Underground garage cleaning robot welt cleaning method and equipment and medium

Abstract

The invention discloses an underground garage cleaning robot welt cleaning method and device and a medium, mainly relates to the technical field of welt cleaning, and is used for solving the problem that the existing scheme is weak in the aspects of path point safety inspection and dynamic adjustment, so that the robot is unstable in performance in a complex environment. Comprising the following steps: for an unsafe point, calculating a translation offset point, and replacing the unsafe point with the translation offset point; smoothing a preset path line related to the replaced path point to obtain an updated path point; according to the global path, the robot performs cleaning by using a pure tracking algorithm; in the cleaning process, according to the current linear speed of the robot, a look-ahead point related to a pure tracking algorithm is dynamically calculated; when reaching the preset edge area, measuring the actual distance with the wall in real time through a distance measuring sensor, and obtaining a welt error in combination with the expected welt distance; according to the welt error, updating and calculating a look-ahead point; and calculating the angular velocity of the robot by using the updated look-ahead point.

Description

Technical Field

[0001] This application relates to the field of edge cleaning technology, and in particular to an edge cleaning method, equipment and medium for an underground parking garage cleaning robot. Background Technology

[0002] Underground parking garages, as typical semi-structured environments, are characterized by narrow spaces, dense obstacles, and poor lighting conditions, which place higher demands on the path planning and edge cleaning capabilities of cleaning robots.

[0003] Existing cleaning robots face two main challenges when cleaning along edges: first, inaccurate distance control from walls or obstacles easily leads to collisions or missed areas; second, path planning algorithms have poor adaptability in dynamic environments and cannot effectively handle sudden obstacles. Most cleaning robots on the market currently employ a single path planning pattern, such as a "bow-shaped" or "square-shaped" pattern. While this achieves basic coverage, it is significantly insufficient in edge-cleaning scenarios. Patent CN111474946B discloses a robot edge-cleaning method, but this method only focuses on maintaining the distance between the robot and the edge, without addressing the adaptive control of the cleaning mechanism. Furthermore, traditional methods are weak in path point safety checks and dynamic adjustments, resulting in unstable robot performance in complex environments. Summary of the Invention

[0004] This application provides a method, equipment, and medium for cleaning robots in underground parking garages to clean along the edges, in order to solve the problem that existing solutions are weak in terms of path point safety checks and dynamic adjustments, which leads to unstable robot performance in complex environments.

[0005] In a first aspect, this application provides a method for edge cleaning with a cleaning robot in an underground parking garage, the method comprising: Select the edge cleaning mode; the edge cleaning mode is divided into left edge cleaning and right edge cleaning. Obtain the cleaning area and divide it into a preset open area and a preset edge area; obtain the global path corresponding to the cleaning area; the path of the preset edge area is generated by the teaching path planning algorithm; Obtain the coordinates of obstacles in the underground parking garage, extract the coordinates of all path points from the global path, and mark unsafe points from all path points based on the Euclidean distance between the path point and the coordinates of the nearest obstacle. For unsafe points, calculate translation offset points and replace unsafe points with translation offset points; smooth the preset path lines involved in the replaced path points to obtain updated path points; Based on the global path, the robot uses a pure tracking algorithm for cleaning. During the cleaning process, the robot dynamically calculates the look-ahead points involved in the pure tracking algorithm based on its current linear velocity. When it reaches the preset edge area, it measures the actual distance to the wall in real time using a ranging sensor, and obtains the edge-fitting error by combining the expected edge-fitting distance. Based on the edge-fitting error, the look-ahead points are updated and calculated. The robot's angular velocity is then calculated using the updated look-ahead points.

[0006] In one implementation of this application, determining the edge-to-edge cleaning mode specifically includes: Trigger and confirm the edge-to-edge cleaning mode through the application software interface.

[0007] In one implementation of this application, the cleaning area is obtained and divided into a preset open area and a preset edge area; the global path corresponding to the cleaning area is obtained, specifically including: The cleaning area is divided into a preset open area and a preset edge area by obtaining the division data of the cleaning area through the preset interface. The path is divided into "bow" shapes in the pre-defined open area, and the path is generated in the pre-defined edge area using a teaching path planning algorithm; Combine the paths of the preset open area and the preset edge area into a global path.

[0008] In one implementation of this application, unsafe points are marked from all path points based on the Euclidean distance between a path point and the coordinates of the nearest obstacle, specifically including: Path points whose Euclidean distance is less than a preset distance threshold are identified as unsafe points.

