Method, system and pool cleaning robot for pool cleaning robot detection and disengagement from stepped platforms

Abstract

This application relates to the field of underwater robot technology and discloses a method, system, and pool cleaning robot for detecting and escaping from stepped platforms. By collecting multimodal environmental perception data, the robot determines the terrain type it is currently in, including normal driving, ordinary pool wall climbing, oscillation in wide stepped areas, and getting stuck in narrow stepped areas. Based on the terrain type, the robot adaptively selects or generates a corresponding stepped escape action sequence from a preset dynamic stepped escape library and executes it. This allows for the execution of corresponding escape commands for different terrain types, reducing human intervention, increasing the success rate of escape, and improving the efficiency of automatic cleaning.

Description

Technical Field

[0001] This application relates to the field of underwater robot technology, and in particular to a method, system, and pool cleaning robot for detecting and detaching from a stepped platform. Background Technology

[0002] Pool cleaning robots are intelligent devices used for automated cleaning of swimming pools. During the cleaning process, robots often encounter stepped areas within the pool. These steps, depending on their width, slope, and other characteristics, can create different types of obstacles for the robot's movement. For example, wide steps (usually wider than 60 cm) can cause the robot to repeatedly fail to climb, leading to abnormal posture over time, while narrow steps (usually between 20 and 60 cm wide) can cause the robot to become stuck in continuous oscillations. Traditional pool cleaning robots are typically equipped only with contact sensors, such as collision switches, or proximity sensors, such as infrared sensors, which perform simple backward-turning actions when encountering obstacles. However, this approach has limited effectiveness on the special terrain of stepped platforms. Therefore, the slope of the stepped platform may not effectively trigger the collision sensors, and it cannot distinguish between steps and ordinary pool walls, resulting in an ineffective cycle of collision, backward movement, and further collisions in the stepped area. The success rate of escaping from these obstacles is low, and manual intervention is usually required. Summary of the Invention

[0003] This application provides a method for detecting and detaching a pool cleaning robot from a stepped platform, aiming to solve the aforementioned technical problems.

[0004] This application provides a method for detecting and detaching from a stepped platform for a pool cleaning robot, including: Collect multimodal environmental perception data, including tactile pressure distribution data, three-dimensional attitude angle data, and chassis height data; The terrain type currently in which the robot is located is determined based on the multimodal environmental perception data; Based on the terrain type, the system adaptively selects or generates a corresponding stair escape sequence from a preset dynamic stair escape library and executes it to leave the current terrain. The dynamic stair escape library includes various escape instructions with different action logics.

[0005] In one embodiment, determining the terrain type where the robot is currently located based on the multimodal environmental perception data includes: The tactile pressure distribution data is analyzed to obtain the spatial pressure distribution state and pressure value change time sequence of the pressure sensor; The collision pattern between the robot and the obstacle is identified based on the spatial pressure distribution state and the pressure value change time sequence, wherein the collision pattern includes at least one of frontal collision, side scraping and bottom pressing; The sum of the absolute values ​​of the changes in pitch angle, roll angle and yaw angle within a preset time window is calculated based on the three-dimensional attitude angle data, and this sum is used as the attitude stability value. The vertical distance from the lower surface of the robot chassis to the bottom of the swimming pool at the current location is obtained based on the chassis height data; The robot's travel state is determined based on the vertical distance, wherein the travel state includes driving up the stairs, hovering at the edge of the stairs, or repeatedly oscillating. The collision mode, the attitude stability value, and the travel state are used as input feature vectors and input to a preset decision model. The decision model outputs a determination result of the robot's current terrain category based on the input feature vectors. The determination result includes normal travel, ordinary pool wall climbing, oscillation in wide-step areas, and getting stuck in narrow-step areas.

[0006] In one embodiment, the decision model is a state decision model based on predefined rules or a trained classification machine learning model.

