A curve control method and system for ground mobile robot

Abstract

The present invention proposes a curve control method and system for a ground mobile robot, the method comprising the following steps: S1, obtaining a walking path through a path planning algorithm; S2, extracting curves with small spans and large curvatures in the walking path, and outputting the entry and exit points of the corresponding curves; S3, controlling the robot to decelerate when it approaches the entry point, controlling the robot to maintain a low speed when entering the curve, and controlling the robot to speed up when it reaches the exit point. This application provides a method for extracting the distance and angle information between each differential path segment in the path generated by path planning, so as to enable the robot to perform deceleration processing during autonomous navigation of the curved path.

Description

Technical Field

[0001] The present invention relates to the field of mobile robot walking control, and in particular to a curve control method and system for a ground mobile robot. Background Art

[0002] Autonomous mobile robots are widely used in various scenarios, such as logistics factories, specialized environments, and military operations. Specific applications include transportation, exploration, search and rescue, data collection, equipment operation, cleanup, and combat operations. Robots often encounter narrow areas during operation, which can affect their performance. High-speed robots are particularly prone to collisions around narrow, steep curves.

[0003] Common local planning methods for mobile robots, such as DWA, TEB, and MPC, have some shortcomings when dealing with narrow-span, high-curvature curves. They lack the ability to plan specifically for these curves. Common local planning methods for narrow-span, high-curvature curves can be described as discrete, segmented processes: target tracking, obstacle avoidance, target tracking, obstacle avoidance, and so on, until the robot navigates the curve. However, the DWA method does not explicitly consider the characteristics of the curve and may not be able to effectively plan paths for curves with high curvature. The TEB method, based on elastic band theory, plans paths by adjusting the robot's trajectory over time. The MPC method uses models to predict the robot's state, typically using linearized models, which struggle to accurately describe the nonlinear characteristics of curves with high curvature.

[0004] In various scenarios involving autonomous mobile robots, path planning and trajectory planning are core steps in achieving autonomous navigation. The target pose at each moment of robot motion is provided by the path planning function, while the target velocity is provided by the trajectory planning function. Specifically, path planning algorithms typically use a graph search to obtain the shortest path from the robot's current position to the target position. This path consists of a set of discrete spatial coordinate points. The trajectory planner uses these spatial coordinate points as local targets and imposes soft and hard constraints on the planning process based on robot kinematics, the distance between the robot and obstacles, robot motion efficiency, and robot hardware parameters. The planner then generates a trajectory in real time to drive the robot's movement. Although the robot can safely navigate narrow, highly curvatured paths such as U-turns after obstacle expansion, the path generation process only considers the shortest path property and does not account for the robot's velocity or spatial dimensions. In other words, the robot is unaware that it is navigating a curve and simply tracks the target path. Therefore, it is easy for the robot to overshoot the curve while traveling straight ahead, deviating from the path points along the curve, and then having to adjust itself again after the state estimate is updated to return to the next target point. This wastes both space and energy in the robot.

[0005] In the prior art, patent application number CN201910180624.9, entitled "A Method and Apparatus for Curving Speed ​​Planning for an Automatic Guidance Device," describes a method that obtains path information in real time, calculates the curvature range of the path, and linearly maps the device's speed to the path curvature to achieve different speed configurations for paths with different curvatures. However, for autonomous mobile robots, the only condition for the robot to turn under normal circumstances is that the global planner has planned the turning path based on an offline map. This task is completed when the robot starts navigation, and the location where the robot needs to turn can be determined based on this path. Therefore, it is unnecessary for the robot to calculate the path curvature in real time during movement. Secondly, the path is a set of discrete coordinate points. To calculate the curvature, curve fitting must first be performed, and then the curvature is calculated by differentiating the fitted function, which is a relatively cumbersome process. Third, there is a time delay between the path analysis process and the control process. The method does not clearly specify how to determine the entry and exit points of the curve, or where to start speed control. Fourth, speed control based solely on path curvature conditions is not comprehensive.

[0006] The purpose of this invention is to design a curve control method and system for a ground mobile robot in response to the above-mentioned problems in the prior art. Summary of the Invention

[0007] The purpose of the present invention is to propose a curve control method and system for a ground mobile robot, a method for extracting the distance and angle information between each segment of the differential path in the path generated by path planning, so as to enable the robot to perform deceleration processing during autonomous navigation of the curved path.