[0009] In one implementation of this application, for an unsafe point, a translation offset point is calculated, and the unsafe point is replaced using the translation offset point; the preset path line involved in the replaced path point is smoothed to obtain the updated path point, specifically including: S1. Obtain the adjacent path points i-1 and i+1 of the unsafe point i; Through the formula: ; ; Calculate the coordinate difference between adjacent path points i-1 and i+1; Through the formula: Calculate the initial translation offset point ; Where (x,y) represents the current unsafe point, When the left edge is attached, the default positive data is used; when the right edge is attached, the default negative data is used. Where (x,y) represents the current unsafe point, Indicates the offset distance; S2. Calculate the Euclidean distance between the initial translation offset point and the coordinates of the nearest obstacle. If it is still an unsafe point, use the initial translation offset point as the updated unsafe point and recalculate the initial translation offset point until the initial translation offset point is no longer an unsafe point. Use the coordinates of points that are not unsafe as the translation offset points; The Savitzky-Golay filtering algorithm is used to smooth the preset path lines involved in the replaced path points to obtain the updated path points.

[0010] In one implementation of this application, during the cleaning process, the look-ahead points involved in the pure tracking algorithm are dynamically calculated based on the robot's current linear velocity, specifically including: Through the formula: Calculate and obtain the look-ahead point ; in, This indicates the preset minimum look-ahead distance. This indicates the preset adjustment coefficient. This indicates the robot's current linear velocity; based on Select a lookahead point from the global path. .

[0011] In one implementation of this application, when a preset edge area is reached, the actual distance to the wall is measured in real time using a distance measuring sensor, and the edge-fitting error is obtained by combining the expected edge-fitting distance; based on the edge-fitting error, the look-ahead point is updated and calculated, specifically including: Through the formula: The edge-fitting error is calculated. ;in, Indicates the actual distance. Indicates the desired edge distance; Determine based on the range of edge-fitting error. , , The corresponding specific values; Through the formula: Calculate the lateral compensation amount ; Through the formula: ; ; Calculate and obtain the look-ahead point Update Preview ; Where δ represents the original lookahead point The tangent direction angle of the path.

[0012] In one implementation of this application, the robot's angular velocity is calculated using the updated look-ahead point, specifically including: Get updated preview points ; Through the formula: Calculate path curvature ; in, It is the robot's current position in the global coordinate system. This is the current forward-looking distance; Through the formula: Calculate the angular velocity; in, This is the robot's current linear velocity.

[0013] Secondly, this application provides an edge-sweeping device for an underground parking garage cleaning robot, the device comprising: processor; And a memory containing executable code, which, when executed, causes the processor to perform an edge-sweeping method for an underground parking garage cleaning robot as described above.

[0014] Thirdly, this application provides a non-volatile computer storage medium storing computer instructions thereon, which, when executed, implement a method for edge cleaning of an underground parking garage cleaning robot as described above.

[0015] As can be seen from the above technical solutions, this application has the following advantages: The edge-following cleaning method for underground parking garage cleaning robots provided in this application improves the robot's cleaning stability and edge-following accuracy in complex environments through multi-dimensional technological optimization. Firstly, it employs a dynamic path point safety detection and adjustment mechanism. By acquiring the coordinates of obstacles in the underground parking garage and marking unsafe points, it calculates translational offset points to replace the original path points and smooths the replaced path points, effectively solving the problem of poor adaptability of traditional path planning algorithms in dynamic environments. This mechanism enables the robot to avoid sudden obstacles in real time, preventing collisions or missed areas due to insufficient path point safety, and is particularly suitable for complex scenarios such as frequent vehicle movement and debris accumulation in underground parking garages. Secondly, combining a pure tracking algorithm with dynamic look-ahead point calculation, the robot adjusts its look-ahead point in real time based on the current linear velocity. During edge-following cleaning, it measures the actual distance using a distance sensor, calculates the error based on the expected edge-following distance, and updates the look-ahead point, achieving precise distance control over walls or obstacles. This design overcomes the shortcomings of traditional single path planning modes (such as "bow-shaped") in edge-following scenarios, reducing collisions or missed areas caused by inaccurate distance control, and improving cleaning coverage and efficiency. Thirdly, by combining the preset edge area with the teaching path planning algorithm, the cleaning area is divided into open area and edge area. The teaching path planning is used to generate the path in the edge area, which further optimizes the path generation logic of edge cleaning, enabling the robot to complete the cleaning task of the edge area more efficiently, and improving the overall stability and reliability of underground garage cleaning. Attached Figure Description

