Robot for automatically cleaning garbage on water surface

Abstract

The invention discloses a robot for automatically cleaning water surface garbage, which comprises the following steps: S1, collecting water surface garbage distribution and obstacle data by using an active optical detection technology and a CMOS (Complementary Metal Oxide Semiconductor) photosensitive device, and constructing a water surface environment data model; s2, performing path planning based on an A * path planning algorithm in combination with a GBFS algorithm and a Dijkstra algorithm, and generating a candidate path set; s3, optimizing a path by using a particle swarm hybrid Poisson jump algorithm, and outputting an optimal path; s4, transmitting the optimal path to an STM32 single-chip microcomputer, and controlling the robot to move through a PWM motor driving module; s5, monitoring and adjusting a path in real time through an ultrasonic obstacle avoidance system; s6, a garbage collection device is started, garbage and oil dirt are collected through a double-roller adsorption system, and water is filtered; s7, detecting the storage amount based on a pressure sensor, and triggering an automatic return path; and S8, the operation report is fed back through the remote monitoring module. Intelligent and automatic water surface garbage cleaning is achieved, and the cleaning efficiency and the environmental adaptability are effectively improved.

Description

Technical Field

[0001] The invention relates to the technical field of intelligent control systems, and in particular to a robot for automatically cleaning garbage on a water surface. Background Art

[0002] At present, with the rapid development of industrialization and urbanization, water surface garbage pollution has become a common environmental problem in rivers, lakes, ports and other water bodies. These surface garbage not only damages the ecological environment and affects the survival of aquatic plants and animals, but also has a serious negative impact on the development of tourism resources, shipping safety and the quality of life of residents. The sources of surface garbage are diverse, including domestic waste, industrial waste, oil pollution, etc. Cleaning up these garbage has become an important issue in water environment management.

[0003] The traditional method of cleaning up garbage on the water surface mainly relies on manual salvage by boat. Although this method can solve the problem of garbage on the water surface to a certain extent, it has many limitations. For places such as small rivers, artificial lakes and water playgrounds, due to the narrow water surface and shallow water depth, it is difficult for ships to pass, resulting in difficulty in garbage salvage. In addition, manual salvage relies on human operation, with low work efficiency and high cost, and workers are exposed to polluted waters for a long time, which poses a safety hazard. In severely polluted or scenic waters, manual salvage may also damage the surrounding ecological environment and beauty, further limiting its scope of application.

[0004] In order to improve the automation level of surface garbage cleaning, some surface garbage cleaning devices have appeared in the prior art, such as large-scale garbage cleaning ships. These devices have excellent performance in endurance and processing capacity, and are suitable for garbage cleaning in large areas of water. However, large-scale garbage cleaning ships are large in size, cannot move flexibly in narrow waters, and have high operating costs. They are not suitable for small rivers, artificial lakes and other scenes. In addition, the automation level of garbage cleaning ships is limited, and manual control is still required, which cannot achieve fully autonomous and intelligent cleaning operations.

[0005] Although the surface garbage cleaning robots on the market have initially achieved automation, most robots still have several outstanding problems. First, these robots are often large in size and complex in structure, and cannot adapt to complex and narrow water environments. Secondly, the existing cleaning robots mainly rely on remote control operations, lack intelligent path planning and obstacle avoidance functions, and cannot independently complete garbage cleaning tasks. In addition, the existing robots have low recognition accuracy for garbage distribution areas and lack efficient path planning and dynamic optimization technology, resulting in low operating efficiency and high energy consumption, which affects the actual application effect. Therefore, how to provide a robot that automatically cleans surface garbage is a problem that technicians in this field urgently need to solve. Summary of the invention

[0006] One purpose of the present invention is to propose a robot for automatically cleaning garbage on the water surface. The present invention adopts active optical detection technology, A* path planning algorithm, particle swarm hybrid Poisson jump algorithm and ultrasonic obstacle avoidance technology, combines a solar power supply system and a garbage collection device, and describes in detail the intelligent implementation method of automatic detection of water surface garbage, path planning, dynamic obstacle avoidance and garbage collection. The robot has the advantages of high efficiency, intelligence and adaptability to complex water environments.

[0007] A robot for automatically cleaning water surface garbage according to an embodiment of the present invention comprises the following steps:

[0008] S1. Collect the distribution data of garbage on the water surface and the location information of obstacles, use active optical detection technology and CMOS photosensitive devices to convert scene information into electrical signals, and build a water surface environment data model through multi-sensor fusion technology;

[0009] S2, using the A* path planning algorithm combined with the GBFS algorithm and the Dijkstra algorithm to perform preliminary path planning on the water surface environment data model and generate a set of candidate paths;

[0010] S3, input the candidate path set into the particle swarm hybrid Poisson jump algorithm, dynamically perturb the particle swarm by introducing the Poisson jump strategy to avoid falling into the local optimal solution, and generate the optimal path according to the path length, energy consumption and obstacle avoidance requirements;

[0011] S4, transmit the optimal path data to the STM32 single-chip microcomputer control system, use the PWM motor drive module to control the robot to move along the optimal path, and use solar power supply technology to provide the required power for operation;

[0012] S5. Monitor obstacles in front of the robot in real time through the ultrasonic obstacle avoidance system. When an obstacle is detected, dynamically adjust the robot's path direction.

