Automatic cleaning method and system for unmanned cleaner
By performing global path planning and focus weight assignment in the unmanned cleaner, a swing path is generated to improve the cleaning effect, solving the problem of increasing the algorithm difficulty of the unmanned cleaner when it is difficult to clean stains.
Patent Information
- Application Number
- CN202510050114.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-13
- Publication Date
- 2025-05-13
AI Technical Summary
When unmanned cleaners encounter stains that are difficult to clean, local path planning adjustments lead to increased algorithm difficulty.
Obtain concerns through global path planning and generate swing paths based on the weight of concerns, increasing the working time of the unmanned cleaner at the concerns, and improving the cleaning effect.
This method improves the cleaning strength of difficult-to-clean stains, reduces algorithm complexity, and avoids the need for path changes.
Smart Images
Figure CN119987419A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to an automatic cleaning method in the technical field of household cleaning, and in particular to an automatic cleaning method and system for an unmanned cleaning device. Background Art
[0002] The behavior of an unmanned cleaner (i.e., a sweeping robot) when it encounters a hard-to-clean stain depends on its design and the technology built into it. Some advanced unmanned cleaners do have the ability to intelligently identify dirty areas and perform repeated cleaning, which usually involves adjustments to local path planning.
[0003] For example, Dreame Technology has introduced an intelligent re-washing and re-mopping function in its S20 and X20 series. When the robot detects that an area is particularly dirty (such as a kitchen floor covered with greasy dirt), it will automatically mark the area and return to the base station to clean the mop after completing the initial cleaning. After that, it will set out again to the marked heavily dirty area for re-mopping to ensure that the area is thoroughly cleaned.
[0004] But this makes the algorithm more difficult. Summary of the invention
[0005] The object of the present invention is to provide an automatic cleaning method and system for an unmanned cleaner, so as to solve the problem that the local path planning adjustment leads to increased algorithm difficulty.
[0006] To achieve the above object, one of the objects of the present invention is to provide an automatic cleaning method for an unmanned cleaning device, which comprises the following steps:
[0007] S1. Perform global path planning;
[0008] S2. Obtain the focus points in the global path and assign weights to the focus points;
[0009] S3, perform local path planning for the focus point;
[0010] S4, generating a swing path according to the weight assigned to the focus point, so as to increase the time that the unmanned cleaner remains working at the focus point and improve the cleaning effect of the focus point;
[0011] S5. The unmanned cleaner moves along the global path, the local path and the swing path, and updates the focus points during the movement.
[0012] As a further improvement of the technical solution, the swing path in S4 is a path obtained by reciprocatingly adjusting the travel angle of the unmanned cleaner on the basis of the local path;
[0013] The weights assigned to the focus points are used to determine the size and frequency of the angle adjustments.
[0014] As a further improvement of the technical solution, the global path planning in S1 includes the input of a map and the output of a global path, wherein:
[0015] Input map M, which contains obstacles, starting point S, end point E and a set of initial points of interest {P1, P2, ..., P k};
[0016] Output the optimal path from the starting point S to the end point E passing through all points of interest.
[0017] As a further improvement of the technical solution, the global path planning in S1 also includes the modification of the A* algorithm, and the specific modification steps are as follows:
[0018] S1.1. For each point of interest P i , calculate the Euclidean distance or Manhattan distance between it and the starting point S and the end point E;
[0019] S1.2, treat the focus point as a temporary target node;
[0020] S1.3, global path calculation:
[0021] S1.3.1. Use the A* algorithm to calculate the best path from the starting point S to the first focus point P1;
[0022] S1.3.2, continue to use the A* algorithm to calculate from the current focus point P i Go to the next focus point P i+1 The best path to the last point of interest;
[0023] S1.3.3. Calculate the last focus point P k The best path to the end point E.
[0024] As a further improvement of the technical solution, for the evaluation function f(n) between any two nodes, define:
[0025] h(n)′=ω·h(n);
[0026] f(n)=g(n)+h(n)′;
[0027] Where g(n) is the actual cost of node n from the starting point S; h(n) is the estimated cost from node n to the next target node, using Euclidean distance or Manhattan distance as the heuristic function; ω is an adjustable parameter greater than 1 to emphasize the focus on the point of interest.