[0007] In one embodiment, identifying the collision pattern between the robot and the obstacle based on the spatial pressure distribution state and the timing of pressure value changes includes: Based on the spatial pressure distribution, pressure concentration areas are identified, including pressure concentration in the central area of ​​the robot's front bumper, pressure concentration on the side of the robot, and pressure concentration in the lower front or middle part of the robot's chassis. The pressure change state is obtained based on the pressure value change time sequence, wherein the pressure change state includes a sharp increase in pressure value and a continuous change in pressure value; If the pressure is concentrated in the central area of ​​the robot's front bumper and the pressure value increases sharply, the collision pattern between the robot and the obstacle is identified as a frontal collision. If the pressure is concentrated on the side of the robot and the pressure value changes continuously, the collision pattern between the robot and the obstacle is identified as side scraping. If the pressure is concentrated at the front or lower middle of the robot chassis, the collision pattern between the robot and the obstacle is identified as bottom pressure.

[0008] In one embodiment, the step of calculating the sum of the absolute values ​​of the changes in pitch angle, roll angle, and yaw angle within a preset time window based on the three-dimensional attitude angle data, and using this sum as the attitude stability value, includes: The three-dimensional attitude angle data is acquired, wherein the three-dimensional attitude angle data includes pitch angle, roll angle and yaw angle; Calculate the absolute values ​​of the angle differences between the current time and the preset time window, namely the pitch angle, roll angle, and yaw angle; The attitude stability value is obtained by adding the absolute values ​​of the angle differences between the three values.

[0009] In one embodiment, determining the robot's travel state based on the vertical distance includes: Determine whether the vertical distance exceeds a first threshold and increases dramatically within a first preset time period, and whether it remains stable within a second preset time period. If it exceeds the first threshold and increases dramatically within a first preset time period, and remains stable within a second preset time period, then determine that the travel state is driving up the stairs. Determine whether the vertical distance exhibits periodic changes within the third time period. If it exhibits periodic changes, and the period of change is synchronized with or has a fixed phase relationship with the periodic changes in the pitch angle, then the travel state is determined to be repetitive oscillation. If there is no periodic change, determine whether the vertical distance exceeds the second threshold and shows vertical distance value fluctuation within the fourth time period. If it exceeds the second threshold and shows vertical distance value fluctuation, determine that the travel state is suspended at the edge of the steps.

[0010] In one embodiment, the step of adaptively selecting or generating a corresponding stair-step escape sequence from a preset dynamic stair-step escape database based on the terrain type includes: Obtain the terrain category, which includes at least wide-step area oscillation and narrow-step area jamming; When the terrain is classified as a wide stepped area oscillation, the corresponding stepped dragging action sequence is to retreat for the first preset duration and then turn at the first angle. When stuck in a narrow stepped terrain, the corresponding sequence of steps to get out of trouble is to adjust the aircraft attitude, then retreat for a second preset duration and turn at a second angle.

[0011] This application also provides a system for detecting and disengaging a pool cleaning robot from a stepped platform, the system being configured to implement the method described in any of the preceding claims, comprising: The data acquisition module is used to collect multimodal environmental perception data, including tactile pressure distribution data, three-dimensional attitude angle data, and chassis height data. The determination module is used to determine the terrain type where the robot is currently located based on the multimodal environmental perception data; The execution module is used to adaptively select or generate a corresponding stair escape action sequence from a preset dynamic stair escape library according to the terrain type and execute it to leave the current terrain. The dynamic stair escape library includes a variety of escape instructions with different action logics.

[0012] This application also provides a pool cleaning robot, including the system described above for detecting and disengaging from a stepped platform.

[0013] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0014] The beneficial effects of this application are as follows: by collecting multimodal environmental perception data, the robot can determine the current terrain type, which includes normal driving, ordinary pool wall climbing, oscillation in wide-step areas, and getting stuck in narrow-step areas. Based on the terrain type, the robot can adaptively select or generate the corresponding step-by-step escape action sequence from the preset dynamic step escape library and execute it. This allows for the execution of corresponding escape commands for different terrain types, reducing manual intervention, increasing the success rate of escape, and improving the efficiency of its automatic cleaning. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of a method flow according to an embodiment of this application; Figure 2 This is a flowchart illustrating a method according to another embodiment; Figure 3 This is a flowchart illustrating a method according to another embodiment; Figure 4 This is a schematic diagram of the system structure according to an embodiment of this application.