[0013] S6. When the robot moves along the path, it starts the water surface garbage collection device, uses the double roller adsorption and separation system to collect garbage and oil, and stores the collected materials in the built-in recovery cabin, and uses the filtering device to preliminarily purify the water body;

[0014] S7. When the robot detects that the storage volume in the recovery cabin reaches the set threshold, the pressure sensor technology is used to trigger the automatic closure of the collection device, and the navigation automatic memory and power protection technology are used to guide the robot back to the initial position along the return path;

[0015] S8. After the robot completes the garbage cleaning task, it will provide the user terminal with a cleaning path, garbage distribution heat map, and collection status report through the remote monitoring module, and enter standby mode.

[0016] Optionally, S2 includes the following steps:

[0017] S21, based on water surface environment data model M f , extract the characteristic data F of the garbage distribution area and obstacle distribution area f , define the current position of the robot as the starting point S, define the target cleaning area as the end point G, and divide the water surface environment into grid cells C(i,j), where i and j are the row and column coordinates of the grid respectively;

[0018] S22. In the preliminary path planning, the core cost function f(n) of the A* algorithm is used to combine the actual cost g(n) with the heuristic cost h(n). The total cost function is expressed as:

[0019] f(n)=g(n)+h(n);

[0020] Among them, g(n) is the actual cost from the starting point S to the current grid node n; h(n) is the heuristic estimated cost from the current grid node n to the target node G;

[0021] S23. In the process of calculating g(n), the actual cost of the grid is dynamically updated by combining the shortest path search characteristics of the Dijkstra algorithm, which is specifically expressed as:

[0022] g(n)=g(p)+d(p,n)·[1+κ·Δ obstacle (n)];

[0023] Among them, g(p) is the actual cost of the parent node p, d(p,n) is the path cost from the parent node p to the current node n; Δ obstacle (n) represents the density of obstacles around node n; κ is the obstacle density penalty coefficient, which reflects the cost weight of the actual path in a complex environment;

[0024] S24. When calculating the heuristic cost h(n), the greedy best first search GBFS algorithm is introduced. By greedily searching the directional characteristics of the target area, the cost function is dynamically adjusted. The specific expression is:

[0025] h(n)=α·d euclid (n,G)+β·d curvature (n,G);

[0026] Among them, d euclid (n,G) is the Euclidean distance from node n to target point G; d curvature (n,G) is the correction amount of path curvature to the heuristic cost, expressed as:

[0027]

[0028] Among them, α and β are weight coefficients, and they satisfy α+β=1;

[0029] S25. Substitute the actual cost g(n) calculated by the Dijkstra algorithm and the heuristic cost h(n) calculated by the GBFS algorithm into the total cost function f(n) of the A* algorithm, and sort the candidate paths based on the dynamic weight adjustment mechanism to generate a set of candidate paths. Where n is the number of paths;

[0030] S26, evaluate the paths in the candidate path set P, and select the path P with the minimum total cost according to the minimum value priority principle of the total cost function f(n) opt The preliminary optimal path is used as output to provide input data for robot execution.

[0031] Optionally, S3 includes the following steps:

[0032] S31, the generated candidate path set As the particle swarm initialization input, define the position of the particles and speed in is the position of the particle in the path coordinate space, Indicates the dynamic change direction of the particle's current position;

[0033] S32. Design the fitness function F(X) based on path length, energy consumption and obstacle avoidance performance. i ), defined as:

[0034] F(X i )=λ1·L(X i )+λ2·E(X i )+λ3·A(X i );

[0035] Among them, L(X i ) represents the total length of the path, d(C k ,C k+1 ) is the Euclidean distance between adjacent nodes on the path; Represents the path energy consumption, which is calculated by combining the obstacle density and path curvature; Δ obstacle (C k ) is node C k Surrounding obstacle density, d curvature is the path curvature; A(X i )=min k [d obstacle (C k)] represents the obstacle avoidance performance, which calculates the minimum distance from the node in the path to the nearest obstacle; λ1, λ2, and λ3 are the weight factors of path length, energy consumption, and obstacle avoidance performance, respectively, satisfying λ1+λ2+λ3=1;

[0036] S33, initialize the particle swarm and set the initial position and initial velocity And introduce the inertia weight w, individual learning factor c1 and global learning factor c2, and update the particle speed and position through the following formula:

[0037]

[0038] in, is the historical optimal position of particle i, is the global optimal position; r1 and r2 are random numbers between [0,1], w is the inertia weight; J poisson (t) is the Poisson jump term, expressed as:

[0039]

[0040] Among them, δ k is the jump amplitude, λ k is the Poisson distribution parameter;

[0041] S34, in each iteration, dynamically adjust the particle's inertia weight w and learning factors c1, c2, and adjust the parameters in combination with the adaptive mechanism;

[0042] S35, introduce the path perturbation mechanism, randomly perturb the particle position through Poisson jump, enhance the global search capability, and select the path with the best fitness;

[0043] S36, according to the fitness function F(X i ) Evaluate the fitness values ​​of all particles and select the particle position X with the smallest fitness value opt As the optimal path P opt , and output the optimal path result to provide path input data for robot execution.