[0028] As a further improvement of the technical solution, in S2, the weight is determined according to the difficulty of cleaning the stain at the focus point, and {σ1, σ2, ..., ρ k}.
[0029] As a further improvement of the technical solution, the sensors used for calculating the weights include 3DToF solid-state lidar sensors, RGB-D vision sensors, humidity / temperature sensors, and pressure sensors;
[0030] The weight calculation formula is as follows:
[0031] σ=α·f size +β·f depth +γ·f meterial +δ·f humidity ;
[0032] In the formula, σ is the initial weight; f size is the stain area ratio factor; f depth is the stain depth factor; f meterial is the influence factor of ground material; f humidity is the ground humidity factor; α, β, γ, and δ are the weight coefficients corresponding to each factor.
[0033] As a further improvement of the technical solution, the specific calculation process of the angle adjustment size and the adjustment frequency is as follows:
[0034]
[0035] In formula 1, σ′ is the normalized weight; σ min , σ max The minimum and maximum weight values of all focus points are used to standardize the weight of each point;
[0036] Based on the normalized weight σ′, the swing amplitude θ and frequency f corresponding to each focus point are calculated:
[0037] θ=θ min +σ′·(θ max -θ min ) Formula 2;
[0038] f=f min +σ′·(f max -f min )Formula 3.
[0039] Among them, θ min is the minimum swing amplitude; θ max is the maximum swing amplitude; f min is the minimum swing frequency; f max is the maximum swing frequency.
[0040] As a further improvement of the technical solution, the updating of the focus point during the traveling process in S5 includes re-determining the focus point or updating the existing focus point;
[0041] Among them, the updated focus is the use of 3DToF solid-state lidar sensors, RGB-D vision sensors, humidity / temperature sensors, and pressure sensors.
[0042] The second object of the present invention is to provide a system for the automatic cleaning method of an unmanned cleaning device as described in any one of the above, comprising:
[0043] A perception module, used to collect environmental information;
[0044] The control module is used to process the data from the perception module and determine the action strategy of the unmanned cleaner;
[0045] Path planning module, used to plan global and local paths;
[0046] The swing path generation module calculates the swing amplitude and frequency according to the weight of the focus point, and guides the unmanned cleaner to perform local fine cleaning without changing the local path;
[0047] A communication module, used for remotely controlling the unmanned cleaner and receiving feedback information;
[0048] as well as,
[0049] The cleaning module is used to perform the actual cleaning task.
[0050] Compared with the prior art, the present invention has the following beneficial effects:
[0051] In the automatic cleaning method and system of the unmanned cleaner, by increasing the focus points, the global path includes the locations where difficult-to-clean stains are most likely to appear in daily life from the beginning. In this way, there is no need to repeat the mopping after identification. Instead, the cleaning intensity of the difficult-to-clean stains is improved by adjusting the swing angle and frequency of the unmanned cleaner itself. Moreover, this swing will not deviate from the original local path, thereby eliminating the path change algorithm and solving the problem of the algorithm being too complicated. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] Figure 1 It is a schematic diagram of the overall method steps of the present invention. DETAILED DESCRIPTION
[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0054] For some difficult-to-clean locations in the home (such as kitchen areas, dining table areas, and balcony areas), current unmanned cleaners will replan paths for these locations, but this requires adjusting the global path or local path, which increases the difficulty of the algorithm. For this reason, this embodiment provides an automatic cleaning method for an unmanned cleaner, see Figure 1 As shown, it includes the following steps:
[0055] S1. Perform global path planning;
[0056] S2. Obtain the focus points in the global path and assign weights to the focus points;
[0057] S3, perform local path planning for the focus point;
[0058] S4, generating a swing path according to the weight assigned to the focus point, so as to increase the time that the unmanned cleaner remains working at the focus point and improve the cleaning effect of the focus point;
[0059] in:
[0060] The swing path is a path obtained by reciprocatingly adjusting the travel angle of the unmanned cleaner based on the local path;
[0061] The weights assigned to the focus points are used to determine the size and frequency of the angle adjustment;
[0062] S5. The unmanned cleaner moves along the global path, the local path and the swing path, and updates the focus points during the movement.
