Garbage feeding, bag breaking and paving system and method based on artificial intelligence bionic manipulator

Abstract

The invention discloses a garbage feeding, bag breaking and paving system and method based on an artificial intelligence bionic mechanical arm. The garbage feeding, bag breaking and paving system comprises a feeding device, a paving device and an intelligent recognition and path optimization device. The feeding device comprises a conical hopper, a truss located above the conical hopper and the mechanical arm in sliding connection with the truss through a bendable and telescopic arm. The paving device comprises a paving platform, a conveying transmission structure and a vibrating screen device; the intelligent identification and path optimization device comprises an image acquisition device and an AI intelligent control module electrically connected with the image acquisition device, the manipulator comprises a bionic structure with two-hand equivalent fingers and capable of crossing the corresponding fingers of the two hands, a force feedback sensor is mounted on the surface of each finger to monitor the tension of the garbage bag in real time, and each finger is of a multi-joint structure. Therefore, bionic actions are carried out. According to the feeding, bag breaking and paving integrated device, traditional multiple-device dispersing and independent treatment of a garbage pit, a chain scraper conveyor, manual auxiliary paving and a bag breaking machine is replaced, and the feeding, bag breaking and paving integrated device is created for the first time.

Description

Technical Field

[0001] The present invention relates to the field of solid waste resource treatment technology, in particular to a bionic manipulator based on artificial intelligence control for automatic garbage loading, intelligent bag breaking and paving integrated processing system and method. That is, the garbage loading, bag breaking and paving system and method based on artificial intelligence bionic manipulator can be widely used in the fields of garbage classification, resource recovery, garbage incineration pretreatment, etc. Background Art

[0002] As the amount of urban solid waste continues to increase, the way of waste treatment is gradually changing from "end-of-line incineration and landfill" to "front-end resource utilization and reduction". A large amount of mixed waste is directly sent to the incineration system without effective pretreatment, which not only reduces the recycling efficiency of recyclables, but also increases the processing cost and carbon emission load.

[0003] In the waste resource processing chain, "front-end loading → bag breaking → paving" is the pre-processing link of traditional waste treatment. As a basic operation unit, the processing efficiency, degree of automation and environmental adaptability directly affect the operation quality of the entire line. However, the current industry still widely adopts the combination of "manual feeding + bag breaking machine + chain conveyor belt paving", which has the following significant shortcomings:

[0004] 1. Dispersed processes and low collaboration efficiency

[0005] Traditional systems rely on serial equipment for segmented operations (such as manual loading, gear bag breaking, chain conveyor paving). The equipment occupies a large area, has many maintenance points, and cannot be coordinated and adjusted in real time. It is easy to cause problems such as stacking, bag jamming, and spilling, causing process bottlenecks. These decentralized processes also have the following problems:

[0006] 1) Problems with chain conveyors: high energy consumption and high operating costs; easy to get clogged, garbage bags pile up on the chain conveyor, affecting processing efficiency; manual paving is labor-intensive.

[0007] 2) Problems with shear or drum bag breaking machines: The bag breaking method is single, some high-value objects cannot be identified and broken, and some garbage bags cannot be completely broken, that is, the characteristics of the garbage bags cannot be intelligently identified, and the bag breaking method is not accurate; high requirements for uniform feeding, easy to accumulate and jam; poor adaptability, and uneven bag breaking effects for different plastic garbage bags.

[0008] 3) Problems with manually assisted paving: The working environment is harsh, and workers need to work in a smelly and dusty environment; the paving is uneven, affecting the accuracy of subsequent screening and resource recovery; more manual assistance is required for paving; the machine is prone to jamming, seriously affecting production efficiency.

[0009] 4) Hopper problems: Traditional hoppers cannot automatically complete automatic feeding. Manual intervention is required continuously to push the garbage onto the chain conveyor, and stacking and pressing are not allowed to prevent blockage of the chain conveyor and affect the process operation.

[0010] 2. High degree of manual dependence and harsh working environment

[0011] In the scenarios of wet garbage and mixed domestic garbage treatment, due to the complex material composition and various garbage bag structures, most of the current bag-breaking and spreading actions need to be assisted by manual labor, with high labor intensity, high health risks, and difficult to guarantee stability.

[0012] 3. Unable to accurately adapt to the rhythm of the sorting system

[0013] The back-end optical sorting, air sorting, and intelligent sorting equipment have high requirements for the thickness of the material layer and the uniformity of particle size distribution. However, the traditional spreading system is difficult to achieve precise regulation, often resulting in an increase in the misselection rate and a decrease in the sorting efficiency.

[0014] 4. Low level of intelligence and lack of flexible control ability

[0015] At present, the AI application in the feeding and bag-breaking processes in the industry is still in its infancy, and a control mechanism for sustainable learning and adaptive material differences has not been formed. The equipment grasping strategy is rigid and the path planning is single, and it is easy to make mistakes or stop when facing situations such as highly mixed and compacted garbage.

[0016] To solve the above problems, the market urgently needs a comprehensive equipment with the ability of "bionic intelligent grasping + path self-adaptation + uniform spreading control", which can complete automatic feeding, precise bag-breaking and efficient spreading operations in a closed and clean environment, and provide a stable and reliable front-end processing basis for subsequent resource extraction and intelligent sorting systems.

[0017] In some existing automated systems for feeding, bag-breaking and spreading, although there are designed manipulators, there is no integrated equipment for feeding, bag-breaking and spreading, and it is even more impossible to intelligently adapt to the shape of garbage bags, resulting in grasping failure or damage to garbage bags, and it is also impossible to identify and remove foreign objects during grasping.

[0018] With the increase in the cost of garbage incineration treatment, the garbage pretreatment system has become a key link in intelligent solid waste sorting. The existing system depends on manual labor or multi-equipment series connection during the bag-breaking and spreading processes, with high energy consumption, heavy pollution and low efficiency, and it is difficult to meet the industrial requirements of resource classification. There is an urgent need for new technologies that are efficient, concise, pollution-free and can meet resource classification and can subvert the existing treatment methods. Summary of the Invention

[0019] The present invention solves the technical problems that in the processes of feeding, bag breaking, and spreading of existing garbage, due to scattered processes, low collaborative efficiency, high energy consumption, high cost, easy blockage, high dependence on labor, poor working environment, poor adaptability to bag breaking, inability to accurately adapt to the rhythm of the sorting system, uneven spreading, machine jamming and other various defects, it is impossible to better realize automation. It provides a garbage feeding, bag breaking and spreading system and method based on an artificial intelligence bionic manipulator, effectively breaking through the technical bottleneck of the traditional pretreatment link, and providing new equipment support and an intelligent path for the urban solid waste resource utilization industry. It can well overcome the above defects, realize the automation of feeding, bag breaking and spreading without on-site manual intervention, and can complete the automated feeding, accurate bag breaking and efficient spreading operations in a closed and clean environment, providing a stable and reliable front-end processing basis for subsequent resource extraction and intelligent sorting systems. It improves the intelligent and automated level of garbage treatment, reduces labor costs, and improves the accuracy of garbage classification and the efficiency of resource recovery. The technical solution of the present invention is as follows:

[0020] A garbage feeding, bag breaking and spreading system applying an artificial intelligence bionic manipulator, comprising:

[0021] A feeding device: comprising a conical hopper with a relatively large top opening for dumping garbage, a truss located above the conical hopper, and a first manipulator slidably connected to the truss through a bendable and telescopic arm;

[0022] A spreading device: comprising an operation spreading platform located outside the conical hopper, a conveying drive structure and a vibrating screen device respectively connected to the operation spreading platform. The truss extends above the operation spreading platform. The operation spreading platform comprises an operation table and a spreading surface on the upper part of the operation table. The spreading surface is a wide conveyor belt for garbage conveyance. The conveying drive structure drives the spreading surface to move horizontally, and the vibrating screen device drives the spreading surface to vibrate. Both the conveying drive structure and the vibrating screen device are located below the spreading surface;

[0023] A second manipulator arranged on the side of the operation table of the operation spreading platform;

[0024] Intelligent recognition and path optimization device, including an image acquisition device and an AI intelligent control module electrically connected to the image acquisition device. The AI intelligent control module includes an identification and analysis module, a path and action optimization module connected to the identification and analysis module, a manipulator control module connected to the path and action optimization module, and a paving surface movement control module electrically connected to the identification and analysis module. The image acquisition device includes a camera, a 3D lidar, and / or a near-infrared detector. The camera, 3D lidar, and / or near-infrared detector are arranged above the conical hopper and the operation paving platform, and transmit the three-dimensional information of the garbage in the conical hopper and on the operation paving platform, including the garbage accumulation situation, garbage type, etc., and the action information of the first and second manipulators to the AI intelligent control module. The identification and analysis module includes an AI vision processing unit and an intelligent analysis unit. The AI vision processing unit receives, identifies, and processes the three-dimensional information and the action information in real time, and transmits the processing result to the intelligent analysis unit. The intelligent analysis unit conducts intelligent analysis on the garbage and the actions of the first and second manipulators. The intelligent analysis includes discovering abnormalities such as foreign objects, large items, and working errors according to the three-dimensional information of the garbage, extracting the material information, volume information, density information, and other garbage bag characteristic information of the garbage bag, and sending the analysis result to the path and action optimization module. The path and action optimization module includes a deep learning unit, which calculates through deep learning the grasping point, grasping force, grasping method, clamping force, movement action on the truss, movement distance, bag-breaking method, bag-breaking action, and optimized path of the first manipulator, and the tearing, sorting, and paving actions of the second manipulator, and transmits them to the manipulator control module to control the grasping and clamping, clamping movement, and bag-breaking actions of the first manipulator, and the tearing, sorting, and paving actions of the second manipulator. The paving surface movement control module controls the movement of the paving surface to convey the paved garbage according to the garbage condition on the operation paving platform analyzed by the identification and analysis module, and also optimizes the manipulator path through an intelligent obstacle avoidance algorithm to avoid collisions between manipulators.