[0044] Optionally, S4 includes the following steps:

[0045] S41, the optimal path P opt ={C1,C2,...,C m} Format the path data and extract each path node C k (x k ,y k ) and generate the path data set D path ={(x1,y1),(x2,y2),...,(xm ,y m )};

[0046] S42, path data set D path Transmitted to the STM32 microcontroller control system through the serial communication interface to define the robot's current position C current (x c ,y c ) and the target position C target (x t ,y t );

[0047] S43. Based on the PWM control principle, design the speed control function V of the motor drive control , control the speed difference of the left and right motors to realize the robot path movement and direction steering, which is specifically expressed as:

[0048]

[0049] Among them, V max is the maximum control speed of the motor; Δ angle is the offset angle between the robot's current position and the target position; ∈ is the allowed error threshold of the offset angle;

[0050] S44, during the robot movement, based on the path data set D path Dynamically update the current position C current (x c ,y c ), and according to the path target node C target )x t ,y t ) Calculate the offset distance d between the current position and the target position offset ;

[0051] S45: In response to the dynamic path deviation during the robot's movement, the PWM control signals of the left and right motors are adjusted in real time, and the duty cycle of the left and right motors is set to D L and D R ;

[0052] S46, compare the actual moving trajectory data of the robot during the movement with the optimal path P opt Compare and generate deviation correction value Δ correction , and update the motor control parameters in real time to correct the robot path;

[0053] S47. Use solar power supply technology to provide the required electricity for operation, monitor the power supply status in real time, and trigger an early warning signal when the power supply is insufficient to ensure the normal working state of the robot.

[0054] Optionally, S5 includes the following steps:

[0055] S51, using an ultrasonic sensor to measure the distance from the current position to the obstacle in real time, and compare it with the obstacle avoidance threshold;

[0056] S52, when the measured distance is less than the threshold, triggering the obstacle avoidance mechanism, and recording the obstacle position and path deviation;

[0057] S53, combining the duty cycle adjustment of the left and right motors, and dynamically adjusting the PWM control signal to change the robot's forward direction;

[0058] S54, during the obstacle avoidance process, the robot path data is updated in real time, and the temporary path is merged with the optimal path to form a corrected path;

[0059] S55: recalculate the path cost from the current position to the target node based on the path fusion result and the current environment data.

[0060] Optionally, the S6 includes the following steps:

[0061] S61. When the robot enters the garbage distribution area, it controls the dual-roller adsorption system to start and monitors the adsorption status of the rollers in real time;

[0062] S62, dynamically adjusting the rolling speed and adsorption strength of the roller according to the running speed and adsorption efficiency of the roller to adapt to the distribution of garbage with different densities;

[0063] S63, during the adsorption process, the garbage storage volume is detected in real time, and when the storage volume reaches a preset threshold, the separation system is triggered to transfer the adsorbed garbage from the roller surface to the recycling cabin;

[0064] S64, integrating the garbage collection data with the path execution data, recording the garbage distribution state corresponding to each path node, and forming garbage collection state data;

[0065] S65. When the garbage collection is completed or the storage capacity of the recycling compartment reaches the upper limit, the roller adsorption system is stopped and the current task status is fed back to the robot control system.

[0066] Optionally, the S7 includes the following steps:

[0067] S71, detecting the storage status of the recovery chamber in real time through a pressure sensor, and triggering the collection device to automatically shut down when the storage volume reaches a threshold;

[0068] S72, recording the current robot position information, and generating a return path in combination with navigation automatic memory technology;

[0069] S73, through the power protection technology, dynamically adjust the movement speed and motor load of the robot return path;

[0070] S74, monitoring the path deviation between the current position and the initial position of the robot in real time, and performing path correction;

[0071] S75. Record the environmental data and task status collected during the return process to provide reference data for subsequent tasks.

[0072] Optionally, the S8 includes the following steps:

[0073] S81, integrating the path data, garbage collection data and storage status data recorded during the task execution process to generate a task operation report;

[0074] S82. Generate a garbage distribution heat map based on the garbage distribution status data, and synchronize the operation report and the heat map to the user terminal;

[0075] S83, displaying the operation report on the user terminal, including the cleaning path, garbage distribution status, task execution time and recycling cabin storage status information;

[0076] S84. After completing the data feedback, the robot enters the standby state and maintains the online connection of the remote monitoring module, waiting for the user to issue new task instructions.