[0016] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0018] In one embodiment, such as Figure 1-4 As shown, this application provides a method for a swimming pool cleaning robot to detect and detach from a stepped platform. This embodiment illustrates the method by applying it to a computer device. It is understood that this method can also be applied to a server, and further to a system including both a computer device and a server, and implemented through the interaction between the computer device and the server. In this embodiment, the method includes the following steps: S1. Collect multimodal environmental perception data, wherein the multimodal environmental perception data includes tactile pressure distribution data, three-dimensional attitude angle data, and chassis height data.

[0019] For example, in this embodiment, the robot's bumper is equipped with a pressure sensor that can capture tactile pressure distribution data in real time, and an inertial measurement unit (IMU) that can collect the robot's three-dimensional attitude angle data, specifically including pitch angle, roll angle, yaw angle and angular velocity; the bottom of the robot is equipped with an ultrasonic ranging sensor that can be used to continuously measure the chassis height data of the center of the chassis from the bottom of the pool.

[0020] S2. Determine the terrain type where the robot is currently located based on the multimodal environmental perception data.

[0021] In this embodiment, as an optional implementation, the main controller on the robot is equipped with a lightweight decision engine. This decision engine can process multimodal environmental perception data to determine the terrain category in which the robot is currently located. The terrain categories include normal driving, ordinary pool wall climbing, oscillation in wide-step areas, and getting stuck in narrow-step areas. For example, the decision engine can make a determination based on a set of pre-set rules, or it can output the determination result of the current terrain category based on the reasoning result of a small neural network. This is not the only limitation.

[0022] S3. Based on the terrain type, adaptively select or generate the corresponding stair escape action sequence from the preset dynamic stair escape library and execute it to leave the current terrain. The dynamic stair escape library includes a variety of escape instructions with different action logics.

[0023] In this embodiment, as an optional implementation, a dynamic staircase escape library is first preset. The library contains various escape commands with different action logics. For example, if the terrain type is a wide staircase area oscillation, the corresponding escape command is to control the robot to move backward for 5 seconds, then turn left 45 degrees, and then move forward at low speed. For a narrow staircase area stuck, the left and right tracks can be controlled to alternate forward and reverse at a small differential speed for 3 seconds to make the body sway, and then the standard backward movement and turning are executed. During this process, the data corresponding to the escape is continuously monitored. If no other terrain type is triggered in the next 10 seconds, the escape is determined to be successful. This allows for the execution of corresponding escape commands for different terrain types, reducing human intervention and increasing the success rate of escape. Furthermore, to enable the robot to adaptively escape later, key data from each successful escape, such as timestamps, executed commands, and the terrain location, can be packaged, compressed, and stored in a dynamic stair escape database. This facilitates subsequent adaptive optimization by the robot. For example, after escaping the stair, key data from the stair escape sequence are collected and packaged. This key data includes the timestamp of the escape time, the determination result of the terrain type that triggered the escape, and the selected and executed stair from the dynamic stair escape database. The sequence of escape actions is generated; the key data packaged is stored as experience records and updated in the dynamic stair escape database. For example, the unique identifier and timestamp of the machine can be used as an index and stored in local memory, which can retain the latest multiple experience records; based on the updated dynamic stair escape database, the decision logic for the cleaning robot to subsequently execute the adaptive selection or generate the corresponding stair escape sequence is optimized. For example, if statistics show that for narrow stair jamming, 80% of the experience records show that the first step was a back-turn and right-turn escape action, then the priority order of back-turn and right-turn can be adjusted forward and recommended first.