[0025] Both the first manipulator and the second manipulator include a bionic structure with two hands having equivalent fingers and the fingers of the two hands being able to cross. The two hands of the first manipulator are also respectively connected to arms that can be bent, telescoped, and moved on the truss. Force feedback sensors are installed on the finger surfaces to monitor the tension of the garbage bag in real time and transmit signals to the motion and rotation control module. Each finger is a multi-joint structure, and multi-joint motion and rotation control structures are attached to each finger and the arm to control the bionic motions of the fingers and the arm. The motion and rotation control module is integrated with the motion and rotation control module; the motion and rotation control module is respectively connected to the recognition and analysis module and the manipulator control module, transmits the sensing signal to the recognition and analysis module, and the manipulator control module transmits the grasping motion and the optimized path control instructions optimized by the path and motion optimization module to the motion and rotation control module to control the multi-joint motion and rotation control structure to operate the motions of the first and second manipulators.

[0026] The path and optimization module further includes an intelligent grasping mode unit that stores different grasping modes of the first manipulator and determines which intelligent grasping mode the first manipulator adopts according to the recognition and analysis of the garbage by the recognition and analysis module. The intelligent grasping modes include a clamping mode for rigid garbage bags, hard plastic bags, etc.; a flexible grasping mode for preventing damage to kitchen waste garbage bags; an adsorption grasping mode for light garbage, paper, plastic bags, polystyrene foam, etc.; a high-frequency vibration blade is equipped at the end of the first manipulator finger, the multi-joint motion and rotation control structure includes a vibration drive structure, and the motion and rotation control module includes vibration drive control for the vibration drive structure; the path and optimization module further includes a bag-breaking method unit that stores different bag-breaking methods of the first manipulator and determines which bag-breaking method the first manipulator adopts according to the recognition and analysis of the garbage by the recognition and analysis module. The manipulator control module transmits the bag-breaking method control instructions optimized by the path and motion optimization module to the motion and rotation control module to control the bag-breaking motion of the multi-joint motion and rotation control structure. The bag-breaking methods include tearing bag-breaking, high-frequency vibration bag-breaking, and twisting and pulling bag-breaking.

[0027] The cross-section of the conical hopper is a polyhedron, circle or ellipse, and the angle of the cone ranges from 15° to 90° with the vertical line. Its inner side wall is coated with wear-resistant and anti-corrosion materials, and the truss spans the side wall area of the conical hopper, that is, the second side wall is lower than other side wall parts, namely the first side wall; it also includes a third manipulator, which is arranged inside the conical hopper to remove abnormal objects such as foreign objects and large pieces. The third manipulator includes a bionic structure with equal fingers on both hands and the corresponding fingers on both hands can cross. Force feedback sensors are installed on the surface of the fingers. Each finger is a multi-joint structure, and each finger is attached with a multi-joint movement and rotation control structure to control the bionic movement of the fingers. An action and rotation control module is integrated on the multi-joint movement and rotation control structure. The force feedback sensor monitors the tension of the garbage bag in real time and transmits the sensing signal to the recognition and analysis module. The action and rotation control module is connected to the recognition and analysis module and the manipulator control module. The manipulator control module transmits the action and optimized path control instructions optimized by the path and action optimization module to the action and rotation control module to control the multi-joint movement and rotation control structure to operate the third manipulator to remove foreign objects and large pieces.

[0028] A fourth manipulator is also arranged on the truss on the operation paving platform. The fourth manipulator includes a bionic structure with equal fingers on both hands and the corresponding fingers on both hands can cross. Force feedback sensors are installed on the surface of the fingers. Each finger is a multi-joint structure, and each finger is attached with a multi-joint movement and rotation control structure to control the bionic movement of the fingers. An action and rotation control module is integrated on the multi-joint movement and rotation control structure. The force feedback sensor monitors the garbage tension in real time and transmits the sensing signal to the recognition and analysis module. The action and rotation control module are respectively connected to the recognition and analysis module and the manipulator control module. The manipulator control module transmits the flicking action and optimized path control instructions optimized by the path and action optimization module to the action and rotation control module to control the multi-joint movement and rotation control structure to operate the flicking action of the fourth manipulator.

[0029] The path and optimization module also includes an intelligent flipping and paving unit, which stores the flipping and paving mode of the first manipulator and starts the flipping and paving of the first manipulator according to the recognition and analysis of the bag-breaking state of the garbage bag by the recognition and analysis module. The manipulator control module transmits the flipping and paving control instructions issued by the path and action optimization module to the action and rotation control module, and controls the first manipulator to flip 180° after bag-breaking through the multi-joint movement and rotation control structure, and evenly spread the garbage on the conveyor belt.

[0030] The first, second, third, and fourth manipulators are each one or more. The path and optimization module further includes a manipulator coordination unit that coordinates the grasping, clamping, moving, bag-breaking, tearing, sorting, spreading, rejecting, and flicking actions of the manipulators. The manipulator coordination unit uses an intelligent reinforcement learning (RL) load balancing algorithm to dynamically allocate grasping tasks according to the garbage flow and optimize the throughput. The manipulator coordination unit further includes a synchronous cooperation mode, which includes multiple first manipulators pulling and breaking the bag simultaneously and different manipulators being responsible for different tasks to cooperate synchronously. The toughness of the garbage bag is detected by the force feedback sensor, and the optimal bag-breaking mode is automatically selected.

[0031] The vibration frequency of the high-frequency vibration blade is 30 - 50 Hz, and the vibration sieve device uses a vibration frequency of 50 - 200 Hz. The feeding device, spreading device, second manipulator, and intelligent recognition and path optimization device are arranged in a closed chamber and are located in the same space.

[0032] A garbage leachate collection system is connected to the bottom of the garbage conical hopper. The garbage leachate collection system includes a diversion slope, a drainage channel, a filtration system, and an automatic extraction system. The diversion slope is a diversion slope that forms an angle of 2° - 15° with the bottom of the conical hopper. The diversion slope is connected to the drainage channel. The drainage channel is an anti-blocking spiral sewage discharge pipe. A metal grid is installed at the bottom of the conical hopper, and a polymer permeable filter layer is coated on the metal grid. The automatic extraction system includes an intelligent liquid level sensor arranged at the lower part of the conical hopper, a extraction pump communicated with the drainage channel, and a sewage treatment system connected to the end of the drainage channel. The leachate flows to the drainage channel through the diversion slope.

[0033] The AI intelligent control manipulator garbage feeding, bag-breaking, and spreading method includes:

[0034] S1. Intelligent feeding step: Use AI vision to automatically identify the three-dimensional information of the garbage poured into the conical hopper, including the stacking situation, garbage type, and garbage bag characteristic information such as the material information, volume information, and density information of the garbage bag. S2. High-efficiency grasping and bag-breaking step of the manipulator: The AI artificial intelligence automatically adjusts the bionic structure manipulator's bionic hands with corresponding fingers crossed to grasp the garbage bag according to the three-dimensional information of the garbage, and moves the garbage bag to the conveyor belt and then breaks the bag. The manipulator includes a bionic structure with equal fingers on both hands and the corresponding fingers on both hands can cross. Force feedback sensors are installed on the surface of each finger to monitor the tension of the garbage bag in real time and transmit signals to the AI artificial intelligence to adjust the grasping and bag-breaking methods and strategies in real time.

[0035] S3. Uniform spreading step of the manipulator: Use the bionic structure manipulator to uniformly spread the broken garbage bag.

