target obstacle circumventing curve trajectory, so as to realize the walking of the robot along the extension direction of the obstacle fitting curve without touching the obstacle.

[0057] As an embodiment, FIG. 1 is a schematic diagram of an obstacle fitting curve U1 fitted in an embodiment, and the profile on the left side of the obstacle #1 is entirely covered by the reference detection region EFGH; the coordinate offset amount of the forward-looking point A2 and the starting point A1 of the obstacle fitting curve in the X-axis direction is equal to the preset obstacle avoidance margin D; the obstacle fitting curve U1 is used to characterize the inwardly concave profile line in the obstacle #1 framed by the reference detection region EFGH, the starting point (equivalent to the starting point of the Bezier curve) of the obstacle fitting curve U1 and the lower endpoint of the inwardly concave profile line framed by the reference detection region EFGH are both the position point A1, and the terminal point (equivalent to the terminal point of the Bezier curve) of the obstacle fitting curve U1 and the upper endpoint of the inwardly concave profile line framed by the reference detection region EFGH are both the position point B1. The dashed curve extending from the forward-looking point A2 to the positive direction of the Y-axis is the trajectory of the actual walking of the robot, so as to walk along the upward extension direction of the obstacle fitting curve U1 in FIG. 1, and preferably, the dashed curve A2B2 extending from the forward-looking point A2 to the positive direction of the Y-axis is parallel to the obstacle fitting curve U1.

[0058] As another embodiment, FIG. 2 is a schematic diagram of a segment of the obstacle fitting curve U2 fitted in an embodiment, the profile of the left side of the obstacle #2 is not entirely covered by the reference detection region EFGH; wherein the coordinate offset of the look-ahead point A5 and the position point A4 in the X-axis direction of the obstacle fitting curve is equal to the preset obstacle avoidance margin D; the obstacle fitting curve U2 is used to represent the outward convex profile line of the obstacle #2 framed by the reference detection region EFGH, the starting point (equivalent to the starting point of the Bezier curve) of the obstacle fitting curve U2 and the lower endpoint of the outward convex profile line framed by the reference detection region EFGH are both the position point A3, and the terminal point (equivalent to the terminal point of the Bezier curve) of the obstacle fitting curve U2 and the upper endpoint of the outward convex profile line framed by the reference detection region EFGH are both the position point B3. The dotted curve extending from the look-ahead point A5 to the positive direction of the Y-axis is the actual walking trajectory of the robot, to walk along the extension direction of the obstacle fitting curve A4B3, that is: the robot walks along the extension direction of the obstacle fitting curve A4B3 on the target obstacle-avoiding curve trajectory A5B4, wherein the obstacle fitting curve A4B3 can be parallel to the target obstacle-avoiding curve trajectory A5B4.

[0059] As an embodiment, in step S4, the method that the robot moves from the current position point to the target obstacle-avoiding curve trajectory includes: the robot calculates the adjacent longitudinal coordinate of the current position point in the positive direction of the body longitudinal axis, and sets the adjacent longitudinal coordinate as the starting look-ahead longitudinal coordinate, the difference between the adjacent longitudinal coordinate of each position point and the longitudinal coordinate of the position point is equal to a fixed coordinate value. Then the robot substitutes the starting look-ahead longitudinal coordinate into the trajectory equation of the obstacle fitting curve, calculates the transverse coordinate of the point in the obstacle fitting curve with the same longitudinal coordinate as the starting look-ahead longitudinal coordinate, obtains the position coordinate of the point in the obstacle fitting curve with the same longitudinal coordinate as the starting look-ahead longitudinal coordinate, and the calculation and the corresponding position coordinate are performed in the same coordinate system. In this embodiment, the points in the obstacle fitting curve are all offset from the target offset direction by the preset obstacle-avoiding margin to obtain the points in the target obstacle-avoiding curve trajectory, that is, the target obstacle-avoiding curve trajectory is obtained by offsetting the obstacle fitting curve from the target offset direction by the preset obstacle-avoiding margin; in the obstacle fitting curve, the transverse coordinate offset between the point with the starting look-ahead longitudinal coordinate and the starting look-ahead point is equal to the preset obstacle-avoiding margin D; therefore, the point in the obstacle fitting curve with the same longitudinal coordinate as the starting look-ahead longitudinal coordinate is offset from the target offset direction by the preset obstacle-avoiding margin to obtain the starting look-ahead point of the robot, and the starting look-ahead point of the robot is determined to be the point on the target obstacle-avoiding curve trajectory; the point in the obstacle fitting curve with the same longitudinal coordinate as the starting look-ahead longitudinal coordinate is marked as the target point, so the target point can be point A1 in FIG. 1, and the target point can also be point A4 in FIG. 2; then the target point is offset from the target offset direction by the preset obstacle-avoiding margin D to obtain the starting look-ahead point, so the starting look-ahead point is point A2 in FIG. 1, and the starting look-ahead point is point A5 in FIG. 2. Then the robot moves from the current position point to the starting look-ahead point to enter the target obstacle-avoiding curve trajectory, so as to start walking along the extension direction of the obstacle fitting curve and maintaining a certain distance from the obstacle fitting curve. Generally, the method that the robot moves from the current position point to the starting look-ahead point can be straight walking according to the direction formed by the starting look-ahead point relative to the current position point, and the walking distance is equal to the straight-line distance between the current position point and the starting look-ahead point.