[0008] To achieve the above object, the technical solution of the present invention is:

[0009] A method for controlling a ground mobile robot on a curve comprises the following steps:

[0010] S1, obtain the walking path through the path planning algorithm;

[0011] S2, extracting a curve with a small span and a large curvature in the walking path, and outputting an entry point and an exit point of the corresponding curve;

[0012] S3, controlling the robot to decelerate when the robot approaches the turning point, controlling the robot to keep walking at a low speed when the robot enters the curve, and controlling the robot to speed up when the robot walks to the turning exit point.

[0013] Furthermore, the walking path includes a number of path points, which are a sequence of plane coordinates in ascending order. The number of the path points is counted. If the number of the path points is lower than a threshold value of the number of path points, a single-curve algorithm is used to extract a curve with a small span and a large curvature in the walking path, and the entry and exit points of the corresponding curve are output. Otherwise, a multi-curve algorithm is used to extract a curve with a small span and a large curvature in the walking path, and the entry and exit points of the corresponding curve are output.

[0014] Furthermore, extracting a curve with a small span and a large curvature in the walking path using a single curve algorithm, and outputting the entry point and exit point of the corresponding curve includes:

[0015] Calculate the directional angle difference between any two micropaths, where the micropath is a line connecting adjacent pathpoints. The direction of the micropath points from the pathpoint with a smaller sequence number to the pathpoint with a larger sequence number. The directional angle of the micropath is bound to the pathpoint with the smaller sequence number, and extract the pathpoint pairs whose angle difference meets the angle threshold condition.

[0016] Calculating the along-path distance of the path point pairs, and extracting the path point pairs whose along-path distance is less than a distance threshold;

[0017] The path point pairs are separated, and the entry and exit points of the corresponding curves are output.

[0018] Furthermore, outputting the entry point and exit point of the corresponding curve includes:

[0019] Establishing a two-dimensional coordinate system to obtain the coordinates of each path point;

[0020] For all the extracted path point pairs whose distance along the path is less than the distance threshold, the coordinates of the starting path points in all the path point pairs are extracted and averaged to obtain the turning point, and the coordinates of the ending path points in all the path point pairs are extracted and averaged to obtain the turning point.

[0021] Furthermore, outputting the entry point and exit point of the corresponding curve includes:

[0022] For all the extracted path point pairs whose distance along the path is less than the distance threshold, the starting path point among all the path point pairs is extracted, and the starting path point with the smallest serial number is used as the turning point; the ending path point among all the path point pairs is extracted, and the ending path point with the largest serial number is used as the turning point.

[0023] Furthermore, extracting a curve with a small span and a large curvature in the walking path using a multi-curve algorithm, and outputting the entry point and exit point of the corresponding curve includes:

[0024] Create a blank image;

[0025] Drawing a binary curve graph on the blank image through all the path points;

[0026] The magnitude and direction of the gradient of each part of the binary curve graph are analyzed to obtain all the curve images in the binary curve graph.

[0027] retaining a plurality of curve segments in the curved road image whose gradient difference is greater than a second angle threshold;

[0028] Extract the two-dimensional coordinates of each curve segment, and take the point with the smallest sequence number in each segment as the entry point, and the point with the largest sequence number in each segment as the exit point.

[0029] Furthermore, before drawing the binary curve graph on the blank image using all the path points, performing: establishing a two-dimensional coordinate system, obtaining the coordinates of each of the path points, and establishing a mapping relationship between the blank image and the two-dimensional coordinate system;

[0030] After retaining a plurality of curve segments in the curved road image whose gradient difference is greater than the second angle threshold, performing: inverse mapping the plurality of curve segments back to the two-dimensional coordinate system according to the mapping relationship.

[0031] Furthermore, after drawing a binary curve graph on the blank image through all the path points, the following steps are performed:

[0032] Smooths non-differentiable path vertices.

[0033] Furthermore, the control method is executed when the robot starts, and if the robot stops during walking, the method returns to step S1.