[0036] The cross-section of the conical hopper described in S1 is polygonal, circular or elliptical, and the angle range of the cone is 15°-90° with the vertical line. Its inner wall is coated with wear-resistant and anti-corrosion materials;

[0037] The manipulator described in S2 is the first manipulator. The two hands of the first manipulator also include flexible and telescopic arms respectively connected to the two hands. The other ends of the arms are controllably movably connected to a truss arranged on the conical hopper. The first manipulators work together in multiple units to optimize task allocation; A third manipulator is also provided inside the conical hopper to handle abnormalities such as foreign objects, large items, and working errors found through AI intelligent analysis;

[0038] The manipulator described in S3 is the second manipulator. The conveyor belt is a wide paving surface arranged on the operation paving platform outside the conical hopper. The truss extends above the paving surface. The second manipulators work together in multiple units to optimize task allocation;

[0039] The paving described in S3 also includes automatically adjusting the angle and force of the bionic structure first manipulator to feed materials to the conveyor belt after bag breaking through AI artificial intelligence, as well as the coordinated operation of multiple first and second manipulators;

[0040] The grasping in S2 includes the following steps:

[0041] S21. Adjust the grasping order of the manipulator through the volume information and density information in the garbage bag feature information;

[0042] S22. Adjust the grasping mode of the manipulator through the material information in the garbage bag feature information. The clamping mode is for rigid garbage bags, hard plastic bags, etc.; The flexible grasping mode is for kitchen waste garbage bags to prevent breakage; The adsorption grasping mode is for light garbage, paper, plastic bags, expanded polystyrene;

[0043] S23. Calculate the optimal grasping points and paths through deep learning (CNN+Transformer), and optimize the manipulator path through an intelligent obstacle avoidance algorithm to avoid collisions between manipulators;

[0044] S24. Grasp the garbage in sequence according to the grasping order, grasping mode, optimal grasping points and optimal grasping paths;

[0045] S25. Judge the toughness of the garbage bag according to the pressure sensor and compare it with the preset toughness value;

[0046] The bag breaking in S2 includes the following steps:

[0047] S26. Adopt corresponding bag-breaking schemes according to the toughness of the garbage bag. When the pressure sensor determines that the toughness of the garbage bag is greater than the preset toughness value, high-frequency vibration bag-breaking is adopted, and micro-vibration cutting is carried out; when the pressure sensor determines that the toughness of the garbage bag is less than the preset toughness value, grasp both ends of the garbage bag with both hands and rotate them in opposite directions to pull and tear, forming torsional tearing to complete the bag-breaking operation;

[0048] The paving in S3 also includes the following steps:

[0049] S31. Turn the bag-broken garbage bag over and pave it on the conveyor belt;

[0050] S32. Automatically adjust the feeding angle and strength for uniform paving, prevent garbage accumulation, and improve the screening efficiency;

[0051] S33. Vibrate the conveyor belt at a vibration frequency of 50 - 200 Hz to prevent the garbage from forming clusters and improve the paving uniformity;

[0052] S34. Sort the recyclable garbage after paving; it also includes the following steps:

[0053] S4: Optimize the task allocation of multiple manipulators through an intelligent scheduling algorithm to improve the garbage processing throughput;

[0054] S5. Provide a slightly inclined (2° - 15°) leachate diversion slope at the bottom of the conical hopper to direct the leachate to the designated drainage channel; install a metal grid + polymer permeable filter layer at the bottom of the conical hopper to intercept solid garbage; adopt an intelligent liquid level sensor, and when the leachate reaches the set value, automatically start the extraction pump to discharge the leachate into the sewage treatment system; the leachate pipeline adopts an anti-blocking spiral sewage discharge pipe.

[0055] The technical effects of the present invention:

[0056] The method of the present invention sets a bionic-structured manipulator, equipped with AI vision to automatically recognize the three-dimensional information of the dumped garbage. Using AI artificial intelligence, according to the three-dimensional information of the garbage, it automatically adjusts the fingers of the two hands of the manipulator to cross-grab the garbage bag in the garbage pile, move it to the conveyor belt for bag-breaking and uniform spreading. That is, through artificial intelligence control of the bionic manipulator, it can fully automate feeding, bag-breaking, and spreading, replacing the mode of the traditional garbage feeding system where a garbage pit + chain conveyor + manual auxiliary spreading + bag-breaking machine work independently in sequence to process garbage. It pioneered an integrated device for feeding, bag-breaking, and spreading, eliminating the bag-breaking machine and manual labor in all processes, greatly reducing energy consumption and costs. By adjusting the operation of the manipulator through artificial intelligence, various defects such as easy blockage of the conveyor belt, poor adaptability and accuracy of bag-breaking, uneven spreading, and machine jamming are eliminated, realizing automation of feeding, bag-breaking, and spreading. It can complete automated feeding, precise bag-breaking, and efficient spreading operations in a closed and clean environment, providing a stable and reliable front-end processing basis for subsequent resource extraction and intelligent sorting systems, improving the intelligent and automated level of garbage treatment, reducing labor costs, and enhancing the accuracy of garbage classification and resource recovery efficiency.The system of the present invention includes a feeding device, a paving device, and an intelligent recognition and path optimization device. The feeding device comprises three components: a conical hopper, a truss, and a first manipulator. The respective structures, positional relationships, and connection relationships enable the bendable and extendable first manipulator to reach into the conical hopper to grasp the garbage dumped into it and move it out of the conical hopper through the truss. The top opening of the conical hopper is relatively large. Firstly, it shows a conical shape with a large top and a small bottom. Secondly, the garbage truck dumps the garbage from top to bottom, so the opening is large, which is convenient for the first manipulator to drop from the truss into the conical hopper to grasp the garbage bags dumped inside. The paving device includes an operation paving platform, a conveying and driving structure, and a vibrating screen device for paving the garbage. The truss extends above the operation paving platform, so that the first manipulator can clamp the garbage bag above the operation paving platform for bag breaking. A second manipulator is arranged on the side of the operation table of the operation paving platform for paving operation. The intelligent recognition and path optimization device includes an image acquisition device and an AI intelligent control module connected to the image acquisition device to transmit data and information. The image acquisition device transmits the three-dimensional information of the garbage in the conical hopper and on the operation paving platform, including the garbage accumulation situation, garbage type, etc., and the action information of the first and second manipulators to the AI intelligent control module. Each module of the AI intelligent control module respectively receives, recognizes, processes, analyzes, and controls in real time. The analysis includes deep learning to calculate the grasping point, grasping force, grasping method, clamping force, movement action on the truss, movement distance, bag breaking method, bag breaking action, and optimized path of the first manipulator, as well as the tearing, sorting, and paving actions of the second manipulator, so as to control the relevant actions of the first and second manipulators, the movement of the paved garbage on the paving surface, and the collision avoidance measures of the manipulators. The first and second manipulators are simulation manipulators specially designed for the system. Their finger structures and module settings can be controlled by the above-mentioned AI intelligent control module to imitate the grasping, holding, pulling and tearing with both hands, tearing, sorting, paving, bending, and extending actions of human hands, so as to cooperate with other components to realize the highly integrated operations of manipulator grasping, intelligent bag breaking, and dynamic paving, with a high degree of integration. The system of the present invention can replace 4 traditional devices, reduce more than 80% of manual intervention, and increase the daily garbage processing capacity of a single line by 50%.

[0057] Dumping garbage into the conical hopper, the cone can utilize the downward movement trend formed by the self-weight of the garbage for feeding control, with low energy consumption and reduced operating costs. Moreover, different cone angles form different movement trends, which can be targeted at different types of garbage; accurately grasping the garbage bag by the manipulator and moving it to the conveyor belt, and intelligent control of bag breaking and paving can solve the problem of accumulation on the operation paving platform, improve the processing efficiency, and save a large amount of energy and manpower;

[0058] Through an image acquisition device and an AI intelligent control module electrically connected to the image acquisition device, AI vision analysis and path optimization are realized. It intelligently analyzes and discovers foreign objects, large objects, working errors, and extracts the material information, volume information, and density information of garbage bags, and sends the analysis results to the path and motion optimization module. Then, through deep learning calculation, the grasping points and optimized paths of the manipulator are obtained, realizing the optimization of the grasping order of garbage bags, controlling the actions of the manipulator to automatically identify the type of garbage bag, the best grasping points, and optimizing the motion path of the manipulator; optimizing the manipulator path through an intelligent obstacle avoidance algorithm to avoid collisions between manipulators.

[0059] Intelligent manipulator grasping + dynamic force feedback: Adapt to different garbage bags to prevent the bag body from bursting.

[0060] Other efficient bag breaking: According to the force feedback sensor to detect the toughness of the garbage bag, combined with high-frequency vibration bag breaking and torsional shear bag breaking, automatically select the optimal bag breaking mode.

[0061] Intelligent paving optimization: Realize the reliability of the paving work through flipping paving, and improve the uniformity of spreading through a vibrating screen.

[0062] The present invention has achieved significant improvements in the following core dimensions:

[0063] 1. Improvement in processing efficiency: The manual assistance processing capacity of each production line of the traditional system is about 1.5 tons per hour, while the system of the present invention realizes the improvement of the single-line processing capacity to 3 tons per hour through the collaborative operation of multiple manipulators, intelligent path planning, and paving depth control, with an efficiency improvement of up to 100%.