[0060] It should be noted that the end point of the obstacle fitting curve is offset from the target offset direction by the preset obstacle avoidance margin to obtain the end point of the target obstacle avoidance curve trajectory; the start point of the obstacle fitting curve is offset from the target offset direction by the preset obstacle avoidance margin to obtain the start point of the target obstacle avoidance curve trajectory; the end point of the obstacle fitting curve and the end point of the target obstacle avoidance curve trajectory are end points of the corresponding curves terminating extension in the reference detection area, and the end point of the start of the extension of the target obstacle avoidance curve trajectory is the start point of the target obstacle avoidance curve trajectory, so that the extension direction of the target obstacle avoidance curve trajectory is from the start point of the target obstacle avoidance curve trajectory to the end point of the target obstacle avoidance curve trajectory.

[0061] On the basis of the above-mentioned embodiments, in step S4, the method of walking along the target obstacle avoidance curve trajectory comprises: after the robot moves to the starting look-ahead point, the robot determines a target look-ahead point located in front of the robot based on the rule that the coordinates unidirectionally change along the extension direction of the target obstacle avoidance curve trajectory from the starting look-ahead point, and then the robot moves to the target look-ahead point. Wherein, the extension direction of the target obstacle avoidance curve trajectory is the same as the extension direction of the obstacle fitting curve, and the tangents at the same longitudinal coordinate position point are parallel to each other.

[0062] Preferably, based on the rule that the coordinates vary unidirectionally along the extension direction of the target obstacle-avoiding curve trajectory, the method for determining the target lookahead point in front of the robot comprises: the robot calculates the adjacent longitudinal coordinate of the starting lookahead point in the positive direction of the body longitudinal axis, then calculates the transverse coordinate of the point on the obstacle fitting curve with the same longitudinal coordinate as the adjacent longitudinal coordinate of the starting lookahead point by substituting the adjacent longitudinal coordinate of the starting lookahead point into the trajectory equation of the obstacle fitting curve, and obtains the target lookahead point of the robot by offsetting the point on the obstacle fitting curve with the same longitudinal coordinate as the adjacent longitudinal coordinate of the starting lookahead point along the target offset direction by the preset obstacle avoidance margin, including the coordinates of the target lookahead point in the robot coordinate system. Alternatively, since the target obstacle-avoiding curve trajectory is obtained by offsetting the obstacle fitting curve along the target offset direction by the preset obstacle avoidance margin, the trajectory equation of the obstacle fitting curve is also subjected to corresponding offset processing, including controlling the transverse coordinate of each control point to offset the preset obstacle avoidance margin along the target offset direction to obtain the trajectory equation of the obstacle fitting curve after offset (also generated according to the offset starting point of the obstacle fitting curve, the offset end point of the obstacle fitting curve, and the aforementioned sequentially marked offset control points to generate an n-order Bezier curve), as the trajectory equation of the target obstacle-avoiding curve trajectory (belongs to the n-order Bezier curve equation); then the adjacent longitudinal coordinate of the starting lookahead point is substituted as the longitudinal coordinate of the target lookahead point into the trajectory equation of the target obstacle-avoiding curve trajectory to obtain the transverse coordinate of the target lookahead point, which is regarded as the target lookahead point of the robot, including the coordinates of the target lookahead point in the robot coordinate.