[0077] The beneficial effects of the present invention are:

[0078] The present invention proposes a robot for automatically cleaning garbage on the water surface. Through active optical detection technology and CMOS photosensitive devices, it collects real-time data on the distribution of garbage on the water surface and obstacle information, and combines multi-sensor fusion technology to build a water surface environment data model, thereby achieving high-precision garbage detection and environmental perception, and effectively improving the adaptability of the system in complex water environments.

[0079] The present invention utilizes the A* path planning algorithm combined with the GBFS algorithm and the Dijkstra algorithm for path planning, and further optimizes the path through the particle swarm hybrid Poisson jump algorithm to avoid the path search from falling into the local optimal solution, thereby generating the optimal path with the minimum energy consumption and excellent obstacle avoidance performance, which greatly improves the execution efficiency of the robot cleaning task and the accuracy of path planning.

[0080] The present invention adopts a double-roller adsorption and separation system to collect garbage and oil on the water surface, and performs preliminary purification of the water body through a filtering device, thereby achieving simultaneous garbage cleaning and water purification, and improving the comprehensive effect of water environment management.

[0081] The present invention uses solar power supply technology and pressure sensor technology to trigger the automatic return function when the garbage storage volume reaches a set threshold, and guides the robot to return through navigation automatic memory technology, effectively realizing the intelligence, autonomy and long endurance of the system, and reducing energy consumption and the need for manual intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0083] Figure 1 A flow chart of a robot for automatically cleaning water surface garbage proposed by the present invention;

[0084] Figure 2 A schematic diagram of a robot for automatically cleaning water surface garbage proposed by the present invention;

[0085] Figure 3 This is a flow chart of a water purification device of a robot for automatically cleaning garbage on the water surface proposed by the present invention. DETAILED DESCRIPTION

[0086] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0087] refer to Figure 1-3 , a robot for automatically cleaning garbage on the water surface, comprising the following steps:

[0088] S1. Collect the distribution data of garbage on the water surface and the location information of obstacles, use active optical detection technology and CMOS photosensitive devices to convert scene information into electrical signals, and build a water surface environment data model through multi-sensor fusion technology;

[0089] S2, using the A* path planning algorithm combined with the GBFS algorithm and the Dijkstra algorithm to perform preliminary path planning on the water surface environment data model and generate a set of candidate paths;

[0090] S3, input the candidate path set into the particle swarm hybrid Poisson jump algorithm, dynamically perturb the particle swarm by introducing the Poisson jump strategy to avoid falling into the local optimal solution, and generate the optimal path according to the path length, energy consumption and obstacle avoidance requirements;

[0091] S4, transmit the optimal path data to the STM32 single-chip microcomputer control system, use the PWM motor drive module to control the robot to move along the optimal path, and use solar power supply technology to provide the required power for operation;

[0092] S5. Monitor obstacles in front of the robot in real time through the ultrasonic obstacle avoidance system. When an obstacle is detected, dynamically adjust the robot's path direction.

[0093] S6. When the robot moves along the path, it starts the water surface garbage collection device, uses the double roller adsorption and separation system to collect garbage and oil, and stores the collected materials in the built-in recovery cabin, and uses the filtering device to preliminarily purify the water body;

[0094] S7. When the robot detects that the storage volume in the recovery cabin reaches the set threshold, the pressure sensor technology is used to trigger the automatic closure of the collection device, and the navigation automatic memory and power protection technology are used to guide the robot back to the initial position along the return path;

[0095] S8. After the robot completes the garbage cleaning task, it will provide the user terminal with a cleaning path, garbage distribution heat map, and collection status report through the remote monitoring module, and enter standby mode.

[0096] In this implementation, S2 includes the following steps:

[0097] S21, based on water surface environment data model M f , extract the characteristic data F of the garbage distribution area and obstacle distribution area f , define the current position of the robot as the starting point S, define the target cleaning area as the end point G, and divide the water surface environment into grid cells C(i,j), where i and j are the row and column coordinates of the grid respectively;

[0098] S22. In the preliminary path planning, the core cost function f(n) of the A* algorithm is used to combine the actual cost g(n) with the heuristic cost h(n). The total cost function is expressed as:

[0099] f(n)=g(n)+h(n);

[0100] Among them, g(n) is the actual cost from the starting point S to the current grid node n; h(n) is the heuristic estimated cost from the current grid node n to the target node G;

[0101] S23. In the process of calculating g(n), the actual cost of the grid is dynamically updated by combining the shortest path search characteristics of the Dijkstra algorithm, which is specifically expressed as:

[0102] g(n)=g(p)+d(p,n)·[1+κ·Δ obstacle (n)];

[0103] Among them, g(p) is the actual cost of the parent node p, d(p,n) is the path cost from the parent node p to the current node n; Δ obstacle (n) represents the density of obstacles around node n; κ is the obstacle density penalty coefficient, which reflects the cost weight of the actual path in a complex environment;