[0024] In one embodiment, determining the terrain category S2 where the robot is currently located based on the multimodal environmental perception data includes: S21. Analyze the tactile pressure distribution data to obtain the spatial pressure distribution state and pressure value change sequence of the pressure sensor; S22. Identify the collision mode between the robot and the obstacle based on the spatial pressure distribution state and the pressure value change sequence, wherein the collision mode includes at least one of frontal collision, side scraping and bottom pressing. S23. Calculate the sum of the absolute values ​​of the changes in pitch angle, roll angle and yaw angle within a preset time window based on the three-dimensional attitude angle data, and use it as the attitude stability value. S24. Obtain the vertical distance from the lower surface of the robot chassis to the bottom of the swimming pool at the current location based on the chassis height data; S25. Determine the robot's travel state based on the vertical distance, wherein the travel state includes driving up the stairs, hovering at the edge of the stairs, or repeatedly oscillating. S26. The collision mode, the attitude stability value, and the travel state are used as input feature vectors and input to a preset decision model. The decision model outputs a determination result of the terrain category where the robot is currently located based on the input feature vectors. The determination result includes normal travel, ordinary pool wall climbing, oscillation in wide-step areas, and getting stuck in narrow-step areas.

[0025] As described in steps S21-S26 above, this embodiment, by fusing tactile pressure distribution, three-dimensional posture angle data, and chassis height data, reduces the probability of misjudgment compared to a single sensor in traditional technologies. By combining the collision mode, posture stability value under motion, and action state, the three factors are incorporated into a preset decision model, which can comprehensively weigh various features, resulting in faster and more accurate judgment results. Compared to the design logic of a simple threshold chain in manual systems, the decision-making process is more intelligent, thereby ensuring that the robot triggers the most suitable subsequent escape process according to different scenarios, improving the success rate of escape and reducing human intervention.

[0026] In one embodiment, identifying the collision pattern S22 between the robot and the obstacle based on the spatial pressure distribution state and the pressure value change time sequence includes: S221. Identify pressure concentration areas based on spatial pressure distribution, wherein the pressure concentration areas include the central area of ​​the robot's front bumper, the side of the robot, and the lower front or middle part of the robot chassis. S222. Obtain the pressure change state according to the pressure value change time sequence, wherein the pressure change state includes a sharp increase in pressure value and a continuous change in pressure value; S223. If the pressure is concentrated in the central area of ​​the robot's front bumper and the pressure value increases sharply, the collision mode between the robot and the obstacle is identified as a frontal collision. S224. If the pressure is concentrated on the side of the robot and the pressure value changes continuously, the collision mode between the robot and the obstacle is identified as side scraping. S225. If the pressure is concentrated in the front or lower middle part of the robot chassis, the collision mode between the robot and the obstacle is identified as bottom pressure.

[0027] As described in steps S221-S225 above, by analyzing the tactile pressure distribution data, it is possible to identify refined collision patterns such as frontal collision, side scraping, and bottom pressure. For example, bottom pressure is a typical feature of a robot track getting stuck on the edge of a step, while side scraping may be because the robot is moving along the edge of the step. By identifying the specific collision patterns, the specific structure of the step can be inferred more accurately, such as its width and the position of the edges, which makes it easier to select the most suitable escape command based on the collision pattern.

[0028] In one embodiment, step S23, which calculates the sum of the absolute values ​​of the changes in pitch angle, roll angle, and yaw angle within a preset time window based on the three-dimensional attitude angle data and uses this sum as the attitude stability value, includes: S231. Obtain the three-dimensional attitude angle data, wherein the three-dimensional attitude angle data includes pitch angle, roll angle and yaw angle; S232. Calculate the absolute values ​​of the angle differences between the current time and the preset time window time, namely the pitch angle, roll angle and yaw angle; S233. Add the absolute values ​​of the angle differences of the three values ​​to obtain the attitude stability value.