[0063] Whenever the robot moves to the target look-ahead point, the next target look-ahead point in front of the robot is determined on the target obstacle-avoiding curve trajectory based on the rule that the coordinates vary unidirectionally along the extension direction of the target obstacle-avoiding curve trajectory, and then the next target look-ahead point is updated as the target look-ahead point, and then the robot moves to the updated target look-ahead point, and the iteration is performed until the robot moves to the end point of the target obstacle-avoiding curve trajectory, and it is determined that the obstacle walking around the obstacle profile points framed by the reference detection area is completed. The iteration action mentioned in the embodiment is to calculate the coordinates of each target look-ahead point in turn along the extension direction of the target obstacle-avoiding curve trajectory or the extension direction of the obstacle fitting curve, and the vertical coordinate difference between the target look-ahead points participating in calculation in adjacent two times is fixed. Based on the rule that the coordinates vary unidirectionally along the extension direction of the target obstacle-avoiding curve trajectory, the method for determining the next target look-ahead point in front of the robot is the same as the calculation method required for the target look-ahead point determined by the same rule, except that the target look-ahead point is updated as the starting look-ahead point to correspond to the calculation of the next target look-ahead point, until the end point of the target obstacle-avoiding curve trajectory is calculated, that is, the next target look-ahead point calculated is the result of offsetting the preset obstacle avoidance margin from the end point of the obstacle fitting curve along the target offset direction, so as to walk around the obstacle in which the obstacle profile points framed by the reference detection area by walking along the extension direction of the obstacle fitting curve or the target obstacle-avoiding curve trajectory.

[0064] It should be noted that the end point of the target obstacle-avoiding curve trajectory in FIG. 1 can be obtained by translating the position point B1 along the negative direction of the X axis by the preset obstacle avoidance margin D, and is located at the end point of the uppermost virtual curve of A2; the end point of the target obstacle-avoiding curve trajectory in FIG. 2 can be obtained by translating the position point B3 along the negative direction of the X axis by the preset obstacle avoidance margin D to obtain the position point B4. Therefore, in order to prevent the robot from colliding with the obstacle during obstacle avoidance, the obstacle fitting curve is offset in the direction of the starting point of the obstacle avoidance by a distance greater than the body radius to obtain the target look-ahead point for the robot to walk around, and then the robot is controlled to reach a certain target look-ahead point in front of it from its current position, and the points in front of the robot are tracked in turn according to the target obstacle-avoiding curve trajectory, so that the robot can track the nodes in the same curve extension direction according to the historical marking of the obstacle profile points, and no collision avoidance or obstacle avoidance is achieved in the tracking process.

[0065] After the robot moves to the end point of the target obstacle-avoiding curve trajectory, if the reference detection area set by the robot covers a new ground area, then the step S3 is performed to fit a Bezier curve for the obstacle profile points framed in the new ground area, to obtain a new obstacle fitting curve; then the step S4 is performed. For example, the reference detection area set by the robot in FIG. 2 covers other obstacle profile points (there are pre-detected obstacle profile points) above the right of the position point B3, and the obstacle profile points above the right of the position point B2 have not been fitted into an obstacle fitting curve. In order to make the robot completely avoid the side obstacle #2, the step S3 is performed to fit a Bezier curve for the obstacle profile points framed in the new ground area, to obtain a new obstacle fitting curve; then the step S4 is performed to obtain a new target obstacle-avoiding curve trajectory, and to make the robot walk along the extension direction of the new obstacle fitting curve in the process of moving along the new target obstacle-avoiding curve trajectory without touching the obstacle. The ground area is parallel to the moving plane of the robot. There can be an overlapping area between the ground areas covered by the reference detection area before and after the movement of the robot, but since the robot has moved to the end point of a target obstacle-avoiding curve trajectory, the reference detection area has moved to the right of the end point B2 of the obstacle fitting curve as a whole, so most of the obstacle profile points framed in the ground area newly covered by the reference detection area have not been fitted by the step S3. Therefore, after the obstacle fitting curve fitted in the step S3 is completed to the end point of a target obstacle-avoiding curve trajectory, a new obstacle fitting curve (the extension direction and the curvature will change) is fitted by the step S3 again, and then the robot walks along the extension direction of the new obstacle fitting curve in the step S4. In some embodiments, in the process of moving of the robot, when the new ground area covered by the reference detection area set by the robot does not overlap with the ground area originally covered by the reference detection area, the step S3 is performed to fit a Bezier curve for the obstacle profile points in the new ground area, to obtain a new obstacle fitting curve, and then the start point of the new obstacle fitting curve is connected with the end point of the obstacle fitting curve fitted in the last execution of the step S3, and then the connected obstacle fitting curve is offset by the preset obstacle avoidance margin in the target offset direction, to obtain a new target obstacle-avoiding curve trajectory.