[0063] Specifically, the global path planning stage includes the input of the map and the output of the global path, wherein the input map M (including obstacles, a starting point S, an end point E, and a set of initial focus points {P1, P2, ..., P k}); Output the optimal path from S to E passing through all points of interest.
[0064] The global path planning stage also includes modifications to the A* algorithm to take into account the focus points. The specific modification steps are as follows:
[0065] S1.1. For each point of interest P i , calculate the Euclidean distance or Manhattan distance between it and the starting point S and the end point E;
[0066] S1.2, consider these focus points as temporary target nodes;
[0067] S1.3, global path calculation:
[0068] S1.3.1. Use the A* algorithm to calculate the best path from the starting point S to the first focus point P1;
[0069] S1.3.2, continue to use the A* algorithm to calculate from the current focus point P i Go to the next focus point P i+1 The best path to the last point of interest;
[0070] S1.3.3. Calculate the last focus point P k The best path to the end point E.
[0071] In the above calculation process, for the evaluation function f(n) between any two nodes, define:
[0072] f(n)=g(n)+h(n);
[0073] Where g(n) is the actual cost of node n from the starting point S; h(n) is the estimated cost from node n to the next target node, and the Euclidean distance or Manhattan distance can be used as the heuristic function;
[0074] Among them, for the path part involving the focus point, the weight of h(n) can be appropriately increased, that is, h(n)′=ω·h(n), where ω is an adjustable parameter greater than 1 to emphasize the attention to the focus point.
[0075] Finally, f(n)=g(n)+h(n)′.
[0076] In S2, the weight is determined according to the difficulty of cleaning the stain at the point of interest, and {σ1, σ2, …, ρ k}, for this purpose, 3DToF solid-state lidar sensor, RGB-D vision sensor, humidity / temperature sensor and pressure sensor are introduced, among which:
[0077] 3DToF solid-state lidar sensors are used to build a three-dimensional model of the environment to help identify the shape and position of objects;
[0078] RGB-D vision sensors not only capture color information, but also provide depth data, which helps to distinguish between different material surfaces and estimate the thickness of stains;
[0079] Humidity / temperature sensors are used to detect changes in ground humidity and assist in determining whether there is a liquid leak or a wet area;
[0080] The pressure sensor is installed near the cleaning brush to measure the contact force between the brush and the ground, reflecting the stubbornness of the stains.
[0081] Based on the data collected by the above four sensors, a stain difficulty scoring system is established. An initial weight σ is given to each focus point according to the stain characteristics and calculated by the following formula:
[0082] σ=α·f size+β·f depth +γ·f meterial +δ·f humidity ;
[0083] In the formula, f size is the stain area ratio factor; f depth is the stain depth factor; f meterial is the influence factor of ground material; f humidity is the ground humidity factor; α, β, γ, and δ are the weight coefficients corresponding to each factor, which can be flexibly adjusted according to the actual application scenario. The purpose of this formula is to quantify the difficulty of stain cleaning so that the cleaning efforts can be strengthened in a targeted manner in subsequent steps.
[0084] In S3, local path planning is based on the location of obstacles and points of interest. For adjacent points of interest, consider whether to merge paths to avoid repeated cleaning of a point of interest by the unmanned cleaner. The specific merging situations include:
[0085] Multiple points of interest are located within the cleaning range Z of the unmanned cleaner on the same path, and the multiple points of interest are merged into one path.
[0086] Furthermore, the weights assigned to the focus points are used to determine the size and frequency of angle adjustment. The specific calculation process is as follows:
[0087] Weight normalization: Since the weights of different focus points may vary greatly, in order to make the swing parameter calculation more reasonable, we usually normalize the original weights, that is, map them to a fixed interval, i.e., 0,1.