[0016] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of a method for cleaning an underground parking garage using a cleaning robot that cleans along the edges, as provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of the internal structure of an edge-sweeping device for an underground parking garage cleaning robot provided in an embodiment of this application. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Those skilled in the art should understand that the embodiments described below are merely preferred embodiments of this disclosure and do not imply that this disclosure can only be implemented through these preferred embodiments. These preferred embodiments are merely used to explain the technical principles of this disclosure and are not intended to limit the scope of protection of this disclosure. Based on the preferred embodiments provided by this disclosure, all other embodiments obtained by those skilled in the art without creative effort should still fall within the scope of protection of this disclosure.

[0021] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0022] The technical solutions proposed in the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0023] The embodiment provides a method for edge cleaning with a cleaning robot in an underground parking garage, such as... Figure 1 As shown in the embodiments of this application, the method mainly includes the following steps: Step 110: Determine the edge cleaning mode.

[0024] The edge cleaning mode is divided into left edge cleaning and right edge cleaning.

[0025] In some embodiments, determining the edge-to-edge cleaning mode specifically includes: Trigger and confirm the edge-to-edge cleaning mode through the application software interface.

[0026] It should be noted that the cleaning robot is equipped with distance sensors on both sides (such as lidar, ultrasonic sensors or infrared sensors) to detect the distance to walls or obstacles in real time.

[0027] Sensor data is sent to the cleaning robot's control module for processing. The control module determines the edge-walking strategy based on the edge-walking mode flag: when the flag is 1, left edge-walking is performed; when the flag is 2, right edge-walking is performed. Sensor detection data is used to calculate the actual distance between the robot and the wall and compare it with a preset safety distance (usually set to 10cm) to generate an edge-walking error signal.

[0028] It should be noted that this application achieves precise control by clearly defining the edge-walking mode (left edge / right edge) and combining it with sensor data, directly improving the edge-walking effect of the cleaning robot. Specifically, the left and right distance sensors (such as lidar, ultrasonic sensors, or infrared sensors) on the cleaning robot detect the distance to walls or obstacles in real time. After the sensor data is sent to the control module, the edge-walking strategy is determined according to the edge-walking mode flag (edge_walking_flag): left edge-walking is executed when the flag is 1, and right edge-walking is executed when the flag is 2. By calculating the actual distance between the robot and the wall and comparing it with a preset safety distance (usually set to 10cm), an edge-walking error signal is generated. This error signal is directly used to adjust the robot's movement trajectory to ensure that a constant distance is maintained from the wall during the cleaning process. This design avoids the insufficient adaptability of traditional single modes in complex environments, reduces collisions or missed areas caused by inaccurate distance control, and is especially suitable for cleaning edge areas such as underground garages. At the same time, triggering the edge-walking mode through the application software interface further simplifies the operation process and improves the convenience of user interaction.

[0029] Step 120: Obtain the cleaning area, divide the cleaning area into a preset open area and a preset edge area; obtain the global path corresponding to the cleaning area.

[0030] The path to the preset edge region is generated by the teaching path planning algorithm.

[0031] It should be noted that a "bow-shaped" path is used in the preset open area to improve cleaning efficiency. The area to be cleaned is drawn in the operating software, and the program will automatically generate a path to cover the cleaning area according to the width of the vehicle. A teaching path is used in the preset edge area to ensure the edge-fitting effect. The path can also be modified by operating the software. In this way, the cleaning blind spots of the single path mode can be eliminated through the path alternation combination strategy.

[0032] This includes obtaining the cleaning area, dividing it into a preset open area and a preset edge area; and obtaining the global path corresponding to the cleaning area, specifically including: The cleaning area is divided into a preset open area and a preset edge area by obtaining the division data of the cleaning area through the preset interface. The path is divided into "bow" shapes in the pre-defined open area, and the path is generated in the pre-defined edge area using a teaching path planning algorithm; Combine the paths of the preset open area and the preset edge area into a global path.