[0064] 2. Enhancement of automation and intelligence levels: The artificial intelligence of this system can adopt deep neural network and image fusion algorithms to realize the automatic recognition, path generation, and dynamic grasping of garbage bags of different specifications and forms, and at the same time has the ability of task self-allocation and trajectory correction, significantly reducing manual intervention and having the ability to operate continuously for 24 hours.

[0065] 3. Strong environmental adaptability: Due to the absence of manual operation, a closed operation cabin design can be adopted, which can be deployed in high-humidity, odor, and pest environments, avoiding the problem of pollution diffusion caused by traditional open operations, and better meeting the requirements of high-standard environmental protection stations.

[0066] 4. Excellent control of paving uniformity: Through the operation of bionic manipulators, the dynamic simulation compensation algorithm can be used to control the thickness of the material layer, so that the coefficient of variation CV value of the paving thickness is controlled within 10%, meeting the process requirements of subsequent air separation, optical sorting, robot sorting, etc.

[0067] 5. Strong coordination ability of the subsequent sorting system: The paving path and speed are adjusted in real time by the AI control system, which can be linked with technologies such as photoelectric recognition, near-infrared sorting, and air flow sorting to achieve closed-loop optimization of the overall waste resource system.

[0068] 6. High system integration, more optimized floor area and operation and maintenance costs: The present invention integrates the three major processes of feeding, bag breaking, and paving into a single control architecture and structural unit, reducing the space occupation by about 30% and the energy consumption by 40% compared with traditional four-stage equipment. There are fewer later maintenance nodes and higher system operation stability. Brief Description of the Drawings

[0069] Figure 1 It is a structural schematic diagram of the waste feeding, bag breaking and paving system based on the artificial intelligence bionic manipulator of the present invention;

[0070] Figure 2 It is a flowchart of the waste feeding, bag breaking and paving method of the AI intelligent control manipulator of the present invention;

[0071] Figure 3 It is a flowchart of the grasping method in the waste feeding, bag breaking and paving method of the AI intelligent control manipulator of the present invention;

[0072] Figure 4 It is a flowchart of the paving method in the waste feeding, bag breaking and paving method provided by the present invention for the AI intelligent control manipulator;

[0073] Figure 5 It is a complete flowchart of the AI intelligent waste feeding, bag breaking and paving method provided by the present invention;

[0074] Figure 6 It is a structural diagram of the waste feeding, bag breaking and paving system based on the artificial intelligence bionic manipulator provided by the present invention;

[0075] Figure 7 It is a block diagram of the AI intelligent control module of the waste feeding, bag breaking and paving system based on the artificial intelligence bionic manipulator provided by the present invention.

[0076] Figure 1 The reference numerals in the figures are as follows:

[0077] 11, conical hopper; 12, truss; 13, first manipulator; 21, operation paving platform; 22, conveying drive structure; 23, vibrating screen device; 3, second manipulator; 4, intelligent recognition and path optimization device; 5, third manipulator; 6, fourth manipulator; 71, diversion slope; 72, drain channel; 73, filtration system; 74, intelligent liquid level sensor; 75, extraction pump. Detailed Embodiments

[0078] The following describes the implementation manners of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific implementation manners. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0079] The following will further elaborate on the present invention Figure 1-7 in conjunction with the accompanying drawings.

[0080] An embodiment of the present invention discloses a garbage feeding, bag-breaking and spreading system and method applying an artificial intelligence bionic manipulator.

[0081] Referring to Figure 1 、 Figure 6 and Figure 7 , a garbage feeding, bag-breaking and spreading system applying an artificial intelligence bionic manipulator includes:

[0082] A feeding device, including a conical hopper 11 with a relatively large opening at the top for dumping garbage, a truss 12 located above the conical hopper 11, and a first manipulator 13 slidably connected to the truss 12 through a bendable and telescopic arm;

[0083] A spreading device, including an operation spreading platform 21 located outside the conical hopper 11, a conveying and driving structure 22 and a vibrating screen device 23 respectively connected to the operation spreading platform 21. The truss 12 extends above the operation spreading platform 21. The operation spreading platform 21 includes an operation table and a spreading surface on the upper part of the operation table. The spreading surface is a wide conveyor belt for garbage transmission. The conveying and driving structure 22 drives the spreading surface to move horizontally, and the vibrating screen device 23 drives the spreading surface to vibrate. Both the conveying and driving structure 22 and the vibrating screen device 23 are located below the spreading surface;

[0084] A second manipulator 3 provided on the side of the operation table of the operation spreading platform 21;

[0085] An intelligent recognition and path optimization device 4, including an image acquisition device (see Figure 6 ), and an AI intelligent control module electrically connected to the image acquisition device. The AI intelligent control module (see Figure 7)It includes an identification and analysis module, a path and action optimization module connected to the identification and analysis module, a manipulator control module connected to the path and action optimization module, and a paving surface movement control module electrically connected to the identification and analysis module. The image acquisition device includes a camera, a 3D lidar, and / or a near-infrared detector. The camera, 3D lidar, and / or near-infrared detector are arranged above the conical hopper and the working paving platform 21, and transmit the three-dimensional information of the garbage in the conical hopper and on the working paving platform 21, including the garbage accumulation situation, garbage type, etc., and the action information of the first and second manipulators 3 to the AI intelligent control module. The identification and analysis module includes an AI vision processing unit and an intelligent analysis unit. The AI vision processing unit receives, identifies, and processes the three-dimensional information and the action information in real time, and transmits the processing result to the intelligent analysis unit. The intelligent analysis unit conducts intelligent analysis on the garbage and the actions of the first and second manipulators 3. The intelligent analysis includes discovering abnormalities such as foreign objects, large items, and working errors based on the three-dimensional information of the garbage, extracting the material information, volume information, density information, and other garbage bag characteristic information of the garbage bag, and sending the analysis result to the path and action optimization module. The path and action optimization module includes a deep learning unit, which predicts the tensile strength of the garbage bag through deep learning, selects an appropriate clamping force to prevent the garbage bag from bursting, dynamically adjusts the grasping sequence to improve the feeding efficiency, and preferentially grasps high-density garbage bags to prevent blockage. Specifically, it includes calculating the optimal grasping point, grasping force, grasping method, clamping force, grasping path, movement action on the truss 12, movement distance, bag-breaking method, bag-breaking action, and optimized path of the first manipulator 13, and the tearing, paving actions of the second manipulator 3, and transmitting them to the manipulator control module to control the automatic grasping, sorting, clamping movement, and bag-breaking actions of the first manipulator 13, and the tearing, sorting, and paving actions of the second manipulator 3. The paving surface movement control module controls the movement of the paving surface to convey the paved garbage according to the garbage condition on the working paving platform 21 analyzed by the identification and analysis module, and also optimizes the manipulator path through an intelligent obstacle avoidance algorithm to avoid collisions between manipulators.

[0086] The first manipulator 13 and the second manipulator 3 both include a bionic structure with double hands having equivalent fingers and the fingers of the double hands being crossable, preferably five fingers each. The double hands of the first manipulator 13 are also respectively connected to arms that can be bent, telescoped, and moved on the truss 12. Force feedback sensors are installed on the finger surfaces to monitor the tension of the garbage bag in real time and transmit signals to the motion and rotation control module. Each finger is a multi-joint structure, and each finger and the arm are attached with a multi-joint motion and rotation control structure to control the fingers and the arm to perform bionic motions. The motion and rotation control module is integrated with the motion and rotation control module; the motion and rotation control module is respectively connected to the recognition and analysis module and the manipulator control module, transmits the sensing signal to the recognition and analysis module, and the manipulator control module transmits the grasping motion and the optimized path control instruction optimized by the path and motion optimization module to the motion and rotation control module to control the multi-joint motion and rotation control structure to operate the motions of the first manipulator 13 and the second manipulator 3.