[0034] Furthermore, the present application also provides a ground mobile robot curve control system, comprising the following modules:

[0035] Path planning module, used to obtain the walking path through the path planning algorithm;

[0036] A key point output module is used to extract curves with small span and large curvature in the walking path and output the entry and exit points of the corresponding curves;

[0037] a speed control module, configured to control the robot to decelerate when the robot approaches the turning point, control the robot to maintain a low speed when the robot enters the curve, and control the robot to resume speed when the robot reaches the turning point;

[0038] The control system executes the curve control method for a ground mobile robot when it is running.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] Based on the reasons that the robot moves too fast or the braking ability of the driving element is insufficient, this application proposes a method for extracting the distance and angle information between each segment of the differential path in the path generated by path planning, so as to enable the robot to perform deceleration processing in the autonomous navigation of the curved part of the path.

[0041] This application determines whether there is a single bend or multiple bends in the walking path by the number of path points, and executes different calculation strategies for single bends or multiple bends. In the case of a single bend, more accurate entry and exit points can be obtained through sufficient calculation amount, while in the case of multiple bends, the calculation method of image analysis can be used to reduce the amount of calculation, thereby obtaining the entry and exit points more quickly.

[0042] This application calculates the directional angle difference and the along-path distance between pairs of path points that meet the angle threshold conditions for each path point in a single curve, thereby screening the corresponding curves and calculating the exit and entry points based on the coordinates of the path point pairs. This application uses image analysis methods to extract the curves of multiple curves and then derive the entry and exit points based on the starting and ending points of the curves. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 For the present invention Figure 1 A schematic diagram of a flow chart of an embodiment of the present invention;

[0044] Figure 2 A schematic diagram of a walking path including a single curve in Embodiment 1 or 2 of the present invention;

[0045] Figure 3 Schematic diagram of analysis of a single curve in Embodiment 1 or 2 of the present invention;

[0046] Figure 4 This is a schematic diagram of a walking path with multiple bends according to the third embodiment of the present invention;

[0047] Figure 5 for Figure 4 Schematic diagram of the analysis results of multiple bends shown. DETAILED DESCRIPTION

[0048] To facilitate understanding by those skilled in the art, the structure of the present invention will now be further described in detail with reference to the embodiments and the accompanying drawings.

[0049] Example 1

[0050] refer to Figure 1 , a curve control method for a ground mobile robot, comprising the following steps:

[0051] S1, obtain the walking path through the path planning algorithm;

[0052] In this embodiment, existing graph search path planning algorithms can be used, such as Dijkstra, A* and A* variants (D*, HybridA*, etc.), refer to Figure 2 In this embodiment, the plane coordinate system is used as an example. The input parameters are the starting point coordinates of the path start and the end point coordinates goal. The output parameters are the point set of the path [(x1, y1),..., (x i ,y i ),...], i=1,2,3,….

[0053] In this embodiment, the point set of the path has a sequence number, for example, the coordinate of the starting point start is 001, the next path point is 002, and so on.

[0054] S2, extracting a curve with a small span and a large curvature in the walking path, and outputting an entry point and an exit point of the corresponding curve;

[0055] In this embodiment, a curve with a small span and a large curvature indicates that there is a sharp bend at this location, and this location requires an early adaptive change to the walking state of the robot.

[0056] S3, controlling the robot to decelerate when the robot approaches the turning point, controlling the robot to keep walking at a low speed when the robot enters the curve, and controlling the robot to speed up when the robot walks to the turning exit point.

[0057] The robot is prone to overshooting in narrow-span, large-curvature curves. There are two main reasons for this problem. First, the robot is too fast or the braking capacity of the drive element is insufficient. Second, the accuracy of the robot system state estimation is insufficient or the frequency of position updates is low. The robot needs to start slowing down here to meet the time required for the robot to calculate the control strategy of the curve's posture, speed, etc. For curves with large spans or small curvatures, the robot's existing algorithms can cope with it and no special processing is required.

[0058] Furthermore, when the robot approaches a corner, its speed is reduced to a specified speed, which is lower than the robot's speed before the corner. When the robot approaches the corner's exit, its speed returns to the default speed of the trajectory planning controller. If the robot is not near the corner's exit, it maintains deceleration until it exits the corner or reaches the target point. The remaining path uses the speed command of the conventional trajectory planning controller.