[0104] S24. When calculating the heuristic cost h(n), the greedy best first search GBFS algorithm is introduced. By greedily searching the directional characteristics of the target area, the cost function is dynamically adjusted. The specific expression is:

[0105] h(n)=α·d euclid (n,G)+β·d curvature (n,G);

[0106] Among them, d euclid (n,G) is the Euclidean distance from node n to target point G; d curvature (n,G) is the correction amount of path curvature to the heuristic cost, expressed as:

[0107]

[0108] Among them, α and β are weight coefficients, and they satisfy α+β=1;

[0109] S25. Substitute the actual cost g(n) calculated by the Dijkstra algorithm and the heuristic cost h(n) calculated by the GBFS algorithm into the total cost function f(n) of the A* algorithm, and sort the candidate paths based on the dynamic weight adjustment mechanism to generate a set of candidate paths. Where n is the number of paths;

[0110] S26, evaluate the paths in the candidate path set P, and select the path P with the minimum total cost according to the minimum value priority principle of the total cost function f(n) opt The preliminary optimal path is used as output to provide input data for robot execution.

[0111] In this implementation, S3 includes the following steps:

[0112] S31, the generated candidate path set As the particle swarm initialization input, define the position of the particles and speed in is the position of the particle in the path coordinate space, Indicates the dynamic change direction of the particle's current position;

[0113] S32. Design the fitness function F(X) based on path length, energy consumption and obstacle avoidance performance. i ), defined as:

[0114] F(X i )=λ1·L(X i )+λ2·E(X i )+λ3·A(X i );

[0115] Among them, L(X i ) represents the total length of the path, d(C k ,C k+1 ) is the Euclidean distance between adjacent nodes on the path; Represents the path energy consumption, which is calculated by combining the obstacle density and path curvature; Δ obstacle (C k ) is node C k Surrounding obstacle density, d curvature is the path curvature; A(X i )=min k [d obstacle (C k )] represents the obstacle avoidance performance, which calculates the minimum distance from the node in the path to the nearest obstacle; λ1, λ2, and λ3 are the weight factors of path length, energy consumption, and obstacle avoidance performance, respectively, satisfying λ1+λ2+λ3=1;

[0116] S33, initialize the particle swarm and set the initial position and initial velocity And introduce the inertia weight w, individual learning factor c1 and global learning factor c2, and update the particle speed and position through the following formula:

[0117]

[0118] in, is the historical optimal position of particle i, is the global optimal position; r1 and r2 are random numbers between [0,1], w is the inertia weight; J poisson (t) is the Poisson jump term, expressed as:

[0119]

[0120] Among them, δ k is the jump amplitude, λ k is the Poisson distribution parameter;

[0121] S34, in each iteration, dynamically adjust the particle's inertia weight w and learning factors c1, c2, and adjust the parameters in combination with the adaptive mechanism;

[0122] S35, introduce the path perturbation mechanism, randomly perturb the particle position through Poisson jump, enhance the global search capability, and select the path with the best fitness;

[0123] S36, according to the fitness function F(X i ) Evaluate the fitness values ​​of all particles and select the particle position X with the smallest fitness value opt As the optimal path P opt, and output the optimal path result to provide path input data for robot execution.

[0124] In this implementation, S4 includes the following steps:

[0125] S41, the optimal path P opt ={C1,C2,...,C m} Format the path data and extract each path node C k (x k ,y k ) and generate the path data set D path ={(x1,y1),(x2,y2),...,(x m ,y m )};

[0126] S42, path data set D path Transmitted to the STM32 microcontroller control system through the serial communication interface to define the robot's current position C current (x c ,y c ) and the target position C target (x t ,y t );

[0127] S43. Based on the PWM control principle, design the speed control function V of the motor drive control , control the speed difference of the left and right motors to realize the robot path movement and direction steering, which is specifically expressed as:

[0128]

[0129] Among them, V max is the maximum control speed of the motor; Δ angle is the offset angle between the robot's current position and the target position; ∈ is the allowed error threshold of the offset angle;

[0130] S44, during the robot movement, based on the path data set D path Dynamically update the current position C current (x c ,y c ), and according to the path target node C target (x t ,y t ) Calculate the offset distance d between the current position and the target position offset ;

[0131] S45: In response to the dynamic path deviation during the robot's movement, the PWM control signals of the left and right motors are adjusted in real time, and the duty cycle of the left and right motors is set to DL and D R ;

[0132] S46, compare the actual moving trajectory data of the robot during the movement with the optimal path P opt Compare and generate deviation correction value Δ correction , and update the motor control parameters in real time to correct the robot path;

[0133] S47. Use solar power supply technology to provide the required electricity for operation, monitor the power supply status in real time, and trigger an early warning signal when the power supply is insufficient to ensure the normal working state of the robot.