[0087] In one embodiment, the top of the conical hopper 11 is designed with a polyhedron, a circle, or an ellipse, and has a large opening for dumping garbage. The diameter of the top of the conical hopper 11 (conical barrel) can be 4 - 12 meters, the height of the cone can be 4 - 10 meters, the conical angle is 35° - 50°, and the width of the garbage outlet at the bottom is 1.5 - 3 m. The truss 12 is located above the conical hopper 11 and its span extends to the outside of the conical hopper 11. The first manipulator 13 is slidably installed thereon and provides enough grasping space for the first manipulator 13. Among them, the truss 12 is connected to the first manipulator 13 through an arm that can be bent, telescoped, and moved on the truss. The telescopic arm enables the manipulator to contact the bottom center of the conical hopper 11 to realize the grasping of the garbage bag. The bendable arm enables the manipulator to adjust the angle to grasp the target garbage bag and complete the feeding operation of grasping, clamping, and clamping movement;

[0088] The operation paving platform 21 is located outside the conical hopper 11. The truss 12 extends above the operation paving platform 21. The operation paving platform 21 is used to carry the garbage after the first manipulator 13 tears the bag. After the first manipulator 13 completes the grasping and clamping action, it moves on the truss 12 through the arm, reaches above the operation paving platform 21, and then performs the bag tearing operation. After the garbage bag is torn, the internal garbage freely falls to the operation paving platform 21 under the action of gravity. The paving surface on the operation paving platform 21 is a wide conveyor belt. The wide surface can spread more garbage horizontally, preventing garbage accumulation. The paving surface is a working surface controlled by the conveying drive structure 22 to move horizontally for garbage conveyance. The wide surface can also prevent garbage from scattering outside the area of the operation paving platform 21. The vibrating screen device 23 drives the paving surface to vibrate to avoid stacking of garbage and achieve a better paving effect. The vibrating screen device 23 is a vibration drive mechanism, and various vibration drive devices in the prior art can be used;

[0089] Figure 1 The second manipulator 3 described in is fixedly arranged at intervals on both sides of the operation table of the operation paving platform 21, on the one hand, used for sorting recyclable garbage, and on the other hand, used for tearing the garbage bag that is not completely torn to optimize the paving effect;

[0090] Figure 6 In, the intelligent recognition and path optimization device 4 includes an image acquisition device and an AI intelligent control module electrically connected to the image acquisition device. The image acquisition device includes a camera, a 3D lidar, and / or a near-infrared detector. The camera is used to capture 2D images through visible light and record visual information such as colors and textures. The 3D lidar is used to emit laser pulses, measure distances through the reflection time, generate high-precision 3D point clouds, and then obtain spatial coordinates and distances. The near-infrared detector is used to detect object characteristics such as thermal radiation and material reflection, and output grayscale images or specific spectral data. The camera captures 2D images, and the lidar and / or the near-infrared detector capture 3D images. Combining them can obtain more comprehensive image information. The camera, 3D lidar, and / or near-infrared detector are arranged above the conical hopper 11 and the operation paving platform 21, and transmit the three-dimensional information of the garbage in the conical hopper and on the operation paving platform 21, including garbage accumulation conditions and garbage types, and the action information of the first and second manipulators 13 and 3 to the AI intelligent control module.

[0091] Figure 7Among them, the AI intelligent control module includes an identification and analysis module, a path and action optimization module connected to the identification and analysis module, a manipulator control module connected to the path and action optimization module, and a paving surface movement control module electrically connected to the identification and analysis module. The identification and analysis module includes an AI vision processing unit and an intelligent analysis unit. The AI vision processing unit receives, identifies, and processes the three-dimensional information and the action information in real time, and transmits the processing results to the intelligent analysis unit. The intelligent analysis unit conducts intelligent analysis on the garbage and the actions of the first and second manipulators 13 and 3. The intelligent analysis includes discovering abnormalities such as foreign objects, large items, and working errors according to the three-dimensional information of the garbage, extracting the material information, volume information, density information, and other garbage bag characteristic information of the garbage bag, and sending the analysis results to the path and action optimization module. The path and action optimization module includes a deep learning unit, which calculates the best grasping point, grasping force, grasping method, grasping path, movement actions on the truss 12, movement distance, bag-breaking method, bag-breaking actions, and optimized path of the first manipulator 13 through deep learning, as well as the tearing and paving actions of the second manipulator 3, and transmits them to the manipulator control module to control the grasping and clamping, clamping movement, and bag-breaking actions of the first manipulator 13, and the tearing, sorting, and paving actions of the second manipulator 3. The paving surface movement control module controls the movement of the paving surface to convey the paved garbage according to the garbage condition on the operation paving platform 21 analyzed by the identification and analysis module. The garbage bag characteristic information is obtained by analyzing the collected image information, and then the actions, movement paths, operation modes, etc. of the corresponding manipulator are calculated through deep learning. Each module and unit are closely connected and cooperate with each other, and sequentially perform grasping and clamping, clamping movement, bag-breaking action sorting, and paving actions to achieve reliable feeding, sorting, and paving operations. The path of the manipulator is also optimized through an intelligent obstacle avoidance algorithm to avoid collisions between the manipulators, improve the fault tolerance of the system, and solve the problems in the prior art such as the accumulation of garbage bags affecting the processing efficiency, the single bag-breaking method, uneven bag-breaking effects on different plastic garbage bags, manual-assisted paving, high labor intensity, and harsh working environment.

[0092] Figure 1 Among them, both the first manipulator 13 and the second manipulator 3 include a bionic structure with dual-handed equal fingers and the corresponding fingers of the two hands can cross, preferably five fingers each. A force feedback sensor is installed on the finger surface to monitor the tension of the garbage bag in real time and transmit the signal to the action and rotation control module. Each finger is a multi-joint structure, and each finger and the arm are attached with a multi-joint action and rotation control structure to control the bionic actions of the fingers and the arm. The action and rotation control module is integrated on the multi-joint action and rotation control structure (see Figure 7);The motion and rotation control module is respectively connected to the recognition and analysis module and the manipulator control module, transmits the sensing signal to the recognition and analysis module, and the manipulator control module transmits the grasping motion and the optimized path control instructions optimized by the path and motion optimization module to the motion and rotation control module to control the operations of the first and second manipulators 13 and 3 of the multi-joint motion and rotation control structure, which can cope with various environments and improve the flexibility of the operation.

[0093] See Figure 1 , Figure 6 and Figure 7 , the path and optimization module further includes an intelligent grasping mode unit, which stores different grasping modes of the first manipulator 13, and determines that the first manipulator 13 adopts different intelligent grasping modes according to the recognition and analysis of the garbage by the recognition and analysis module. The intelligent grasping modes include clamping modes for rigid garbage bags, hard plastic bags, etc.; flexible grasping modes for preventing damage to kitchen waste garbage bags; adsorption grasping modes for light garbage, paper, plastic bags, expanded polystyrene, etc. The finger tip of the first manipulator 13 is equipped with a high-frequency vibration blade. The multi-joint motion and rotation control structure includes a vibration driving structure (see Figure 7 ), and the motion and rotation control module includes vibration driving control of the vibration driving structure. The path and optimization module further includes a bag-breaking method unit ( Figure 7 ), which stores different bag-breaking methods of the first manipulator 13, and determines that the first manipulator 13 adopts different bag-breaking methods according to the recognition and analysis of the garbage by the recognition and analysis module, improving the flexibility of the bag-breaking work. The manipulator control module transmits the bag-breaking method control instructions optimized by the path and motion optimization module to the motion and rotation control module to control the bag-breaking action of the multi-joint motion and rotation control structure. The bag-breaking methods include tearing bag-breaking, high-frequency vibration bag-breaking, and twisting and pulling bag-breaking.

[0094] In Figure 7 , the intelligent grasping mode unit stores different grasping modes of the first manipulator 13, and determines that the first manipulator 13 adopts different intelligent grasping modes according to the recognition and analysis of the garbage by the recognition and analysis module, improving the flexibility of the grasping work.

[0095] See Figure 1 , Figure 6 and Figure 7, the cross-section of the conical hopper 11 is polygonal, circular or elliptical, the angle range of the cone is 15° - 90° with the vertical line, its inner wall is coated with wear-resistant and anti-corrosive material, and the truss 12 spans the side wall area of the conical hopper, that is, the second side wall is lower than other side wall parts, namely the first side wall; it further includes a third manipulator 5, the third manipulator 5 is arranged inside the conical hopper to remove foreign objects, large pieces and other abnormalities, the third manipulator 5 includes a bionic structure with equal fingers on both hands and the corresponding fingers on both hands can cross, preferably five fingers on each hand, a force feedback sensor is installed on the finger surface, each finger is a multi-joint structure, and each finger is attached with a multi-joint movement and rotation control structure, see Figure 7 , to control the bionic movement of the fingers, the multi-joint movement and rotation control structure is integrated with an action and rotation control module, the force feedback sensor monitors the tension of the garbage bag in real time and transmits the sensing signal to the recognition and analysis module, the action and rotation control module is connected to the recognition and analysis module and the manipulator control module, and the manipulator control module transmits the action and optimized path control instructions optimized by the path and action optimization module to the action and rotation control module to control the multi-joint movement and rotation control structure to operate the third manipulator 5 to remove foreign objects and large pieces.

[0096] In Figure 1 , the cross-section of the conical hopper 11 is circular or elliptical to avoid dead angles at the grasping positions and improve the reliability of the system. The conical design with inclined side walls enables the garbage bag to freely roll into the bottom center of the conical hopper 11, the angle range of the cone is 15° - 90° with the vertical line, the inner wall is coated with wear-resistant and anti-corrosive material, on the one hand, it increases the durability, and on the other hand, it improves the smoothness of the inner wall. The third manipulator 5 is arranged inside the conical hopper to remove foreign objects, large pieces and other abnormalities. The starting sequence of the third manipulator 5 takes precedence over the first manipulator 13 and the second manipulator 3 to remove foreign objects and large pieces before the feeding operation, further improving the reliability of the system operation.