[0033] It's worth noting that the zoned path planning strategy directly improves the cleaning coverage and edge-fitting accuracy of the cleaning robot. Specifically, the cleaning area is divided into a preset open area and a preset edge area. In the open area, a "bow-shaped" path is used, automatically generated based on the robot's width to cover the cleaning area, achieving efficient cleaning through high-frequency reciprocating motion. In the edge area, a teach-path planning algorithm is used to generate a path, which can be modified in conjunction with the operating software to achieve precise alignment with walls. By combining these two paths into a global path, the cleaning blind spots created by a single "bow-shaped" path in edge areas are eliminated, solving the problem of insufficient adaptability of traditional single-path modes in complex environments. Simultaneously, the introduction of the teach-path allows the robot to adjust its edge-fitting distance according to the actual environment, avoiding collisions or missed areas due to inaccurate distance control, making it particularly suitable for cleaning edge areas such as underground parking garages.

[0034] The cleaning efficiency was further optimized through an alternating path combination strategy. In open areas, the automatic generation of a "bow-shaped" path reduces manual intervention and increases cleaning speed; in edge areas, the modifiability of the taught path allows users to adjust path points according to actual needs, such as shortening path spacing in areas with dense obstacles to enhance cleaning effectiveness. The combination of the two paths not only achieves efficient coverage in open areas but also ensures edge-fitting accuracy in edge areas, improving the overall comprehensiveness and reliability of cleaning.

[0035] Step 130: Obtain the coordinates of obstacles in the underground parking garage and extract the coordinates of all path points from the global path; mark unsafe points from all path points according to the Euclidean distance between the path point and the coordinates of the nearest obstacle; for unsafe points, calculate the translation offset points and replace the unsafe points with the translation offset points; smooth the preset path lines involved in the replaced path points to obtain the updated path points.

[0036] It should be noted that for unsafe points, this step involves translating and offsetting along the normal direction of the path tangent. The offset algorithm first calculates the tangent direction vector at the path point, then selects the left or right normal direction according to the edge-fitting mode, and gradually translates the path point in a fixed step size (e.g., 0.05 meters) until the safe distance requirement is met or the maximum offset range (e.g., 0.5 meters) is reached.

[0037] Specifically, unsafe points are marked from all path points based on the Euclidean distance between the path point and the coordinates of the nearest obstacle, including: Path points whose Euclidean distance is less than a preset distance threshold are identified as unsafe points.

[0038] For unsafe points, calculate translation offset points and replace the unsafe points with these offset points; smooth the preset path lines involved in the replaced path points to obtain updated path points, specifically including: S1. Obtain the adjacent path points i-1 and i+1 of the unsafe point i; Through the formula: ; ; Calculate the coordinate difference between adjacent path points i-1 and i+1; Through the formula: Calculate the initial translation offset point ; Where (x,y) represents the current unsafe point, When the left edge is attached, the default positive data is used; when the right edge is attached, the default negative data is used. S2. Calculate the Euclidean distance between the initial translation offset point and the coordinates of the nearest obstacle. If it is still an unsafe point, use the initial translation offset point as the updated unsafe point and recalculate the initial translation offset point until the initial translation offset point is no longer an unsafe point. Use the coordinates of points that are not unsafe as the translation offset points; The Savitzky-Golay filtering algorithm is used to smooth the preset path lines involved in the replaced path points to obtain the updated path points.

[0039] It's worth noting that dynamic path point safety detection and offset processing directly improves the cleaning robot's obstacle avoidance capabilities and path stability in complex environments. Specifically, by acquiring the coordinates of obstacles in the underground parking garage, the coordinates of all path points are extracted from the global path, and the Euclidean distance between each path point and the nearest obstacle is calculated. Path points with distances less than a preset threshold are marked as unsafe points. For unsafe points, a translation offset is made along the normal direction of the path tangent. The path point is gradually translated with a fixed step size until the safe distance requirement is met or the maximum offset range is reached. This design enables the robot to avoid sudden obstacles in real time, preventing collisions or missed areas due to insufficient path point safety, and is particularly suitable for complex scenarios such as frequent vehicle movement and debris accumulation in underground parking garages. Simultaneously, the Savitzky-Golay filtering algorithm smooths the translation offset points, ensuring the continuity and executability of the updated path points, reducing bumps and stuttering during robot movement, and improving cleaning efficiency.