[0066] In summary, on the target obstacle-avoiding curve trajectory, the robot starts from the current position point, searches for a look-ahead point in front of the robot along the extension direction of the target obstacle-avoiding curve trajectory, and takes the look-ahead point as the next target obstacle-avoiding position point on the target obstacle-avoiding curve trajectory, until the end point of the target obstacle-avoiding curve trajectory is searched and walked to, so as to achieve collision-free obstacle avoidance or obstacle avoidance. On this basis, whenever a new obstacle profile point is framed in the detection area, a new obstacle fitting curve is fitted in the new ground area (a semi-closed area surrounded by the connecting line of the new control points) in a Bézier curve fitting manner, so as to facilitate connection and offset to form a new target obstacle-avoiding curve trajectory which is relatively smooth, and to realize continuous walking of the robot along the profile of the side obstacle while tracking the side obstacle of the robot in a collision-free manner relying on the fitting curve.

[0067] As an embodiment, as shown in FIG. 1, the reference detection area EFGH has framed all the obstacle profile points on the right profile of obstacle #1; point B1 is located on the boundary HG of the reference detection area EFGH, at this time, the starting point of the obstacle fitting curve U1 fitted by performing step S3 is position point A1, and the end point of the obstacle fitting curve U1 fitted by performing step S3 is position point B1. In this embodiment, the vertical distance from position point A1 to the boundary EF of the reference detection area EFGH is set as the maximum target offset allowed in the positive direction of the longitudinal coordinate axis for the reference detection area EFGH to frame all the obstacle profile points on the right profile of obstacle #1 (all the discrete points distributed between position point A1 and position point B1); during movement of the robot, if the longitudinal coordinate offset of the reference detection area EFGH in the positive direction of the longitudinal coordinate axis does not exceed the maximum target offset, the obstacle profile points framed by the reference detection area EFGH before and after movement of the robot (which can be two look-ahead points adjacent in the longitudinal coordinate) do not change, and the look-ahead points moved by the robot in step S4 are nodes with sequentially increased longitudinal coordinates in the same target obstacle-avoiding curve trajectory. Therefore, under the premise that the longitudinal coordinate offset of the reference detection area EFGH in the positive direction of the longitudinal coordinate axis does not exceed the maximum target offset, after the robot moves through the end point B2 of the target obstacle-avoiding curve trajectory, the obstacle fitting curve obtained in step S3 before and after movement of the robot does not change, and the target obstacle-avoiding curve trajectory A2B2 formed in FIG. 1 is parallel to the obstacle fitting curve U1 obtained in step S3 before and after movement of the robot.

[0068] When the longitudinal coordinate offset of the reference detection region EFGH in the positive direction of the longitudinal coordinate axis has exceeded the maximum target offset, and the robot has already moved past the end point of the target obstacle-avoiding curve trajectory, the reference detection region EFGH does not frame all the obstacle contour points on the right contour of the obstacle #1, the number of obstacle contour points currently framed by the reference detection region EFGH is reduced relative to the number of obstacle contour points (all the obstacle contour points on the right contour of the obstacle #1) framed by the reference detection region EFGH in FIG. 1, the number of control points required by the obstacle fitting curve fitted in the current execution of step S3 can be reduced, therefore, the obstacle fitting curve fitted in step S3 executed when the reference detection region EFGH has framed all the obstacle contour points on the right contour of the obstacle #1 is different from the obstacle fitting curve fitted in the current execution of step S3.