[0097] See Figure 1 , Figure 6 and Figure 7 , on the truss 12 on the operation paving platform 21, a fourth manipulator 6 is further provided. The fourth manipulator 6 includes a bionic structure with equal fingers on both hands and the corresponding fingers on both hands can cross, preferably five fingers on each hand, a force feedback sensor is installed on the finger surface, each finger is a multi-joint structure, and each finger is attached with a multi-joint movement and rotation control structure, see Figure 7, to control the fingers to perform bionic actions. The multi-joint action and rotation control structure is integrated with an action and rotation control module. The force feedback sensor monitors the garbage tension in real time and transmits the sensing signal to the recognition and analysis module. The action and rotation control module are respectively connected to the recognition and analysis module and the manipulator control module. The manipulator control module transmits the optimized flicking action and optimized path control instructions obtained by the path and action optimization module to the action and rotation control module to control the multi-joint action and rotation control structure to operate the flicking action of the fourth manipulator 6.

[0098] In Figure 1 , the flicking action of the fourth manipulator 6 can achieve fast sorting operations, which is suitable for recyclable garbage with small volume and light weight. The fourth manipulator 6 works synchronously with the second manipulator 3, taking into account the comprehensiveness and efficiency of the sorting work.

[0099] See Figure 1 、 Figure 6 and Figure 7 , the path and optimization module further includes an intelligent flipping and spreading unit. See Figure 7 , this unit stores the flipping and spreading mode of the first manipulator 13, and starts the flipping and spreading of the first manipulator 13 according to the recognition and analysis of the bag-breaking state of the garbage bag by the recognition and analysis module. The manipulator control module transmits the flipping and spreading control instructions issued by the path and action optimization module to the action and rotation control module, and controls the first manipulator 13 to flip 180° after bag-breaking through the multi-joint action and rotation control structure to evenly spread the garbage onto the conveyor belt.

[0100] In this embodiment, the intelligent flipping and spreading unit stores the flipping and spreading mode of the first manipulator 13, which is used to control the first manipulator 13 to flip 180° after bag-breaking to evenly spread the garbage onto the conveyor belt, further improving the evenness and comprehensiveness of the spreading.

[0101] Referring to Figure 1 、 Figure 6 and Figure 7 , each of the first, second, third, and fourth manipulators 6 is one or more. The path and optimization module further includes a manipulator coordination unit to coordinate the grasping, clamping, moving, bag-breaking, tearing, sorting, spreading, removing, and flicking actions of each manipulator; the manipulator coordination unit uses an intelligent reinforcement learning (RL) load balancing algorithm to dynamically allocate grasping tasks according to the garbage flow and optimize the throughput; the manipulator coordination unit further includes a synchronous cooperation mode, and the synchronous cooperation mode includes multiple first manipulators 13 pulling and breaking the bag simultaneously and different manipulators being responsible for different tasks to perform synchronous cooperation.

[0102] In this embodiment, the manipulator coordination unit adopts an intelligent reinforcement learning (RL) load balancing algorithm, which can perceive the changes in the system state in real time, dynamically allocate grasping tasks and adjust strategies according to the garbage flow, adapt to complex environments such as load fluctuations and heterogeneous resources, and has significant advantages. The A algorithm + Dijkstra algorithm can also be adopted to calculate the shortest path, reduce the movement route, and reduce the manipulator conflict.

[0103] See Figure 1 , Figure 6 and Figure 7 , the vibration frequency of the high-frequency vibration blade is 30 - 50 Hz, and the vibrating screen device 23 adopts a vibration frequency of 50 - 200 Hz.

[0104] The feeding device, paving device, each manipulator, and the intelligent recognition and path optimization device 4 of the present invention are all arranged in a closed chamber and located in the same space. This avoids environmental problems such as secondary dust, odor diffusion, and mosquito breeding, significantly improves the operation cleanliness and environmental protection standards, and meets high-demand scenarios at home and abroad (such as underground transfer, wet garbage treatment centers, etc.). The modular partition and function integration of the closed structure can form independent structural protection points. The vibration blade adopts a low-frequency vibration mode of 30 - 50 Hz to ensure the reliability of bag breaking with high impact force and adaptability; the low-frequency vibration has a longer period, allowing the blade to fully reset between two vibrations, ensuring the coherence and stability of the cutting action. Since the garbage bag materials are diverse, the low-frequency vibration can adapt to different toughness materials by adjusting the amplitude rather than the frequency, avoiding blade jamming or material rebound that may be caused by high-frequency vibration. The low-frequency vibration has a lower requirement for the driving motor power, reducing energy consumption; at the same time, the vibration inertia is smaller, reducing the fatigue loss of the structure and extending the service life;

[0105] The vibrating screen device 23 adopts a high-frequency vibration mode of 50 - 200 Hz. A screen is set under the paving surface, or the paving surface is the screen structure, and efficient material distribution and anti-blocking are achieved through refined vibration. The high-frequency vibration can enhance the jumping movement of garbage particles, promote the automatic stratification of materials with different densities, and improve the efficiency of subsequent sorting operations. The high-frequency vibration causes rapid and small-amplitude jitter on the surface of the screen, effectively preventing wet garbage or fibrous materials from adhering to the screen holes and reducing the manual cleaning frequency. The frequency design of both is based on physical characteristics and scenario requirements, taking into account efficiency, durability, and energy consumption control, forming a complete garbage pretreatment solution.

[0106] See Figure 1, a garbage leachate collection system is connected to the bottom of the garbage conical hopper 11. The garbage leachate collection system includes a diversion slope 71, a drainage channel 72, a filtration system 73 and an automatic extraction system. The diversion slope 71 is a diversion slope 71 that is inclined 2°-15° to the bottom of the conical hopper 11. The diversion slope 71 is connected to the drainage channel 72. The drainage channel 72 is an anti-clogging spiral sewage pipe. A metal grid is installed at the bottom of the conical hopper 11, and a polymer permeable filter layer is coated on the metal grid. The automatic extraction system includes an intelligent liquid level sensor 74 arranged at the lower part of the conical hopper 11, an extraction pump 75 communicated with the drainage channel 72, and a sewage treatment system connected to the end of the drainage channel 72. The leachate flows through the diversion slope 71 to the drainage channel 72.

[0107] In Figure 1 , a diversion slope 71 with an inclination of 2°-15° can be formed at the bottom of the conical hopper 11. 2°-15° belongs to a gentle slope, and the garbage bags will not change their stacking form because of this, but the liquid will have a downward movement trend to drain the garbage leachate into the drainage channel 72. The drainage channel 72 is an anti-clogging spiral sewage pipe to avoid blockage. A metal grid is installed at the bottom of the conical hopper 11, and a polymer permeable filter layer is coated on the metal grid to isolate solid particles. When the intelligent liquid level sensor 74 detects that the garbage leachate at the bottom of the conical hopper 11 is higher than the preset liquid level, the extraction pump 75 is controlled to start to realize the cleaning of the garbage leachate.

[0108] The intelligent recognition and path optimization device 4 of the present invention can adopt an independently trained intelligent path planning model, combine image recognition, grasping point prediction and behavior evaluation mechanisms to realize the autonomous classification judgment and dynamic paving strategy of irregular garbage materials. The AI algorithm not only ensures the paving uniformity and accurate order, but also has the functions of fault self-calibration and operation behavior optimization. It can form a protection for behavior control logic, a protection for parameter control range and a protection for the AI model training framework. Each bionic manipulator has the ability of multi-degree-of-freedom linkage, simulating actions such as human hand grasping, rotating, tearing, and unfolding; and is equipped with a force feedback and dynamic grasping adjustment module to ensure that the bag-breaking action is both efficient and avoids damaging the internal recyclables. With the dual advantages of structure and motion control, it has high stability and industrial implementation adaptability.

[0109] See Figure 2 , the AI intelligent control mechanical hand feeding and bag-breaking paving method includes:

[0110] S1. Intelligent feeding step: Use AI vision to automatically identify the three-dimensional information of the garbage dumped into the conical hopper, including the stacking situation, garbage type, and garbage bag characteristic information such as garbage bag material information, volume information, and density information; dynamically collect target garbage images through the conical hopper 11.

[0111] S2. High-efficiency grasping and bag-breaking steps of the manipulator: The AI artificial intelligence automatically adjusts the bionic structure manipulator with bionic two hands crossing their ten fingers to grasp the garbage bag according to the three-dimensional information of the garbage, and moves the garbage bag to the conveyor belt for bag-breaking. The manipulator includes a bionic structure with equal fingers on both hands and the corresponding fingers on both hands can cross, preferably five fingers on each hand. Force feedback sensors are installed on the surface of each finger to monitor the tension of the garbage bag in real time and transmit signals to the AI artificial intelligence to adjust the grasping and bag-breaking methods and strategies in real time;

[0112] S3. Uniform spreading step of the manipulator: Use the bionic structure manipulator to uniformly spread the garbage bag after bag-breaking.