[0029] As described in steps S231-S233 above, during the robot's operation, it can be determined whether it is in a stepped area by using three-dimensional attitude angle data. For example, when the robot is moving normally in the pool, the number of times it encounters the wall during each wall climbing process can be recorded, and the sum of the absolute values ​​of the robot's yaw angle, pitch angle, and roll angle changes can be collected every 1 second. If the attitude stability value does not meet the preset conditions, it indicates that it is in a stepped area.

[0030] In one embodiment, step S24, which determines the robot's travel state based on the vertical distance, includes: S241. Determine whether the vertical distance exceeds the first threshold and increases stepwise within the first preset time period, and whether it remains stable within the second preset time period. If it exceeds the first threshold and increases stepwise within the first preset time period, and remains stable within the second preset time period, then determine that the travel state is driving up the stairs. Among them, a step increase refers to a sudden jump in vertical distance within a short period of time, such as within 3 seconds, from a stable level to a significantly higher level, with almost no gradual transition in between, and a direct leap. After the leap, the vertical distance remains stable. In this case, it can be determined that it is driving up a staircase.

[0031] S242. Determine whether the vertical distance exhibits periodic changes within the third time period. If it exhibits periodic changes, and the period of change is synchronized with or has a fixed phase relationship with the periodic changes in the pitch angle, then determine that the travel state is repetitive oscillation. For example, since vertical distance is easily affected by water surface fluctuations or uneven pool bottoms, and attitude data alone cannot distinguish whether the robot is climbing the pool wall or oscillating on the steps, this embodiment combines vertical distance and three-dimensional attitude angle data. Only when both vertical distance and pitch angle change periodically and synchronously are they determined to be repetitive oscillations, the accuracy of the determination can be improved, providing a basis for subsequent escape.

[0032] S243. If there is no periodic change, determine whether the vertical distance exceeds the second threshold and shows vertical distance value fluctuation in the fourth time period. If it exceeds the second threshold and shows vertical distance value fluctuation, determine that the travel state is suspended at the edge of the steps.

[0033] As described in steps S241-S243 above, this embodiment can finely classify the trapped situation by determining its movement state as driving up stairs, repeated oscillation, and hovering at the edge of stairs. It can identify the situation in the initial stage when the robot gets stuck in a regular loop, rather than responding after it gets stuck. This saves time for implementing preventive, low-energy early escape intervention. Different escape commands can be executed for different movement states, resulting in a high success rate of escape.

[0034] In one embodiment, step S3, which adaptively selects or generates a corresponding stair-step escape sequence from a preset dynamic stair-step escape database based on the terrain type, includes: S31. Obtain the terrain category, which includes at least wide-step area oscillation and narrow-step area jamming; S32. When the terrain type is a wide stepped area oscillation, the corresponding stepped dragging action sequence is to retreat for the first preset time and then turn at the first angle.

[0035] For example, when oscillating in a wide stepped area, the first preset duration can be 3-8 seconds. During this duration, a backward movement is performed to ensure that the robot can completely detach from the inclined surface or edge of the step. After the backward movement ends, the robot can be controlled to perform a stationary or small-radius turning movement. The first angle can be 45 degrees to 90 degrees. This can prevent the robot's axis of travel from forming a large angle with the edge of the step that caused the oscillation when it moves forward again, thus making it more likely to bypass the area.

[0036] S33. When stuck in a narrow stepped area, the corresponding step escape sequence is to adjust the aircraft attitude, then retreat for a second preset duration and turn at a second angle.

[0037] For example, when stuck in a narrow stepped area, before performing a regular backward movement, the robot's left and right tracks or wheels can be controlled to perform short, rapid alternating forward and reverse movements. For example, the left track rotates forward for 0.5 seconds while the right track rotates backward for 0.5 seconds, and then alternates. The duration can be 2-4 seconds. In this way, the robot's posture can be adjusted through active swaying, freeing it from the stuck position. Then, the robot can perform a second preset backward movement for a second preset duration and a second turning angle. The second preset duration can be set slightly shorter than the first duration, and the second angle can be set higher than the first angle, for example, 90-120 degrees, so as to deviate more from the original stuck direction and avoid getting stuck in the same narrow area again.