[0040] Path safety was further optimized through iterative calculations and smoothing. Specifically, after calculating the initial translation offset point, the Euclidean distance between it and the nearest obstacle's coordinates was determined. If it was still an unsafe point, the initial translation offset point was recalculated until the safety requirements were met. This iterative process ensured the reliability of the translation offset point and avoided residual risks caused by insufficient offset in a single operation. Simultaneously, the application of the Savitzky-Golay filtering algorithm effectively eliminated noise introduced by the translation offset point, making the updated path points smoother and conforming to robot kinematic constraints. By replacing the original unsafe point with the translation offset point and combining it with the global path for subsequent cleaning, real-time obstacle avoidance and path adjustment in dynamic environments were achieved, improving the robot's adaptability and cleaning performance in complex environments.

[0041] Step 140: Based on the global path, the robot uses a pure tracking algorithm to perform cleaning. During the cleaning process, the robot dynamically calculates the look-ahead points involved in the pure tracking algorithm based on its current linear velocity. When it reaches the preset edge area, the robot measures the actual distance to the wall in real time using a ranging sensor, and obtains the edge-fitting error by combining the expected edge-fitting distance. The look-ahead points are updated based on the edge-fitting error. The robot's angular velocity is calculated using the updated look-ahead points.

[0042] In some embodiments, during the cleaning process, the look-ahead points involved in the pure tracking algorithm are dynamically calculated based on the robot's current linear velocity, specifically including: Through the formula: Calculate and obtain the look-ahead point ; in, This indicates the preset minimum look-ahead distance. This indicates the preset adjustment coefficient. This indicates the robot's current linear velocity; based on Select a lookahead point from the global path. .

[0043] It should be noted that, to address the issue of decreased control accuracy in edge-following cleaning caused by environmental disturbances or inaccurate robot dynamics models in traditional PID control, this invention employs an adaptive PID controller to handle this error. This controller can adjust its performance based on the system's operating status (such as error rate). It adjusts its proportional, integral, and derivative gains online (and their rate of change) to maintain good control performance in uncertain or changing environments (such as changes in ground friction coefficient or changes in robot load leading to altered dynamic characteristics). Its output is the lateral compensation quantity. .

[0044] Upon reaching the preset edge area, the actual distance to the wall is measured in real time using a distance sensor, and the edge-fitting error is obtained by combining this with the expected edge-fitting distance. Based on the edge-fitting error, the look-ahead point is updated and calculated, specifically including: Through the formula: The edge-fitting error is calculated. ;in, Indicates the actual distance. Indicates the desired edge distance; Determine based on the range of edge-fitting error. , , The corresponding specific values; Through the formula: Calculate the lateral compensation amount ; Through the formula: ; ; Calculate and obtain the look-ahead point Update Preview ; Where δ represents the original lookahead point The tangent direction angle of the path.

[0045] Using the updated look-ahead point, the robot's angular velocity is calculated, specifically including: Get updated preview points ; Through the formula: Calculate path curvature ; in, It is the robot's current position in the global coordinate system. This is the current forward-looking distance; Through the formula: Calculate the angular velocity; in, This is the robot's current linear velocity.

[0046] It's worth noting that by dynamically adjusting the look-ahead point and using adaptive PID control, the robot's trajectory tracking accuracy and stability during edge-following cleaning are improved. Specifically, during cleaning, the look-ahead point is dynamically calculated based on the robot's current linear velocity. By presetting a minimum look-ahead distance and an adjustment coefficient in relation to the current linear velocity, the look-ahead point is ensured to always match the robot's motion state. When the robot reaches a preset edge area, the actual distance to the wall is measured in real time using a distance sensor. This distance is combined with the desired edge-following distance to obtain the edge-following error, and the look-ahead point is updated based on the error range. This design allows the robot to perceive environmental changes in real time and dynamically adjust the look-ahead point position, avoiding trajectory deviation or collision risks caused by a fixed look-ahead point. Simultaneously, the adaptive PID controller adjusts the proportional, integral, and derivative gains online based on the edge-following error and its rate of change, effectively solving the problem of decreased control accuracy caused by environmental disturbances or inaccurate robot dynamics models in traditional PID control. This ensures stable edge-following cleaning performance, especially under complex ground conditions or load variations.