[0113] The dumping in S1 means that the garbage truck dumps the garbage into the conical hopper 11 with a relatively large opening at the top. The cross-section of the conical hopper 11 is polygonal, circular or elliptical, and the angle range of the cone is 15° - 90° with the vertical line. Its inner wall is coated with wear-resistant and anti-corrosion materials;

[0114] The manipulator in S2 is the first manipulator 13. The two hands of the first manipulator 13 also include bendable and telescopic arms respectively connected to the two hands. The other ends of the arms are controllably movably connected to the truss 12 arranged on the conical hopper 11. The first manipulators 13 work together in multiple numbers to optimize the task allocation; A third manipulator 5 is also arranged in the conical hopper 11 to handle abnormalities such as foreign objects, large items, and working errors found by AI intelligent analysis;

[0115] The manipulator in S3 is the second manipulator 3. The conveyor belt is a wide spreading surface arranged on the operation spreading platform 21 outside the conical hopper 11. The truss 12 extends above the spreading surface. The second manipulators 3 work together in multiple numbers to optimize the task allocation;

[0116] The spreading in S3 also includes automatically adjusting the angle and force of the bionic structure first manipulator 13 feeding materials to the conveyor belt after bag-breaking by the AI artificial intelligence, as well as the cooperation of multiple first and second manipulators 3.

[0117] Refer to Figure 3 、 Figure 4 for the AI intelligent control method of feeding, bag-breaking and spreading by the manipulator,

[0118] The grasping in S2 includes the following steps:

[0119] S21. Adjust the grasping order of the manipulator according to the volume information and density information in the garbage bag feature information;

[0120] S22. Adjust the grasping mode of the manipulator according to the material information in the garbage bag feature information. The clamping mode is for rigid garbage bags, hard plastic bags, etc.; the flexible grasping mode is for kitchen waste garbage bags to prevent breakage; the adsorption grasping mode is for light garbage, paper, plastic bags, and expanded polystyrene;

[0121] S23. Calculate the optimal grasping points and paths through deep learning (CNN + Transformer), and optimize the manipulator path through an intelligent obstacle avoidance algorithm to avoid collisions between manipulators;

[0122] S24. Grasp the garbage in sequence according to the grasping order, grasping mode, optimal grasping points, and optimal grasping paths;

[0123] S25. Judge the toughness of the garbage bag according to the pressure sensor and compare it with the preset toughness value;

[0124] The bag breaking of S2 includes the following steps:

[0125] S26. Adopt corresponding bag breaking schemes according to the toughness of the garbage bag. When the pressure sensor judges that the toughness of the garbage bag is greater than the preset toughness value, use high-frequency vibration to break the bag through micro-vibration cutting; when the pressure sensor judges that the toughness of the garbage bag is less than the preset toughness value, grasp both ends of the garbage bag with both hands and rotate them in opposite directions to pull and tear, forming a torsional tear to complete the bag breaking operation;

[0126] The spreading of S3 also includes the following steps:

[0127] S31. Turn over and spread the bag-broken garbage bag on the conveyor belt;

[0128] S32. Automatically adjust the feeding angle and strength for uniform spreading to prevent garbage accumulation and improve the screening efficiency;

[0129] S33. Vibrate the conveyor belt at a vibration frequency of 50 - 200 Hz to prevent the garbage from forming clusters and improve the spreading uniformity;

[0130] S34. Sort the recyclable garbage after spreading.

[0131] Refer to Figure 5 , and also includes the following steps:

[0132] S4: Optimize the task allocation of multiple manipulators through an intelligent scheduling algorithm to improve the garbage processing throughput;

[0133] S5. At the bottom of the conical hopper 11, there is a slightly inclined (2° - 15°) leachate diversion slope 71 to direct the leachate to the designated drainage channel 72; a metal grid + polymer permeable filter layer is installed at the bottom of the conical hopper 11 to intercept solid waste; an intelligent liquid level sensor 74 is used. When the leachate reaches the set value, the extraction pump 75 is automatically started to discharge the leachate into the sewage treatment system; the leachate pipeline uses an anti-blocking spiral sewage pipe.

[0134] The data of the system and method of the present invention compared with the existing traditional manual + chain plate + bag breaker in terms of recognition and control accuracy are shown in Table 1 below:

[0135] Control objective Recognition accuracy / control error range Recognition deviation of garbage bag grasping point ≤±5cm Grasping success rate ≥95.8％ Control deviation of tearing path ≤±10° Track offset error of paving path ≤3cm Control delay (end-to-end processing time) ≤80ms

[0136] Table 1

[0137] Therefore, the system of the present invention still has high recognition rate, high movement accuracy and stable operation ability in a high-complexity material environment, which constitutes the core advantages and technical barriers of the present invention different from traditional equipment.

[0138] The AI recognition accuracy rate ≥ 95%, the path deviation rate ≤ 2°, and the grasping error is controlled within ±5 cm; the system throughput is 8 tons / hour, and the paving uniformity CV value < 10%.

[0139] The performance / cost / efficiency comparison table of the system and method of the present invention compared with the existing traditional manual + chain plate + bag breaker is shown in Table 2 below:

[0140]

[0141] Table 2

[0142] In summary, compared with the existing traditional garbage bag-breaking paving system, the present invention shows significant improvements in terms of treatment efficiency, automation level, energy consumption control and recycling effect, and has significant technological progressiveness and industrial application prospects.

Claims

1. A garbage feeding, bag-breaking and spreading system based on an artificial intelligence bionic manipulator, comprising: A feeding device: including a conical hopper with a relatively large top opening for dumping garbage, a truss located above the conical hopper, and a first manipulator slidably connected to the truss through a flexible and telescopic arm; A spreading device: including an operation spreading platform located outside the conical hopper, a conveying drive structure and a vibrating screen device respectively connected to the operation spreading platform. The truss extends above the operation spreading platform. The operation spreading platform includes an operation table and a spreading surface on the upper part of the operation table. The spreading surface is a wide conveyor belt for garbage transmission. The conveying drive structure drives the spreading surface to move horizontally, and the vibrating screen device drives the spreading surface to vibrate. Both the conveying drive structure and the vibrating screen device are located below the spreading surface; A second manipulator arranged on the side of the operation table of the operation spreading platform; An intelligent recognition and path optimization device, including an image acquisition device and an AI intelligent control module electrically connected to the image acquisition device. The AI intelligent control module includes an identification and analysis module, a path and action optimization module connected to the identification and analysis module, a manipulator control module connected to the path and action optimization module, and a spreading surface movement control module electrically connected to the identification and analysis module. The image acquisition device includes a camera, a 3D lidar and / or a near-infrared detector. The camera, 3D lidar and / or near-infrared detector are arranged above the conical hopper and the operation spreading platform, and transmit the three-dimensional information of the garbage in the conical hopper and on the operation spreading platform, including the garbage accumulation situation, garbage type, etc., and the action information of the first and second manipulators to the AI intelligent control module. The identification and analysis module includes an AI vision processing unit and an intelligent analysis unit. The AI vision processing unit receives, identifies and processes the three-dimensional information and the action information in real time, and transmits the processing result to the intelligent analysis unit. The intelligent analysis unit conducts intelligent analysis on the garbage and the actions of the first and second manipulators. The intelligent analysis includes discovering abnormalities such as foreign objects, large items, and working errors according to the three-dimensional information of the garbage, extracting the material information, volume information, density information and other garbage bag feature information of the garbage bag, and sending the analysis result to the path and action optimization module. The path and action optimization module includes a deep learning unit, which calculates through deep learning the grasping point, grasping force, grasping method, clamping force, movement action on the truss, movement distance, bag-breaking method and bag-breaking action of the first manipulator, and the tearing, sorting and spreading actions of the second manipulator, and transmits them to the manipulator control module to control the grasping and clamping, clamping movement and bag-breaking actions of the first manipulator, and the tearing, sorting and spreading actions of the second manipulator. The spreading surface movement control module controls the movement of the spreading surface according to the garbage condition on the operation spreading platform analyzed by the identification and analysis module to convey the spread garbage, and also optimizes the manipulator path through an intelligent obstacle avoidance algorithm to avoid collisions between manipulators; Both the first manipulator and the second manipulator include a bionic structure with dual-handed equivalent fingers and crossable fingers on both hands. The two hands of the first manipulator are respectively connected to arms that can be bent, telescoped, and moved on the truss. Force feedback sensors are installed on the finger surfaces to monitor the tension of the garbage bag in real time and transmit signals to the motion and rotation control module. Each finger is a multi-joint structure, and multi-joint motion and rotation control structures are attached to each finger and the arm to control the bionic actions of the fingers and the arm. The motion and rotation control module is integrated on the multi-joint motion and rotation control structure; The motion and rotation control module is respectively connected to the recognition and analysis module and the manipulator control module, transmits the sensing signal to the recognition and analysis module, and the manipulator control module transmits the grasping action and optimized path control instructions optimized by the path and motion optimization module to the motion and rotation control module to control the operations of the first and second manipulators by the multi-joint motion and rotation control structure.