[0038] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0039] Based on the same inventive concept, this application also provides a system for detecting and detaching a pool cleaning robot from a stepped platform. The solution provided by this system is similar to the solution described in the above method. Therefore, the specific limitations of one or more system embodiments for detecting and detaching a pool cleaning robot from a stepped platform provided below can be found in the limitations of the method for detecting and detaching a pool cleaning robot from a stepped platform described above, and will not be repeated here.

[0040] In one exemplary embodiment, a system is provided for a pool cleaning robot to detect and detach from a stepped platform, comprising: The data acquisition module is used to collect multimodal environmental perception data, including tactile pressure distribution data, three-dimensional attitude angle data, and chassis height data. The determination module is used to determine the terrain type where the robot is currently located based on the multimodal environmental perception data; The execution module is used to adaptively select or generate a corresponding stair escape action sequence from a preset dynamic stair escape library according to the terrain type and execute it to leave the current terrain. The dynamic stair escape library includes a variety of escape instructions with different action logics.

[0041] The system for detecting and detaching from the stepped platform for a pool cleaning robot provided in this embodiment can execute the above-described method embodiment. Its implementation principle and technical effect are similar, and will not be repeated here.

[0042] Based on the same inventive concept, this application also provides a swimming pool cleaning robot. The solution provided by this swimming pool cleaning robot is similar to the solution described in the above method. Therefore, the specific limitations of the one or more embodiments of the swimming pool cleaning robot can be found in the above-described limitations of the method for detecting and detaching from the stepped platform, and will not be repeated here.

[0043] An embodiment of this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for detecting and detaching a pool cleaning robot from a stepped platform.

[0044] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in this application and in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0045] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0046] The above description is only a preferred embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural changes made based on the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for detecting and detaching from a stepped platform for a pool cleaning robot, characterized in that, include: Collect multimodal environmental perception data, including tactile pressure distribution data, three-dimensional attitude angle data, and chassis height data; The terrain type currently in which the robot is located is determined based on the multimodal environmental perception data; Based on the terrain type, the system adaptively selects or generates a corresponding stair escape sequence from a preset dynamic stair escape library and executes it to leave the current terrain. The dynamic stair escape library includes various escape instructions with different action logics.

2. The method for detecting and detaching from a stepped platform for a pool cleaning robot according to claim 1, characterized in that, The determination of the terrain type where the robot is currently located based on the multimodal environmental perception data includes: The tactile pressure distribution data is analyzed to obtain the spatial pressure distribution state and pressure value change time sequence of the pressure sensor; The collision pattern between the robot and the obstacle is identified based on the spatial pressure distribution state and the pressure value change time sequence, wherein the collision pattern includes at least one of frontal collision, side scraping and bottom pressing; The sum of the absolute values ​​of the changes in pitch angle, roll angle and yaw angle within a preset time window is calculated based on the three-dimensional attitude angle data, and this sum is used as the attitude stability value. The vertical distance from the lower surface of the robot chassis to the bottom of the swimming pool at the current location is obtained based on the chassis height data; The robot's travel state is determined based on the vertical distance, wherein the travel state includes driving up the stairs, hovering at the edge of the stairs, or repeatedly oscillating. The collision mode, the attitude stability value, and the travel state are used as input feature vectors and input to a preset decision model. The decision model outputs a determination result of the robot's current terrain category based on the input feature vectors. The determination result includes normal travel, ordinary pool wall climbing, oscillation in wide-step areas, and getting stuck in narrow-step areas.

3. The method for detecting and detaching from a stepped platform for a pool cleaning robot according to claim 2, characterized in that, The decision model is either a state decision model based on predefined rules or a trained classification machine learning model.