[0059] Furthermore, the walking path includes a number of path points, which are a sequence of plane coordinates in ascending order. The number of the path points is counted, and a single-curve algorithm is used to extract curves with small spans and large curvatures in the walking path, and the entry and exit points of the corresponding curves are output.

[0060] In this embodiment, a single curve indicates that there is only one curve in the walking path, and the subsequent steps and algorithms provided in this embodiment can be used to reduce the amount of calculation. The threshold value of the number of path points can be a number calculated based on actual experience and is not limited here.

[0061] This embodiment determines the number of bends in the walking path by the number of path points. In combination with the actual environmental space, the number of path points is used to determine the number of bends. It can be considered that when the number of path points is below a certain value, that is, the number of path points is small, the path the robot walks is short, and there is a high probability that there will not be multiple bends, and the calculation method corresponding to the single bend provided subsequently requires a small amount of calculation. In this case, there is enough time for the robot to complete the calculation, and it is assumed that there is only one bend in the walking path. If there are too many path points, there are a high probability that there are multiple bends in the walking path, and the calculation method provided by the single bend takes too long, and it is necessary to use the algorithm provided by Example 3 for calculation.

[0062] Furthermore, if there is a single curve in the walking path, extracting a curve with a small span and a large curvature in the walking path, and outputting the entry point and exit point of the corresponding curve includes:

[0063] S2.1. Calculate the directional angle difference between any two micropaths. A micropath is a line connecting adjacent pathpoints. A micropath points from a smaller pathpoint to a larger pathpoint. The directional angle of each micropath is bound to the pathpoint with the smaller number. Extract pathpoint pairs whose angular difference satisfies the angle threshold.

[0064] In this step, the direction angle of the micropath is the angle between the micropath and the positive direction of a specified coordinate axis of the global coordinate system, such as Figure 2 As shown, this embodiment adopts the angle αi between the micropath and the X-axis direction. The direction angle of the micropath is bundled with the path point with a small sequence number, for example, the angle αi of the micropath formed by the third path point and the fourth path point is i , then the angle α i Bind to the 3rd waypoint.

[0065] Then, calculate the direction angle difference threshold between any two micro-paths deg For example, the directional angle difference between the micro-path composed of the 3rd and 4th path points and the micro-path composed of the 8th and 9th path points. The angle threshold condition can be set as: 180° ≥ threshold deg ≥90°, that is, when the directional angle difference between any two micro-paths is within this range, it is considered that the path between the two points may have a curve with a small span and a large curvature. At this time, it is necessary to determine the existence of a curve with a small span and a large curvature through subsequent steps.

[0066] The angle of any micro-segment in the path relative to the global coordinate system can be calculated using the following formula:

[0067]

[0068] Here, ε is a small quantity to avoid unsolvable solutions when calculating the inverse tangent. Considering the properties of the tangent function, to prevent ambiguity in angle descriptions, angular directions that are obtuse (>90°) clockwise or counterclockwise relative to the positive x-axis are converted using π and -π after the inverse tangent calculation.

[0069] S2.2, calculating the along-path distances of the path point pairs, and extracting the path point pairs whose along-path distances are less than a distance threshold;

[0070] In this step, the basis for determining whether the path is a curve with a small span and a large curvature is that the heading change that meets the angle threshold condition must occur within a relatively short path span. If it occurs between two path points with a relatively long path span, and the robot heading between the two relatively long path points makes a large turn, the path distance between the two points is relatively far, and the curve can be transitioned more smoothly. At this time, it cannot be determined that it is a curve with a small span and a large curvature, and the robot does not need to perform speed processing.

[0071] Set the distance threshold s This value needs to be determined based on the actual robot application environment. If it is too large, the robot efficiency will be insufficient; if it is too small, the robot is still prone to overshoot.

[0072] Calculating the distance along the path of the path point pair can be achieved by the following steps:

[0073] 1. Calculate the Euclidean distance between every two adjacent points in the path;

[0074]

[0075] 2. Add up the distances of each micropath in the path;

[0076] Any two path points (x a ,y a ) and (x b ,y b ) along the path is

[0077] That is, when points a and b are the starting point and end point respectively, S is the length of the entire path.

[0078] S2.3, separating the path point pairs and outputting the entry and exit points of the corresponding curves.