[0069] It should be noted that the rule that the coordinate varies in one direction along the extension direction of the target obstacle-avoiding curve trajectory includes: from the starting look-ahead point, the coordinate offset of the longitudinal coordinate of the target look-ahead point determined in the current execution relative to the longitudinal coordinate of the target look-ahead point determined in the last execution is equal to the value 1, then the extension direction of the target obstacle-avoiding curve trajectory is the direction in which the longitudinal coordinate of the look-ahead point in the target obstacle-avoiding curve trajectory increases, wherein the coordinate offset of the longitudinal coordinate of the initially determined target look-ahead point relative to the longitudinal coordinate of the starting look-ahead point is equal to the value 1; wherein the positive direction of the longitudinal axis of the body is the positive direction of the longitudinal coordinate axis in the coordinate system set with the center of the body of the robot as the origin.

[0070] In some embodiments, the difference between the longitudinal coordinate of the adjacent position point of the starting look-ahead point in the positive direction of the longitudinal axis of the body and the longitudinal coordinate of the starting look-ahead point is equal to the value 1, and the adjacent position point of the starting look-ahead point in the positive direction of the longitudinal axis of the body has the same horizontal coordinate as the starting look-ahead point, so that the absolute value of the difference between the longitudinal coordinates of the two look-ahead points obtained in step S4 in succession is equal to the value 1.

[0071] Specifically, the adjacent position points of the starting lookahead point include: the position points with an absolute value of the difference of the horizontal coordinate of the starting lookahead point equal to 1 and an absolute value of the difference of the vertical coordinate of the starting lookahead point equal to 1, denoted as adjacent corner points, there are four kinds, distributed at the upper left corner position, the lower left corner position, the upper right corner position and the lower right corner position of the starting lookahead point. The position points with an absolute value of the difference of the horizontal coordinate of the starting lookahead point equal to 1 and an absolute value of the difference of the vertical coordinate of the starting lookahead point equal to 0, denoted as horizontal adjacent position points, there are two kinds, distributed on the left and right sides of the starting lookahead point and the connecting line is parallel to the positive direction of the body horizontal axis. The position points with an absolute value of the difference of the horizontal coordinate of the starting lookahead point equal to 0 and an absolute value of the difference of the vertical coordinate of the starting lookahead point equal to 1, denoted as vertical adjacent position points, there are two kinds, distributed on the upper and lower sides of the starting point of the around-obstacle walking and the connecting line is parallel to the positive direction of the body vertical axis. Thus, the four adjacent corner points, the two vertical adjacent position points and the two horizontal adjacent position points surround one starting lookahead point to form an eight-neighbor domain, equivalent to the layout form of a nine-guanyuan grid, which is conducive to helping the robot walk out of the trajectory with strong continuity in step S4; wherein the positive direction of the body vertical axis is the positive direction of the vertical coordinate axis in the coordinate system with the center of the robot body as the origin, which can be further understood as the moving direction of the robot being the positive direction of the vertical coordinate axis in the coordinate system with the center of the robot body as the origin.

[0072] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments of the method. In the embodiments provided in the present application, the references to memory, storage, database or other media can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable memory (PROM), electrically programmable memory (DPROM), electrically erasable programmable memory (DDPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory.

[0073] The technical features of the above embodiments can be combined in any way. In order to make the description concise, not all possible combinations of the above embodiments are described, but as long as the combinations of the technical features do not exist, they should be considered as the scope of the present application.

[0074] The above embodiments only express several embodiments of the present application, and the description is more specific and detailed, but it cannot be understood as a limitation of the scope of the patent. It should be noted that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of the present application.