[0047] The robot's motion control performance was further optimized through look-ahead point updates and angular velocity calculations. Specifically, lateral compensation was calculated based on the edge-fitting error, and the look-ahead point was updated using the tangent angle of the path at the original look-ahead point, enabling the robot to more accurately fit against the wall edge. The robot's angular velocity was calculated using the updated look-ahead point, and the correlation between path curvature and current linear velocity was used to ensure that the angular velocity matched the robot's motion state. This design allows the robot to smoothly adjust its direction during cleaning, reducing bumps or deviations caused by sudden changes in angular velocity, making it particularly suitable for edge-fitting cleaning in narrow spaces such as underground parking garages. Simultaneously, the lateral compensation output of the adaptive PID controller further enhances the robot's anti-interference capability, maintaining a stable edge-fitting distance even when encountering changes in the ground friction coefficient or dynamic characteristics, thus improving cleaning efficiency and safety.

[0048] As described above, this embodiment improves the robot's cleaning stability and edge-following accuracy in complex environments through multi-dimensional technological optimization. Firstly, it employs a dynamic path point safety detection and adjustment mechanism. By acquiring the coordinates of obstacles in the underground parking garage and marking unsafe points, it calculates translational offset points to replace the original path points and smooths the replaced path points, effectively solving the problem of poor adaptability of traditional path planning algorithms in dynamic environments. This mechanism enables the robot to avoid sudden obstacles in real time, preventing collisions or missed areas due to insufficient path point safety, and is particularly suitable for complex scenarios such as frequent vehicle movement and debris accumulation in underground parking garages. Secondly, combining a pure tracking algorithm with dynamic look-ahead point calculation, the robot adjusts its look-ahead point in real time based on its current linear velocity. During edge-following cleaning, it measures the actual distance using a distance sensor, calculates the error based on the expected edge-following distance, and updates the look-ahead point, achieving precise distance control over walls or obstacles. This design overcomes the shortcomings of traditional single path planning modes (such as the "bow-shaped" approach) in edge-following scenarios, reducing collisions or missed areas caused by inaccurate distance control and improving cleaning coverage and efficiency. Thirdly, by combining the preset edge area with the teaching path planning algorithm, the cleaning area is divided into open area and edge area. The teaching path planning is used to generate the path in the edge area, which further optimizes the path generation logic of edge cleaning, enabling the robot to complete the cleaning task of the edge area more efficiently, and improving the overall stability and reliability of underground garage cleaning.

[0049] The above are method embodiments of this application. Based on the same inventive concept, this application also provides an edge-sweeping device for an underground parking garage cleaning robot. Figure 2 As shown, the device includes: a processor; and a memory storing executable code thereon, which, when executed, causes the processor to perform an edge-sweeping method for an underground parking garage cleaning robot as described in the above embodiment.

[0050] Specifically, the server determines the edge-following cleaning mode; this mode is further divided into left-following and right-following; it acquires the cleaning area, dividing it into a preset open area and a preset edge area; it acquires the global path corresponding to the cleaning area; the path for the preset edge area is generated by a teaching path planning algorithm; it acquires the coordinates of obstacles in the underground parking garage, extracting the coordinates of all path points from the global path; based on the Euclidean distance between the path point and the nearest obstacle coordinates, it marks unsafe points from all path points; for unsafe points, it calculates translation offset points and replaces unsafe points using these offset points; it smooths the preset path lines involved in the replaced path points to obtain updated path points; based on the global path, the robot uses a pure tracking algorithm for cleaning; during the cleaning process, it dynamically calculates the look-ahead points involved in the pure tracking algorithm based on the robot's current linear velocity; when reaching the preset edge area, it measures the actual distance to the wall in real time using a ranging sensor, and obtains the edge-following error based on the expected edge-following distance; it updates the look-ahead points based on the edge-following error; and it calculates the robot's angular velocity using the updated look-ahead points.

[0051] In addition, this application embodiment also provides a non-volatile computer storage medium storing executable instructions, which, when executed, implement the above-described method for edge cleaning by an underground parking garage cleaning robot.

[0052] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An under-ground garage cleaning robot skirting cleaning method characterized by, The method comprises: determining a skirting cleaning mode; wherein the skirting cleaning mode is divided into left skirting and right skirting; obtaining a cleaning area, dividing the cleaning area into a preset open area and a preset edge area, obtaining a global path corresponding to the cleaning area; wherein the path of the preset edge area is generated by a teaching path planning algorithm; obtaining an underground garage obstacle coordinate, extracting all path point coordinates from the global path, marking unsafe points from all path points according to the Euclidean distance between the path points and the nearest obstacle coordinate; for the unsafe points, calculating a translation offset point, and replacing the unsafe points with the translation offset point; performing smoothing processing on the preset path line involved in the replaced path point to obtain updated path points; according to the global path, the robot performs cleaning by using a pure tracking algorithm; in the cleaning process, the look-ahead point involved in the pure tracking algorithm is dynamically calculated according to the current linear speed of the robot; when reaching the preset edge area, the actual distance from the wall is measured in real time by using a distance measuring sensor, and the skirting error is obtained by combining the expected skirting distance; the look-ahead point is updated according to the skirting error; the angular velocity of the robot is calculated by using the updated look-ahead point.