4. The method for detecting and detaching from a stepped platform for a pool cleaning robot according to claim 2, characterized in that, Identifying collision patterns between the robot and obstacles based on the spatial pressure distribution and the temporal sequence of pressure value changes includes: Based on the spatial pressure distribution, pressure concentration areas are identified, including pressure concentration in the central area of ​​the robot's front bumper, pressure concentration on the side of the robot, and pressure concentration in the lower front or middle part of the robot's chassis. The pressure change state is obtained based on the pressure value change time sequence, wherein the pressure change state includes a sharp increase in pressure value and a continuous change in pressure value; If the pressure is concentrated in the central area of ​​the robot's front bumper and the pressure value increases sharply, the collision pattern between the robot and the obstacle is identified as a frontal collision. If the pressure is concentrated on the side of the robot and the pressure value changes continuously, the collision pattern between the robot and the obstacle is identified as side scraping. If the pressure is concentrated at the front or lower middle of the robot chassis, the collision pattern between the robot and the obstacle is identified as bottom pressure.

5. The method for detecting and detaching from a stepped platform for a pool cleaning robot according to claim 2, characterized in that, The step of calculating the sum of the absolute values ​​of the changes in pitch angle, roll angle, and yaw angle within a preset time window based on the three-dimensional attitude angle data, and using this sum as the attitude stability value, includes: The three-dimensional attitude angle data is acquired, wherein the three-dimensional attitude angle data includes pitch angle, roll angle and yaw angle; Calculate the absolute values ​​of the angle differences between the current time and the preset time window, namely the pitch angle, roll angle, and yaw angle; The attitude stability value is obtained by adding the absolute values ​​of the angle differences between the three values.

6. The method for detecting and detaching from a stepped platform for a pool cleaning robot according to claim 2, characterized in that, Determining the robot's travel status based on the vertical distance includes: Determine whether the vertical distance exceeds a first threshold and increases dramatically within a first preset time period, and whether it remains stable within a second preset time period. If it exceeds the first threshold and increases dramatically within a first preset time period, and remains stable within a second preset time period, then determine that the travel state is driving up the stairs. Determine whether the vertical distance exhibits periodic changes within the third time period. If it exhibits periodic changes, and the period of change is synchronized with or has a fixed phase relationship with the periodic changes in the pitch angle, then the travel state is determined to be repetitive oscillation. If there is no periodic change, determine whether the vertical distance exceeds the second threshold and shows vertical distance value fluctuation within the fourth time period. If it exceeds the second threshold and shows vertical distance value fluctuation, determine that the travel state is suspended at the edge of the steps.

7. The method for detecting and detaching from a stepped platform for a pool cleaning robot according to claim 1, characterized in that, The step of adaptively selecting or generating a corresponding stair-step escape sequence from a preset dynamic stair-step escape database based on the terrain type includes: Obtain the terrain category, which includes at least wide-step area oscillation and narrow-step area jamming; When the terrain is classified as a wide stepped area oscillation, the corresponding stepped dragging action sequence is to retreat for the first preset duration and then turn at the first angle. When stuck in a narrow stepped terrain, the corresponding sequence of steps to get out of trouble is to adjust the aircraft attitude, then retreat for a second preset duration and turn at a second angle.

8. A system for detecting and detaching from a stepped platform for a pool cleaning robot, characterized in that, The system is used to implement the method as described in any one of claims 1-7, comprising: The data acquisition module is used to collect multimodal environmental perception data, including tactile pressure distribution data, three-dimensional attitude angle data, and chassis height data. The determination module is used to determine the terrain type where the robot is currently located based on the multimodal environmental perception data; The execution module is used to adaptively select or generate a corresponding stair escape action sequence from a preset dynamic stair escape library according to the terrain type and execute it to leave the current terrain. The dynamic stair escape library includes a variety of escape instructions with different action logics.

9. A swimming pool cleaning robot, characterized in that, This includes the system for detecting and disengaging from a stepped platform for a pool cleaning robot as described in claim 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images