[0088] This can be done by transforming Eq.
[0089]
[0090] In formula 1, σ′ is the normalized weight; σ min , σ max It is the minimum and maximum weight value of all focus points, which is used to standardize the weight of each point.
[0091] Next, the swing amplitude θ and frequency f corresponding to each focus point are calculated based on the normalized weight σ′. Linear interpolation is used here to allocate the swing parameters according to the weight ratio, that is:
[0092] θ=θ min +σ′·(θ max -θ min ) Formula 2;
[0093] f=f min +σ′·(f max -fmin )Formula 3.
[0094] Among them, the minimum / maximum swing amplitude is: θ min ,θ max Define the maximum steering angles that the unmanned cleaner can take when it is at the least important and most important points of concern, respectively;
[0095] Min. / max. swing frequency: f min 、f max Defines the range of variation in the number of times the cleaner can complete a full oscillation cycle (from side to side and back again) in the same period of time.
[0096] It can be seen that as the normalized weight σ′ increases, the swing amplitude θ and frequency f also increase, thereby ensuring that the unmanned cleaner can spend more time and energy cleaning in more important locations.
[0097] Moreover, by increasing the swing amplitude θ, the cleaning range of the unmanned cleaner can be expanded.
[0098] Finally, after determining the swing amplitude θ and frequency f, we can start to implement the swing path planning. Assuming that the unmanned cleaner is currently located at a point P(x, y) in the coordinate system and is moving in a certain direction. Then, in the next time step, it should move according to the new direction angle φ, and this new direction angle is determined by the current direction plus or minus the swing amplitude:
[0099] φ(t+1)=φ(t)+(-1) t·f θ;
[0100] Here, |t·f| is a floor operator used to control the alternating direction of the swing; t represents a discretized time variable, which usually corresponds to the timestamp of each iterative update.
[0101] In addition, in S5 , the focus point is re-determined or the existing focus point is updated through the above four sensors.
[0102] The present invention also provides a system for the above-mentioned automatic cleaning method of the unmanned cleaner, which comprises:
[0103] The perception module is responsible for collecting environmental information, including but not limited to obstacle detection and ground type recognition;
[0104] The control module processes the data from the perception module and determines the action strategy of the unmanned cleaner;
[0105] Path planning module, responsible for planning global and local paths;
[0106] The swing path generation module calculates the swing amplitude and frequency according to the weight of the focus point, and guides the unmanned cleaner to perform local fine cleaning without changing the local path;
[0107] The communication module allows users to remotely control the unmanned cleaner through a mobile phone app or other smart devices and receive feedback information;
[0108] and, the cleaning module, which performs the actual cleaning tasks, such as vacuuming and mopping.
[0109] Among them, the main control chip uses a high-performance SoC (System on Chip), such as Allwinner Technology's MR813 or Rockchip RK3399, which support complex AI algorithm processing and multi-sensor fusion.
[0110] Equipped with 3DToF solid-state lidar sensor, RGB-D vision sensor, humidity / temperature sensor and pressure sensor.
[0111] Select high-efficiency brushless DC motors and their supporting drivers, such as TI's DRV8876, to ensure sufficient torque output and smooth operating performance.
[0112] Use a dedicated battery management system (BMS), such as the BQ24773 charge management chip and BQ4050 fuel gauge from Texas Instruments, to ensure a safe and reliable power supply.
[0113] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.
Claims
1. An automatic cleaning method for a robotic cleaner, characterized in that, It includes the following steps: S1. Perform global path planning; S2. Obtain the focus points in the global path and assign weights to the focus points; S3, perform local path planning for the focus point; S4, generating a swing path according to the weight assigned to the focus point, so as to increase the time that the unmanned cleaner remains working at the focus point and improve the cleaning effect of the focus point; S5. The unmanned cleaner moves along the global path, the local path and the swing path, and updates the focus points during the movement.