Claims

1. A robot obstacle avoidance control method based on Bezier curve fitting, characterized by, The robot obstacle-surrounding control method comprises: Step S1, the robot emits linear laser to its front side; Step S2, the robot collects the obstacle contour points detected by the linear laser in real time and obtains the position coordinates of the obstacle contour points; Step S3, a reference detection area is set on the side of the robot close to the obstacle, and the position coordinates of the obstacle contour points framed by the reference detection area are used to perform Bezier curve fitting to generate an obstacle fitting curve; Step S4, the robot offsets the obstacle fitting curve by a preset obstacle-avoiding margin to obtain a target obstacle-surrounding curve trajectory, and then moves from the current position point to the target obstacle-surrounding curve trajectory and moves along the target obstacle-surrounding curve trajectory.

2. The method of claim 1, wherein, In step S2, the obstacle contour points detected by the linear laser are formed by the reflection of the linear laser emitted by at least one linear laser emitter in the contour of the obstacle, and the obstacle contour points detected by the linear laser are used to connect to form the contour of the obstacle detected by the linear laser; The linear laser emitter is installed at the front end of the side of the robot; The robot comprises a semicircular head, a semicircular body and two symmetrically arranged wheels, the two symmetrically arranged wheels are connected through an axle, and the axle is arranged at the boundary line between the head and the body; the semicircular head and the semicircular body are configured into a circular body; the center of the axle is the center of the body of the robot; the length of the axle is less than the diameter of the body; The moving direction of the robot is directed to the front side of the robot, and the moving direction of the robot is parallel to the positive direction of the body longitudinal axis, and the positive direction of the body longitudinal axis is perpendicular to the axle; The positive direction of the body transverse axis is in the direction parallel to the axle and towards the side obstacle of the robot, and the side obstacle is an obstacle located outside the side of the body of the robot, and the side of the body of the robot includes the left side or the right side of the body.

3. The method of claim 2, wherein, In step S3, a reference detection area is set on the side of the robot close to the obstacle to cover the contour of the obstacle; The straight-line distance between the boundary point farthest from the edge of the side of the body close to the obstacle and the center of the body in the positive direction of the body transverse axis in the reference detection area is greater than the radius of the body to cover the obstacle contour points outside the body; The reference detection area is set to cover part or all of the passable area between the body and the side obstacle; the straight-line distance covered by the reference detection area in the positive direction of the body longitudinal axis does not exceed the maximum detection distance of the linear laser emitted by the linear laser emitter; The area covered by the reference detection area outside the body is greater than the area covered by the reference detection area inside the body.

4. The method of claim 3, wherein, When the reference detection area is a polygonal area, at least one area boundary of the reference detection area is parallel or perpendicular to the axle. When the reference detection area overlaps with the coverage area of the robot body, the coverage area of the reference detection area in the head part is greater than that in the body part; and in the area covered by the reference detection area outside the robot body, the coverage area of the reference detection area on the head side is greater than that on the body side.

5. The method of claim 4, wherein, When the reference detection area is a rectangular area, the setting method of the reference detection area comprises: selecting a position point at a first preset longitudinal distance from the center of the robot body in the positive direction of the longitudinal axis of the robot body and at a first preset transverse distance from the center of the robot body in the positive direction of the transverse axis of the robot body as the upper left corner point; selecting a position point at a second preset longitudinal distance from the center of the robot body in the reverse direction of the longitudinal axis of the robot body and at a first preset transverse distance from the center of the robot body in the positive direction of the transverse axis of the robot body as the lower left corner point; selecting a position point at a first preset longitudinal distance from the center of the robot body in the positive direction of the longitudinal axis of the robot body and at a second preset transverse distance from the center of the robot body in the positive direction of the transverse axis of the robot body as the upper right corner point; selecting a position point at a second preset longitudinal distance from the center of the robot body in the reverse direction of the longitudinal axis of the robot body and at a second preset transverse distance from the center of the robot body in the positive direction of the transverse axis of the robot body as the lower right corner point; then connecting the lower left corner point with the upper left corner point, connecting the upper left corner point with the upper right corner point, connecting the upper right corner point with the lower right corner point, and connecting the lower right corner point with the lower left corner point to obtain the reference detection area; wherein the first preset transverse distance is set to be greater than half the length of the wheel shaft and less than the radius of the robot body; the boundary length of the reference detection area in the positive direction of the transverse axis of the robot body is equal to the preset obstacle avoidance margin; the first preset longitudinal distance is set to be greater than or equal to the radius of the robot body but less than the maximum detection distance of the linear laser emitted by the linear laser emitter; the second preset longitudinal distance is set to be less than the radius of the robot body or half the length of the wheel shaft.