[0079] In steps S2.1 to S2.2, two path points (x a ,y a ) and (x b ,y b ), the corresponding two path points form a path point pair, each path point pair contains a starting path point (x a ,y a ) and an ending path point (x b ,y b ), where a <b。

[0080] In step S2.2, a pair of path points is obtained. This step separates the pair of path points to obtain the starting path point and the ending path point.

[0081] refer to Figure 3 , the above steps yielded Figure 3 The path point bundled with direction 1 is shown as the turning point, and the path point bundled with direction 2 is shown as the turning point.

[0082] Furthermore, outputting the entry point and exit point of the corresponding curve includes:

[0083] Establishing a two-dimensional coordinate system to obtain the coordinates of each path point;

[0084] The coordinates of the starting path point in all the path point pairs are extracted and averaged to obtain the turning point, and the coordinates of the ending path point in all the path point pairs are extracted and averaged to obtain the turning point.

[0085] In the above steps, after the above calculations, a set of turning points that meet the angle threshold conditions and distance threshold conditions can be obtained. and exit point The two sets of data are one-to-one corresponding. In order to make the information obtained by the robot during movement clearer and more concise and reduce the number of calculations, the two sets of data are processed as follows:

[0086] 1. Turn-in point Take the average of all the turning points as the only turning point:

[0087]

[0088] 2. Turn exit point Take the average of all the exit points as the only exit point:

[0089]

[0090] Furthermore, the control method is executed when the robot starts, and if the robot stops during walking, the method returns to step S1.

[0091] In this embodiment, in order to prevent the path update from overwriting the already calculated deceleration points (turning points) and speed recovery points (turning points), and to prevent the calculation during the robot's movement from taking up too much time for speed generation (the longer the path, the more time it takes), the collection and calculation of the path points are only performed when the robot stops and only once, such as when the robot starts walking, or when the robot encounters an obstacle that it cannot bypass and stops and replans the path. In other words, the path points are not analyzed during the robot's operation. When the robot reaches the target posture and stops, and is about to go to the next posture, the path point data is collected and analyzed again.

[0092] Example 2

[0093] This step is different from the first embodiment in that the calculation method of outputting the entry point and exit point of the corresponding curve is different, and other contents are basically the same as the first embodiment.

[0094] The number of the path points is greater than a path point number threshold, and a multi-curve algorithm is used to extract curves with a small span and a large curvature in the walking path, and an entry point and an exit point of the corresponding curve are output.

[0095] Furthermore, extracting a curve with a small span and a large curvature in the walking path using a multi-curve algorithm, and outputting the entry point and exit point of the corresponding curve includes:

[0096] Extract the starting path point from all path point pairs, and use the starting path point with the smallest serial number as the turning point; extract the ending path point from all path point pairs, and use the ending path point with the largest serial number as the turning point.

[0097] For example, if the starting points of the path point pair are the second, third, and fourth path points, and the ending points are the seventh, eighth, and tenth path points, the second path point is output as the turning point, and the tenth path point is output as the turning point.

[0098] Example 3

[0099] If the number of the path points is greater than the path point number threshold, it is determined that there are multiple bends in the walking path.

[0100] In this embodiment, if the number of path points is greater than the path point number threshold, it means that there are multiple bends in the path point, and the key path point identification method for a multi-bend path is different from that for a single-bend path. Since the path length is longer and there are more path points, if the distance or angle along the path between any two path points is calculated according to the single-bend algorithm, the consumption of computing resources is quite large. In addition, even if the required parameters can be calculated within the allowed time, for the obtained key point list, directly calculating the mean cannot obtain the results of multiple bend entry and exit points, and using K-means for clustering will significantly increase the computational burden and cannot guarantee the reliability of the clustering results. Therefore, this article proposes an image processing recognition method based on OpenCV for multi-bend paths to obtain key path points.

[0101] Therefore, this embodiment uses the following steps to perform calculations.

[0102] Furthermore, if there are multiple bends in the walking path, extracting bends with a small span and a large curvature in the walking path, and outputting entry and exit points of the corresponding bends includes:

[0103] S2.4, create a blank image;

[0104] S2.5, drawing a binary curve graph on the blank image through all the path points;

[0105] In this embodiment, the robot's walking path can be as follows: Figure 4 As shown, through step S2.5, we can get Figure 5 The results shown.