[0134] In this implementation, S5 includes the following steps:

[0135] S51, using an ultrasonic sensor to measure the distance from the current position to the obstacle in real time, and compare it with the obstacle avoidance threshold;

[0136] S52, when the measured distance is less than the threshold, triggering the obstacle avoidance mechanism, and recording the obstacle position and path deviation;

[0137] S53, combining the duty cycle adjustment of the left and right motors, and dynamically adjusting the PWM control signal to change the robot's forward direction;

[0138] S54, during the obstacle avoidance process, the robot path data is updated in real time, and the temporary path is merged with the optimal path to form a corrected path;

[0139] S55: recalculate the path cost from the current position to the target node based on the path fusion result and the current environment data.

[0140] In this implementation, S6 includes the following steps:

[0141] S61. When the robot enters the garbage distribution area, it controls the dual-roller adsorption system to start and monitors the adsorption status of the rollers in real time;

[0142] S62, dynamically adjusting the rolling speed and adsorption strength of the roller according to the running speed and adsorption efficiency of the roller to adapt to the distribution of garbage with different densities;

[0143] S63, during the adsorption process, the garbage storage volume is detected in real time, and when the storage volume reaches a preset threshold, the separation system is triggered to transfer the adsorbed garbage from the roller surface to the recycling cabin;

[0144] S64, integrating the garbage collection data with the path execution data, recording the garbage distribution state corresponding to each path node, and forming garbage collection state data;

[0145] S65. When the garbage collection is completed or the storage capacity of the recycling compartment reaches the upper limit, the roller adsorption system is stopped and the current task status is fed back to the robot control system.

[0146] In this implementation, S7 includes the following steps:

[0147] S71, detecting the storage status of the recovery chamber in real time through a pressure sensor, and triggering the collection device to automatically shut down when the storage volume reaches a threshold;

[0148] S72, recording the current robot position information, and generating a return path in combination with navigation automatic memory technology;

[0149] S73, through the power protection technology, dynamically adjust the movement speed and motor load of the robot return path;

[0150] S74, monitoring the path deviation between the current position and the initial position of the robot in real time, and performing path correction;

[0151] S75. Record the environmental data and task status collected during the return process to provide reference data for subsequent tasks.

[0152] In this implementation, S8 includes the following steps:

[0153] S81, integrating the path data, garbage collection data and storage status data recorded during the task execution process to generate a task operation report;

[0154] S82. Generate a garbage distribution heat map based on the garbage distribution status data, and synchronize the operation report and the heat map to the user terminal;

[0155] S83, displaying the operation report on the user terminal, including the cleaning path, garbage distribution status, task execution time and recycling cabin storage status information;

[0156] S84. After completing the data feedback, the robot enters the standby state and maintains the online connection of the remote monitoring module, waiting for the user to issue new task instructions.

[0157] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A robot for automatically cleaning up garbage on the water surface, characterized in that: The steps include: S1. Collect the distribution data of garbage on the water surface and the location information of obstacles, use active optical detection technology and CMOS photosensitive devices to convert scene information into electrical signals, and build a water surface environment data model through multi-sensor fusion technology; S2, using the A* path planning algorithm combined with the GBFS algorithm and the Dijkstra algorithm to perform preliminary path planning on the water surface environment data model and generate a set of candidate paths; S3, input the candidate path set into the particle swarm hybrid Poisson jump algorithm, dynamically perturb the particle swarm by introducing the Poisson jump strategy to avoid falling into the local optimal solution, and generate the optimal path according to the path length, energy consumption and obstacle avoidance requirements; S4, transmit the optimal path data to the STM32 single-chip microcomputer control system, use the PWM motor drive module to control the robot to move along the optimal path, and use solar power supply technology to provide the required power for operation; S5. Monitor obstacles in front of the robot in real time through the ultrasonic obstacle avoidance system. When an obstacle is detected, dynamically adjust the robot's path direction. S6. When the robot moves along the path, it starts the water surface garbage collection device, uses the double roller adsorption and separation system to collect garbage and oil, and stores the collected materials in the built-in recovery cabin, and uses the filtering device to preliminarily purify the water body; S7. When the robot detects that the storage volume in the recovery cabin reaches the set threshold, the pressure sensor technology is used to trigger the automatic closure of the collection device, and the navigation automatic memory and power protection technology are used to guide the robot back to the initial position along the return path; S8. After the robot completes the garbage cleaning task, it will provide the user terminal with a cleaning path, garbage distribution heat map, and collection status report through the remote monitoring module, and enter standby mode.