6. The method of claim 3, wherein, In the step S3, the method for performing Bezier curve fitting by using the position coordinates of the obstacle contour points framed in the reference detection area comprises: determining the starting point of the obstacle fitting curve and the ending point of the obstacle fitting curve among the obstacle contour points framed in the reference detection area along the positive direction of the longitudinal axis of the robot body; and marking each obstacle contour point distributed between the starting point and the ending point along the positive direction of the longitudinal axis of the robot body in the reference detection area as a control point required for fitting the Bezier curve in sequence, wherein the obstacle fitting curve belongs to the Bezier curve; the number of control points is equal to the sum of the order of the Bezier curve and the value 1; the order of the Bezier curve is represented by n; generating the trajectory equation of the obstacle fitting curve based on the n-order Bezier curve according to the starting point of the obstacle fitting curve, the ending point of the obstacle fitting curve, and the control points marked in sequence between the starting point of the obstacle fitting curve and the ending point of the obstacle fitting curve, wherein the starting point of the obstacle fitting curve and the ending point of the obstacle fitting curve both belong to the control points.

7. The method of claim 6, wherein, The trajectory equation of the obstacle fitting curve: wherein B x (t) is the abscissa of a point in the obstacle fitting curve, B y (t) is the ordinate of a point in the obstacle fitting curve, i denotes the ordinal number of a control point required for the obstacle fitting curve, and t is a ratio equal to i and (n+1); P 0x denotes the abscissa of the start position of the obstacle fitting curve within the reference detection area, P 0y denotes the ordinate of the start position of the obstacle fitting curve within the reference detection area; P nx denotes the abscissa of the end position of the obstacle fitting curve within the reference detection area, P ny denotes the ordinate of the end position of the obstacle fitting curve within the reference detection area; wherein P ix represents the horizontal coordinate of the obstacle profile point within the reference detection area for setting the control point P i represents the vertical coordinate of the obstacle profile point within the reference detection area for setting the control point P iy represents the horizontal coordinate of the obstacle profile point within the reference detection area for setting the control point P i represents the vertical coordinate of the obstacle profile point within the reference detection area for setting the control point P 8. The method of claim 7, wherein, In the process of the Bezier curve fitting, the number of the value categories of t is equal to the number of the points fitted by the obstacle fitting curve; In the positive direction of the longitudinal axis of the machine body, the relative position of two adjacent control points changes, and the curvature of the obstacle fitting curve changes; two adjacent control points use P i and P i+1 represent, wherein i ∈ [0, n-1].

9. The method of claim 4, wherein, The starting point of the obstacle fitting curve is the closest obstacle profile point to the boundary of the reference detection area in the opposite direction of the positive direction of the body longitudinal axis, or the boundary point of the reference detection area through which the profile of the obstacle represented by the obstacle fitting curve passes in the opposite direction of the positive direction of the body longitudinal axis; The ending point of the obstacle fitting curve is the closest obstacle profile point to the boundary of the reference detection area in the positive direction of the body longitudinal axis, or the boundary point of the reference detection area through which the profile of the obstacle represented by the obstacle fitting curve passes in the positive direction of the body longitudinal axis.

10. The method of claim 9, wherein, In step S4, if the profile of the obstacle represented by the obstacle fitting curve is located in the positive direction of the body transverse axis of the robot, the opposite direction of the positive direction of the body transverse axis is configured as the target offset direction, and each point in the obstacle fitting curve is offset by a preset obstacle avoidance margin along the target offset direction to obtain the target obstacle-avoiding curve trajectory. The preset obstacle avoidance margin is set to be equal to the sum of the body radius and the preset obstacle-avoiding distance.