2. The underground garage cleaning robot skirting cleaning method of claim 1, wherein, determining a skirting cleaning mode, specifically comprising: triggering the determination of the skirting cleaning mode through an application software interface.

3. The underground garage cleaning robot skirting cleaning method of claim 1, wherein, obtaining a cleaning area, dividing the cleaning area into a preset open area and a preset edge area, and obtaining a global path corresponding to the cleaning area, specifically comprising: obtaining the division data of the cleaning area through a preset interface, and then dividing the cleaning area into a preset open area and a preset edge area; determining that the preset open area adopts a "bow-shaped" division path, and that the preset edge area generates a path by using a teaching path planning algorithm; combining the path of the preset open area and the path of the preset edge area into a global path.

4. The underground garage cleaning robot skirting cleaning method of claim 1, wherein, marking unsafe points from all path points according to the Euclidean distance between the path points and the nearest obstacle coordinate, specifically comprising: determining that the path point with a Euclidean distance less than a preset distance threshold is an unsafe point.

5. The underground garage cleaning robot skirting cleaning method of claim 1, wherein, for the unsafe points, calculating a translation offset point, and replacing the unsafe points with the translation offset point; performing smoothing processing on the preset path line involved in the replaced path point to obtain updated path points, specifically comprising: S1, obtaining adjacent path points i-1 and i+1 of the unsafe point i; The coordinate difference between adjacent path points i-1 and i+1 is calculated by the formula: ; ; through the formula: , calculating an initial translation offset point ; wherein (x, y) represents the current unsafe point, when being left-justified, the preset positive data, and when being right-justified, the preset negative data. S2, calculating the Euclidean distance between the initial translation offset point and the nearest obstacle coordinate, when it is still an unsafe point, taking the initial translation offset point as the updated unsafe point, and calculating the initial translation offset point again until the initial translation offset point is not an unsafe point; taking the coordinate that is not an unsafe point as the translation offset point; using the Savitzky-Golay filtering algorithm to perform smoothing processing on the preset path line involved in the replaced path point to obtain updated path points.

6. The underground garage cleaning robot skirting cleaning method of claim 1, wherein, in the cleaning process, the look-ahead point involved in the pure tracking algorithm is dynamically calculated according to the current linear speed of the robot, specifically comprising: through the formula: , the prospective point is calculated ; wherein, represents a preset minimum look-ahead distance, represents a preset adjustment coefficient, represents a current linear speed of the robot; based on selecting a lookahead point from a global path .

7. The underground garage cleaning robot skirting cleaning method of claim 1, wherein, when reaching the preset edge area, the actual distance from the wall is measured in real time by using a distance measuring sensor, and the skirting error is obtained by combining the expected skirting distance; the look-ahead point is updated according to the skirting error, specifically comprising: through the formula: , the calculated tuck error ; wherein, represents the actual distance, represents the desired tuck distance; According to the falling range of the edge error, determine , , corresponding specific values; through the formula: , the lateral compensation amount is calculated ; through the formula: ； ； computing an update lookahead point of the lookahead point ; wherein δ denotes the original look-ahead point of the tangent direction angle of the path at the point.

8. The underground garage cleaning robot skirting cleaning method of claim 1, wherein, Using the updated look-ahead point, the angular velocity of the robot is calculated, specifically comprising: acquiring the updated lookahead point ; By the equation: , calculating path curvature ; wherein, is the current position of the robot in the global coordinate system, is the current look-ahead distance; By the equation: , calculating an angular velocity; wherein, is the current linear velocity of the robot.

9. An under-ground garage cleaning robot skirting cleaning apparatus characterized by, The device comprises: a processor; and a memory having stored thereon executable code that, when executed, causes the processor to perform a method of edge cleaning for an underground garage cleaning robot as claimed in any of claims 1-8.

10. A non-transitory computer storage medium, comprising, having stored thereon computer instructions that, when executed, implement a method of edge cleaning for an underground garage cleaning robot as claimed in any of claims 1-8.

Images