2. The robot for automatically cleaning water surface garbage according to claim 1, characterized in that: The S2 comprises the following steps: S21, based on water surface environment data model M f , extract the characteristic data F of the garbage distribution area and obstacle distribution area f , define the current position of the robot as the starting point S, define the target cleaning area as the end point G, and divide the water surface environment into grid cells C(i,j), where i and j are the row and column coordinates of the grid respectively; S22. In the preliminary path planning, the core cost function f(n) of the A* algorithm is used to combine the actual cost g(n) with the heuristic cost h(n). The total cost function is expressed as: f(n)=g(n)+h(n); Among them, g(n) is the actual cost from the starting point S to the current grid node n; h(n) is the heuristic estimated cost from the current grid node n to the target node G; S23. In the process of calculating g(n), the actual cost of the grid is dynamically updated by combining the shortest path search characteristics of the Dijkstra algorithm, which is specifically expressed as: g(n)=g(p)+d(p,n)·[1+κ·Δ obstacle (n)]; Among them, g(p) is the actual cost of the parent node p, d(p,n) is the path cost from the parent node p to the current node n; Δ obstacle (n) represents the density of obstacles around node n; κ is the obstacle density penalty coefficient, which reflects the cost weight of the actual path in a complex environment; S24. When calculating the heuristic cost h(n), the greedy best first search GBFS algorithm is introduced. By greedily searching the directional characteristics of the target area, the cost function is dynamically adjusted. The specific expression is: h(n)=α·d euclid (n,G)+β·d curvature (n,G); Among them, d euclid (n,G) is the Euclidean distance from node n to target point G; d curvature (n,G) is the correction amount of the path curvature to the heuristic cost, expressed as: Among them, α and β are weight coefficients, and they satisfy α+β=1; S25. Substitute the actual cost g(n) calculated by the Dijkstra algorithm and the heuristic cost h(n) calculated by the GBFS algorithm into the total cost function f(n) of the A* algorithm, and sort the candidate paths based on the dynamic weight adjustment mechanism to generate a set of candidate paths. Where n is the number of paths; S26, evaluate the paths in the candidate path set P, and select the path P with the minimum total cost according to the minimum value priority principle of the total cost function f(n) opt The preliminary optimal path is used as output to provide input data for robot execution.

3. The robot for automatically cleaning water surface garbage according to claim 1, characterized in that: The S3 comprises the following steps: S31, the generated candidate path set As the particle swarm initialization input, define the position of the particles and speed in is the position of the particle in the path coordinate space, Indicates the dynamic change direction of the particle's current position; S32. Design the fitness function F(X) based on path length, energy consumption and obstacle avoidance performance. i ), defined as: F(X i )=λ1·L(X i )+λ2·E(X i )+λ3·A(X i ); Among them, L(X i ) represents the total length of the path, d(C k ,C k+1 ) is the Euclidean distance between adjacent nodes on the path; Represents the path energy consumption, which is calculated by combining the obstacle density and path curvature; Δ obstacle (C k ) is node C k Surrounding obstacle density, d curvature is the path curvature; A(X i )=min k [d obstacle (C k ) ] represents the obstacle avoidance performance, and calculates the minimum distance from the node in the path to the nearest obstacle; λ1, λ2, and λ3 are the weight factors of path length, energy consumption, and obstacle avoidance performance, respectively, satisfying λ1+λ2+λ3=1; S33, initialize the particle swarm and set the initial position and initial velocity And introduce the inertia weight w, individual learning factor c1 and global learning factor c2, and update the particle speed and position through the following formula: in, is the historical optimal position of particle i, is the global optimal position; r1 and r2 are random numbers between [0,1], w is the inertia weight; J poisson (t) is the Poisson jump term, expressed as: Among them, δ k is the jump amplitude, λ k is the Poisson distribution parameter; S34, in each iteration, dynamically adjust the particle's inertia weight w and learning factors c1, c2, and adjust the parameters in combination with the adaptive mechanism; S35, introduce the path perturbation mechanism, randomly perturb the particle position through Poisson jump, enhance the global search capability, and select the path with the best fitness; S36, according to the fitness function F(X i ) Evaluate the fitness values ​​of all particles and select the particle position X with the smallest fitness value opt As the optimal path P opt , and output the optimal path result to provide path input data for robot execution.

4. The robot for automatically cleaning water surface garbage according to claim 1, characterized in that: The S4 comprises the following steps: S41, the optimal path P opt ={C1,C2,...,C m } Format the path data and extract each path node C k (x k ,y k ) and generate the path data set D path ={(x1,y1),(x2,y2),...,(x m ,y m )}; S42, path data set D path Transmitted to the STM32 microcontroller control system through the serial communication interface to define the robot's current position C current (x c ,y c ) and the target position C target (x t ,y t ); S43. Based on the PWM control principle, design the speed control function V of the motor drive control , control the speed difference of the left and right motors to realize the robot path movement and direction steering, which is specifically expressed as: Among them, V max is the maximum control speed of the motor; Δ angle is the offset angle between the robot’s current position and the target position; ∈ is the error threshold allowed for the offset angle; S44, during the robot movement, based on the path data set D path Dynamically update the current position C current (x c ,y c ), and according to the path target node C target (x t ,y t ) Calculate the offset distance d between the current position and the target position offset ; S45: In response to the dynamic path deviation during the robot's movement, the PWM control signals of the left and right motors are adjusted in real time, and the duty cycle of the left and right motors is set to D L and D R ; S46, compare the actual moving trajectory data of the robot during the movement with the optimal path P opt Compare and generate deviation correction value Δ correction , and update the motor control parameters in real time to correct the robot path; S47. Use solar power supply technology to provide the required electricity for operation, monitor the power supply status in real time, and trigger an early warning signal when the power supply is insufficient to ensure the normal working state of the robot.