[0106] S2.6, analyzing the magnitude and direction of the gradient of each portion of the binary curve graph to obtain all the curve images in the binary curve graph;

[0107] S2.7, retaining a number of curve segments in the curved road image whose gradient difference is greater than a second angle threshold;

[0108] In this step, by analyzing the amplitude and direction of the gradient of each part of the binary curve image, all the curve part images imgpart in the binary curve image are obtained and compared with the set second angle threshold threshold global Do the comparison and retain the curve image segment points_curve whose gradient difference grad_bias is greater than the second angle threshold i .

[0109]

[0110] Methods for analyzing the magnitude and direction of the gradient of each portion of the binary curve graph include, but are not limited to, using Sobel, Hough, Canny operators, or shape analysis tools such as moments. Using parallel computing includes, but is not limited to, accelerating with GPUs or multi-threaded programming.

[0111] At the same time, the calculation process of this step can be accelerated using parallel computing methods.

[0112] The extracted curves with small span and large curvature can be referred to Figure 5 .

[0113] S2.8, extract the two-dimensional coordinates of each curve segment, and use the point with the smallest sequence number in each segment as the entry point, and the point with the largest sequence number in each segment as the exit point.

[0114] In the above steps, all curve images are extracted through analysis as curves with small span and large curvature. At this time, the two-dimensional coordinates of the two ends of each curve can be directly extracted as the entry point or exit point.

[0115] Furthermore, before drawing the binary curve graph on the blank image using all the path points, performing: establishing a two-dimensional coordinate system, obtaining the coordinates of each of the path points, and establishing a mapping relationship between the blank image and the two-dimensional coordinate system;

[0116] After retaining a plurality of curve segments in the curved road image whose gradient difference is greater than the second angle threshold, performing: inverse mapping the plurality of curve segments back to the two-dimensional coordinate system according to the mapping relationship.

[0117] In this step, the curve can be separated from the background through edge segmentation, and the mapping relationship between image points and coordinate points can be established:

[0118] (x i ,y i )→(pixel_x i ,pixel_y i );

[0119] Among them, (x i ,y i ) is the coordinate point, (pixel_x i ,pixel_y i ) is an image point.

[0120] Then, the image points of different curve segments are inversely mapped back to coordinate points, and the point with the smallest index among each curve segment is taken as the key point for entering the curve, and the point with the largest index is taken as the key point for exiting the curve.

[0121] (pixel_x i ,pixel_y i)→(x i ,y i );

[0122]

[0123]

[0124] Furthermore, after drawing a binary curve graph on the blank image through all the path points, the following steps are performed:

[0125] Smooths non-differentiable path vertices.

[0126] Example 4

[0127] A ground mobile robot curve control system includes the following modules:

[0128] Path planning module, used to obtain the walking path through the path planning algorithm;

[0129] A key point output module is used to extract curves with small span and large curvature in the walking path and output the entry and exit points of the corresponding curves;

[0130] a speed control module, configured to control the robot to decelerate when the robot approaches the turning point, control the robot to maintain a low speed when the robot enters the curve, and control the robot to resume speed when the robot reaches the turning point;

[0131] The control system executes the curve control method for a ground mobile robot when it is running.

[0132] The working principle of this embodiment is the same as that of any one of the embodiments 1 to 3, and will not be described again here.

[0133] It should be noted that in the claims, any reference signs placed between parentheses shall not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claim. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means may be embodied by one and the same item of hardware. The use of the words first, second, third etc. does not indicate any order. These words may be interpreted as names.

[0134] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0135] In the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," "connect," "fixed," etc. should be understood broadly. For example, they may refer to fixed connection, detachable connection, or integration; mechanical connection or electrical connection; direct connection or indirect connection through an intermediate medium; internal communication between two components or interaction between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0136] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms should not be understood as necessarily referring to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0137] The above are preferred embodiments of the present invention. Any changes made according to the technical solution of the present invention, as long as the resulting functions and effects do not exceed the scope of the technical solution of the present invention, shall fall within the scope of protection of the present invention.