11. The method of claim 10, wherein: In step S4, the method for the robot to move from the current position point to the target obstacle-avoiding curve trajectory includes: calculating the adjacent longitudinal coordinate of the current position point in the positive direction of the body longitudinal axis, and setting the adjacent longitudinal coordinate as the starting look-ahead longitudinal coordinate; then substituting the starting look-ahead longitudinal coordinate into the trajectory equation of the obstacle fitting curve to calculate the transverse coordinate of the point in the obstacle fitting curve with the same longitudinal coordinate as the starting look-ahead longitudinal coordinate, and obtaining the position coordinate of the point in the obstacle fitting curve with the same longitudinal coordinate as the starting look-ahead longitudinal coordinate; then offsetting the point in the obstacle fitting curve with the same longitudinal coordinate as the starting look-ahead longitudinal coordinate by the preset obstacle avoidance margin along the target offset direction to obtain the starting look-ahead point of the robot, and determining that the starting look-ahead point of the robot is a point on the target obstacle-avoiding curve trajectory; then the robot moves from the current position point to the starting look-ahead point to enter the target obstacle-avoiding curve trajectory.

12. The method of claim 11, wherein: In step S4, the method for moving along the target obstacle-avoiding curve trajectory includes: after the robot moves to the starting look-ahead point, the robot determines the target look-ahead point located in front of the robot based on the rule that the coordinates change unidirectionally along the extension direction of the target obstacle-avoiding curve trajectory, and then the robot moves to the target look-ahead point; whenever the robot moves to the target look-ahead point, the next target look-ahead point located in front of the robot is determined based on the rule that the coordinates change unidirectionally along the extension direction of the target obstacle-avoiding curve trajectory, the next target look-ahead point is updated as the target look-ahead point, and then the robot moves to the updated target look-ahead point, and the iteration is repeated until the robot moves to the ending point of the target obstacle-avoiding curve trajectory, and it is determined that the walking around the obstacle represented by the profile of the obstacle profile point framed by the reference detection area is completed. The end point of the obstacle fitting curve is offset from the preset obstacle avoidance margin along the target offset direction to obtain an end point of the target obstacle-avoiding curve trajectory; the start point of the obstacle fitting curve is offset from the preset obstacle avoidance margin along the target offset direction to obtain a start point of the target obstacle-avoiding curve trajectory; the end point of the obstacle fitting curve and the end point of the target obstacle-avoiding curve trajectory are end points at which the corresponding curves terminate extending in the reference detection area, and the end point at which the target obstacle-avoiding curve trajectory starts extending is the start point of the target obstacle-avoiding curve trajectory, so that the extending direction of the target obstacle-avoiding curve trajectory is from the start point of the target obstacle-avoiding curve trajectory to the end point of the target obstacle-avoiding curve trajectory.

13. The method of claim 12, wherein: After the robot moves to the end point of the target obstacle-avoiding curve trajectory, if the reference detection area set by the robot covers a new ground area, the Bezier curve fitting is performed in the new ground area according to the step S3 to obtain a new obstacle fitting curve; and then the step S4 is performed.

14. The method of claim 12, wherein: In the process of robot movement, when a new ground area covered by the reference detection area set by the robot has no overlapping area with a ground area originally covered by the reference detection area, the Bezier curve fitting is performed in the new ground area according to the step S3 to obtain a new obstacle fitting curve, the start point of the new obstacle fitting curve is connected with the end point of the obstacle fitting curve fitted last time, and the connected obstacle fitting curve is offset from the preset obstacle avoidance margin along the target offset direction to obtain a new target obstacle-avoiding curve trajectory.

15. The method of claim 12, wherein: The rule that the coordinates change in one direction along the extending direction of the target obstacle-avoiding curve trajectory comprises: From the starting look-ahead point, a coordinate offset amount of the vertical coordinate of the target look-ahead point determined this time relative to the vertical coordinate of the target look-ahead point determined last time is equal to a value 1; wherein a coordinate offset amount of the vertical coordinate of the initially determined target look-ahead point relative to the vertical coordinate of the starting look-ahead point in the positive direction of the body vertical axis is equal to the value 1; The positive direction of the body vertical axis is a vertical coordinate axis positive direction in a coordinate system with the body center of the robot as an origin. The positive direction of the body vertical axis is a vertical coordinate axis positive direction in a coordinate system with the body center of the robot as an origin.

Images