5. The robot for automatically cleaning water surface garbage according to claim 1, characterized in that: The S5 comprises the following steps: S51, using an ultrasonic sensor to measure the distance from the current position to the obstacle in real time, and compare it with the obstacle avoidance threshold; S52, when the measured distance is less than the threshold, triggering the obstacle avoidance mechanism, and recording the obstacle position and path deviation; S53, combining the duty cycle adjustment of the left and right motors, and dynamically adjusting the PWM control signal to change the robot's forward direction; S54, during the obstacle avoidance process, the robot path data is updated in real time, and the temporary path is merged with the optimal path to form a corrected path; S55: recalculate the path cost from the current position to the target node based on the path fusion result and the current environment data.

6. The robot for automatically cleaning water surface garbage according to claim 1, characterized in that: The S6 comprises the following steps: S61. When the robot enters the garbage distribution area, it controls the dual-roller adsorption system to start and monitors the adsorption status of the rollers in real time; S62, dynamically adjusting the rolling speed and adsorption strength of the roller according to the running speed and adsorption efficiency of the roller to adapt to the distribution of garbage with different densities; S63, during the adsorption process, the garbage storage volume is detected in real time, and when the storage volume reaches a preset threshold, the separation system is triggered to transfer the adsorbed garbage from the roller surface to the recycling cabin; S64, integrating the garbage collection data with the path execution data, recording the garbage distribution state corresponding to each path node, and forming garbage collection state data; S65. When the garbage collection is completed or the storage capacity of the recycling compartment reaches the upper limit, the roller adsorption system is stopped and the current task status is fed back to the robot control system.

7. The robot for automatically cleaning water surface garbage according to claim 1, characterized in that: The S7 comprises the following steps: S71, detecting the storage status of the recovery chamber in real time through a pressure sensor, and triggering the collection device to automatically shut down when the storage volume reaches a threshold; S72, recording the current robot position information, and generating a return path in combination with navigation automatic memory technology; S73, through the power protection technology, dynamically adjust the movement speed and motor load of the robot return path; S74, monitoring the path deviation between the current position and the initial position of the robot in real time, and performing path correction; S75. Record the environmental data and task status collected during the return process to provide reference data for subsequent tasks.

8. The robot for automatically cleaning water surface garbage according to claim 1, characterized in that: The S8 comprises the following steps: S81, integrating the path data, garbage collection data and storage status data recorded during the task execution process to generate a task operation report; S82. Generate a garbage distribution heat map based on the garbage distribution status data, and synchronize the operation report and the heat map to the user terminal; S83, displaying the operation report on the user terminal, including the cleaning path, garbage distribution status, task execution time and recycling cabin storage status information; S84. After completing the data feedback, the robot enters the standby state and maintains the online connection of the remote monitoring module, waiting for the user to issue new task instructions.

9. A robot system for automatically cleaning up garbage on the water surface, characterized in that: Includes the following modules: Environmental data acquisition module, used to collect surface garbage distribution data and obstacle location information through active optical detection technology and CMOS photosensitive devices, and convert scene information into electrical signals; The path planning module is used to perform optimal path planning based on the A* path planning algorithm combined with the GBFS algorithm and the Dijkstra algorithm, generate a set of candidate paths, and select the path with the minimum total cost as the optimal path through a dynamic weight adjustment mechanism; The path optimization module is used to input the candidate path set into the particle swarm hybrid Poisson jump algorithm, perturb the path through the Poisson jump strategy, screen the path with the best fitness and output the optimal path result; The control execution module includes an STM32 single-chip microcomputer control system and a PWM motor drive module, which is used to transmit the optimal path data to the robot drive system, dynamically adjust the speed difference between the left and right motors, realize the robot's path movement and direction steering, and provide the required power for operation through solar power supply technology; The obstacle avoidance module is used to monitor the obstacles in front of the robot in real time through the ultrasonic obstacle avoidance system. When an obstacle is detected, the robot path direction is dynamically adjusted and the path is corrected according to the actual environment. The garbage collection module includes a double roller adsorption and separation system, which is used to collect garbage and oil on the water surface, store the collected materials in the built-in recovery tank, and conduct preliminary purification of the water through the filtration device; The pressure detection and navigation module includes a pressure sensor and a navigation automatic memory and power protection device, which is used to detect the storage volume of the recovery chamber, trigger the collection device to automatically shut down when the storage volume reaches a threshold, and guide the robot to return to the initial position along the return path; Feedback and monitoring module, including remote monitoring module and user terminal, is used to generate cleaning path, garbage distribution heat map and collection status operation report, and feed back the operation report to the user terminal to realize visual monitoring of cleaning tasks; The power supply module is used to provide power for system operation through solar power supply technology and monitor the power supply status in real time.

Images