Claims

1. A curve control method for a ground mobile robot, characterized in that: The following steps are involved: S1, obtain the walking path through the path planning algorithm; S2, extracting a curve with a small span and a large curvature in the walking path, and outputting an entry point and an exit point of the corresponding curve; S3, controlling the robot to decelerate when the robot approaches the turning point, controlling the robot to maintain a low speed when the robot enters the curve, and controlling the robot to speed up when the robot reaches the turning exit point; The walking path includes a plurality of path points, each of which is a plane coordinate sequence in ascending order. The number of the path points is counted. If the number of the path points is lower than a path point number threshold, a single-curve algorithm is used to extract a curve with a small span and a large curvature in the walking path, and the entry and exit points of the corresponding curve are output. Otherwise, a multi-curve algorithm is used to extract a curve with a small span and a large curvature in the walking path, and the entry and exit points of the corresponding curve are output. A single-curve algorithm is used to extract a curve with a small span and a large curvature in the walking path, and the entry and exit points of the corresponding curve are outputted, including: Calculate the directional angle difference threshold between any two micropaths deg The micropath is a line connecting adjacent path points. The direction of the micropath points from the path point with a smaller sequence number to the path point with a larger sequence number. The direction angle of the micropath is bound to the path point with a smaller sequence number, and the path point pairs whose angle difference meets the angle threshold condition are extracted; the angle threshold condition is set to: 180° ≥ threshold deg ≥90°; Calculating the along-path distance of the path point pairs, and extracting the path point pairs whose along-path distance is less than a distance threshold; Separating the path point pairs and outputting the entry and exit points of the corresponding curves; The multi-curve algorithm is used to extract curves with small span and large curvature in the walking path, and the entry and exit points of the corresponding curves are outputted, including: Create a blank image; Drawing a binary curve graph on the blank image through all the path points; Analyzing the magnitude and direction of the gradient of each part of the binary curve graph to obtain all the curve images in the binary curve graph; retaining a plurality of curve segments in the curved road image whose gradient difference is greater than a second angle threshold; Extract the two-dimensional coordinates of each curve segment, and take the point with the smallest sequence number in each segment as the entry point, and the point with the largest sequence number in each segment as the exit point.

2. The method for controlling a ground mobile robot on a curve according to claim 1, wherein: Output the corresponding corner entry and exit points including: Establishing a two-dimensional coordinate system to obtain the coordinates of each path point; For all the extracted path point pairs whose distance along the path is less than the distance threshold, the coordinates of the starting path points in all the path point pairs are extracted and averaged to obtain the turning point, and the coordinates of the ending path points in all the path point pairs are extracted and averaged to obtain the turning point.

3. The curve control method of a ground mobile robot according to claim 1, characterized in that: Output the corresponding corner entry and exit points including: Establishing a two-dimensional coordinate system to obtain the coordinates of each path point; For all the extracted path point pairs whose distance along the path is less than the distance threshold, the starting path point among all the path point pairs is extracted, and the starting path point with the smallest serial number is used as the turning point; the ending path point among all the path point pairs is extracted, and the ending path point with the largest serial number is used as the turning point.

4. The method for controlling a ground mobile robot on a curve according to claim 1, wherein: Before drawing the binary curve graph on the blank image through all the path points, performing: establishing a two-dimensional coordinate system, obtaining the coordinates of each of the path points, and establishing a mapping relationship between the blank image and the two-dimensional coordinate system; After retaining a plurality of curve segments in the curved road image whose gradient difference is greater than the second angle threshold, performing: inverse mapping the plurality of curve segments back to the two-dimensional coordinate system according to the mapping relationship.

5. The method for controlling a ground mobile robot on a curve according to claim 1, wherein: After drawing a binary curve graph on the blank image through all the path points, execute: Smooths non-differentiable path vertices.

6. The method for controlling a ground mobile robot on a curve according to claim 1, wherein: The control method is executed when the robot starts walking. If the robot stops during walking, the method returns to step S1.

7. A ground mobile robot curve control system, characterized by: Includes the following modules: Path planning module, used to obtain the walking path through the path planning algorithm; A key point output module is used to extract curves with small span and large curvature in the walking path and output the entry and exit points of the corresponding curves; a speed control module, configured to control the robot to decelerate when the robot approaches the turning point, control the robot to maintain a low speed when the robot enters the curve, and control the robot to resume speed when the robot reaches the turning point; The control system executes the curve control method of a ground mobile robot according to any one of claims 1 to 6 when running.

Images