2. The automatic cleaning method of the unmanned cleaner according to claim 1, characterized in that, The swing path in S4 is a path obtained by reciprocatingly adjusting the travel angle of the unmanned cleaner based on the local path; The weights assigned to the focus points are used to determine the size and frequency of the angle adjustments.
3. The automatic cleaning method of the unmanned cleaner according to claim 1, wherein The global path planning in S1 includes the input of the map and the output of the global path, wherein: Input map M, which contains obstacles, starting point S, end point E and a set of initial points of interest {P1, P2, ..., P k }; Output the optimal path from the starting point S to the end point E passing through all points of interest.
4. The automatic cleaning method of the unmanned cleaner according to claim 3, characterized in that, The global path planning in S1 also includes the modification of the A* algorithm, and the specific modification steps are as follows: S1.
1. For each point of interest P i , calculate the Euclidean distance or Manhattan distance between it and the starting point S and the end point E; S1.2, treat the focus point as a temporary target node; S1.3, global path calculation: S1.3.
1. Use the A* algorithm to calculate the best path from the starting point S to the first focus point P1; S1.3.2, continue to use the A* algorithm to calculate from the current focus point P i Go to the next focus point P i+1 The best path to the last point of interest; S1.3.
3. Calculate the last focus point P k The best path to the end point E.
5. The automatic cleaning method of an unmanned cleaning device according to claim 4, characterized in that: For the evaluation function f(n) between any two nodes, define: h(n)′=ω·h(n); f(n)=g(n)+h(n)′; Where g(n) is the actual cost of node n from the starting point S; h(n) is the estimated cost from node n to the next target node, using Euclidean distance or Manhattan distance as the heuristic function; ω is an adjustable parameter greater than 1, which is used to emphasize the focus on the point of interest.
6. The automatic cleaning method of an unmanned cleaning device according to claim 1, characterized in that: In S2, the weight is determined according to the difficulty of cleaning the stain at the focus point, and {σ1, σ2, ..., ρ k }.
7. The automatic cleaning method of an unmanned cleaning device according to claim 6, characterized in that: The sensors used for calculating the weights include 3DToF solid-state lidar sensors, RGB-D vision sensors, humidity / temperature sensors, and pressure sensors; The weight calculation formula is as follows: σ=α·f size +β·f depth +γ·f meterial +δ·f humidity ; In the formula, σ is the initial weight; f size is the stain area ratio factor; f depth is the stain depth factor; f meterial is the influence factor of ground material; f humidity is the ground humidity factor; α, β, γ, and δ are the weight coefficients corresponding to each factor.
8. The automatic cleaning method of an unmanned cleaning device according to claim 7, characterized in that: The specific calculation process of the angle adjustment size and adjustment frequency is as follows: In Equation 1, σ′ is the normalized weight; σ min , σ max are the minimum and maximum weight values among all the focus points, which are used to standardize the weights of each point; Based on the normalized weight σ′, the swing amplitude θ and frequency f corresponding to each focus point are calculated: θ = θ min + σ′·(θ max - θ min ) Equation 2; f=f min +σ′·(f max -f min )Formula 3. Among them, θ min is the minimum swing amplitude; θ max is the maximum swing amplitude; f min is the minimum swing frequency; f max is the maximum swing frequency.
9. The automatic cleaning method of an unmanned cleaning device according to claim 7, characterized in that: The updating of the focus point during the traveling process in S5 includes re-determining the focus point or updating the existing focus point; Among them, the updated focus is the use of 3DToF solid-state lidar sensors, RGB-D vision sensors, humidity / temperature sensors, and pressure sensors.
10. A system for the automatic cleaning method of an unmanned cleaner according to any one of claims 1 to 9, characterized in that: It includes: Perception module, used to collect environmental information; The control module is used to process the data from the perception module and determine the action strategy of the unmanned cleaner; Path planning module, used to plan global and local paths; The swing path generation module calculates the swing amplitude and frequency according to the weight of the focus point, and guides the unmanned cleaner to perform local fine cleaning without changing the local path; A communication module, used for remotely controlling the unmanned cleaner and receiving feedback information; as well as, The cleaning module is used to perform the actual cleaning task.