2. The garbage feeding, bag-breaking and spreading system based on the artificial intelligence bionic manipulator according to claim 1, wherein, The path and optimization module further includes an intelligent grasping mode unit that stores different grasping modes of the first manipulator and determines which intelligent grasping mode the first manipulator adopts according to the recognition and analysis of the garbage by the recognition and analysis module. The intelligent grasping modes include a clamping mode for rigid garbage bags, hard plastic bags, etc.; a flexible grasping mode for preventing damage to kitchen waste garbage bags; an adsorption grasping mode for light garbage, paper, plastic bags, expanded polystyrene, etc. High-frequency vibration blades are equipped at the finger tips of the first manipulator. The multi-joint motion and rotation control structure includes a vibration drive structure, and the motion and rotation control module includes vibration drive control for the vibration drive structure. The path and optimization module further includes a bag-breaking method unit that stores different bag-breaking methods of the first manipulator and determines which bag-breaking method the first manipulator adopts according to the recognition and analysis of the garbage by the recognition and analysis module. The manipulator control module transmits the bag-breaking method control instructions optimized by the path and motion optimization module to the motion and rotation control module to control the bag-breaking actions of the multi-joint motion and rotation control structure. The bag-breaking methods include tearing bag-breaking, high-frequency vibration bag-breaking, and twisting and pulling bag-breaking.

3. The garbage feeding, bag-breaking and spreading system based on the artificial intelligence bionic manipulator according to claim 2, wherein The cross-section of the conical hopper is a polyhedron, circular or elliptical, and the angle of the cone ranges from 15° to 90° with respect to the vertical line. Its inner sidewall is coated with wear-resistant and anti-corrosive materials, and the truss spans the sidewall area of the conical hopper, that is, the second sidewall is lower than other sidewall parts, namely the first sidewall; it further includes a third manipulator, which is arranged inside the conical hopper to remove abnormal objects such as foreign matters and large pieces. The third manipulator includes a bionic structure with equal hands and corresponding fingers that can cross each other. Force feedback sensors are installed on the surface of the fingers. Each finger is a multi-joint structure, and each finger is attached with a multi-joint movement and rotation control structure to control the bionic movement of the fingers. An action and rotation control module is integrated on the multi-joint movement and rotation control structure. The force feedback sensor monitors the tension of the garbage bag in real time and transmits the sensing signal to the recognition and analysis module. The action and rotation control module is connected to the recognition and analysis module and the manipulator control module. The manipulator control module transmits the action and optimized path control instructions optimized by the path and action optimization module to the action and rotation control module to control the multi-joint movement and rotation control structure to operate the third manipulator to remove foreign matters and large pieces.

4. The garbage feeding and bag-breaking spreading system based on the artificial intelligence bionic manipulator according to claim 3, wherein A fourth manipulator is further arranged on the truss on the operation paving platform. The fourth manipulator includes a bionic structure with equal hands and corresponding fingers that can cross each other. Force feedback sensors are installed on the surface of the fingers. Each finger is a multi-joint structure, and each finger is attached with a multi-joint movement and rotation control structure to control the bionic movement of the fingers. An action and rotation control module is integrated on the multi-joint movement and rotation control structure. The force feedback sensor monitors the garbage tension in real time and transmits the sensing signal to the recognition and analysis module. The action and rotation control module is respectively connected to the recognition and analysis module and the manipulator control module. The manipulator control module transmits the flicking action and optimized path control instructions optimized by the path and action optimization module to the action and rotation control module to control the multi-joint movement and rotation control structure to operate the flicking action of the fourth manipulator.

5. The garbage feeding, bag-breaking and spreading system based on the artificial intelligence bionic manipulator according to claim 4, wherein The path and optimization module further includes an intelligent flipping and paving unit, which stores the flipping and paving mode of the first manipulator and starts the flipping and paving of the first manipulator according to the recognition and analysis of the bag-breaking state of the garbage bag by the recognition and analysis module. The manipulator control module transmits the flipping and paving control instructions issued by the path and action optimization module to the action and rotation control module, and controls the first manipulator to flip 180° after bag-breaking through the multi-joint movement and rotation control structure to evenly spread the garbage on the conveyor belt.

6. The garbage feeding, bag-breaking and spreading system based on the artificial intelligence bionic manipulator according to claim 5, wherein The first, second, third, and fourth manipulators are each one or more. The path and optimization module further includes a manipulator coordination unit that coordinates the grasping, clamping, moving, bag-breaking, tearing, sorting, spreading, rejecting, and flicking actions of the manipulators. The manipulator coordination unit uses an intelligent reinforcement learning (RL) load balancing algorithm to dynamically allocate grasping tasks according to the waste flow and optimize the throughput. The manipulator coordination unit further includes a synchronous cooperation mode, and the synchronous cooperation mode includes multiple first manipulators pulling and breaking bags simultaneously and different manipulators being responsible for different tasks for synchronous cooperation. The toughness of the garbage bag is detected by the force feedback sensor, and the optimal bag-breaking mode is automatically selected.

7. The garbage feeding, bag-breaking and spreading system based on the artificial intelligence bionic manipulator according to claim 6, wherein, The vibration frequency of the high-frequency vibration blade is 30 - 50 Hz, and the vibration frequency of the vibrating screen device is 50 - 200 Hz. The feeding device, spreading device, second manipulator, and intelligent recognition and path optimization device are arranged in a closed chamber and are located in the same space.

8. The garbage feeding, bag-breaking and spreading system based on the artificial intelligence bionic manipulator according to claim 7, wherein A garbage leachate collection system is connected to the bottom of the garbage conical hopper. The garbage leachate collection system includes a diversion slope surface, a drainage channel, a filtration system, and an automatic extraction system. The diversion slope surface is a diversion slope surface that forms an inclination of 2° - 15° with the bottom of the conical hopper. The diversion slope surface is connected to the drainage channel. The drainage channel is an anti-blocking spiral sewage discharge pipe. A metal grid is installed at the bottom of the conical hopper, and a polymer permeable filter layer is coated on the metal grid. The automatic extraction system includes an intelligent liquid level sensor arranged at the lower part of the conical hopper, an extraction pump communicated with the drainage channel, and a sewage treatment system connected to the end of the drainage channel. The leachate flows through the diversion slope surface to the drainage channel.

9. AI intelligent control manipulator garbage feeding and bag-breaking paving method, characterized in that Including: S1. Intelligent feeding step: Use AI vision to automatically identify the three-dimensional information of the garbage dumped into the conical hopper, including the stacking situation, garbage type, and garbage bag characteristic information such as the material information, volume information, and density information of the garbage bag. S2. Efficient manipulator grasping and bag-breaking step: The AI artificial intelligence automatically adjusts the bionic structure manipulator's bionic hands to cross-grasp the garbage bag according to the three-dimensional information of the garbage, and moves the garbage bag to the conveyor belt for bag-breaking after that. The manipulator includes a bionic structure with equal fingers on both hands and the corresponding fingers on both hands can cross. Force feedback sensors are installed on the surface of each finger to monitor the tension of the garbage bag in real time and transmit signals to the AI artificial intelligence to adjust the grasping and bag-breaking methods and strategies in real time. S3. Uniform spreading step by the manipulator: Use the bionic structure manipulator to uniformly spread the broken garbage bag.

10. The AI intelligent control manipulator garbage feeding, bag-breaking, and spreading method according to claim 9, wherein: The cross-section of the conical hopper in S1 is polygonal, circular, or elliptical, and the conical angle range is 15° - 90° with respect to the vertical line, and its inner wall is coated with wear-resistant and corrosion-resistant materials. The manipulator described in S2 is the first manipulator. The two hands of the first manipulator also include bendable and telescopic arms respectively connected to the two hands. The other ends of the arms are controllably movably connected to a truss arranged on the conical hopper. The first manipulators work cooperatively in multiple numbers to optimize task allocation. A third manipulator located in the conical hopper is also provided to handle abnormalities such as foreign objects, large pieces, and working errors found through AI intelligent analysis. The manipulator described in S3 is the second manipulator. The conveyor belt is a wide paving surface arranged on the operation paving platform outside the conical hopper. The truss extends above the paving surface. The second manipulators work cooperatively in multiple numbers to optimize task allocation. The paving in S3 also includes automatically adjusting the angle and strength of the first manipulator with a bionic structure to feed materials onto the conveyor belt after bag breaking through AI artificial intelligence, as well as the cooperative operation of multiple first and second manipulators. The grasping in S2 includes the following steps: S21. Adjust the grasping sequence of the manipulator according to the volume information and density information in the garbage bag feature information. S22. Adjust the grasping mode of the manipulator according to the material information in the garbage bag feature information. The clamping mode is for rigid garbage bags, hard plastic bags, etc.; the flexible grasping mode is for kitchen waste garbage bags to prevent breakage; the adsorption grasping mode is for light garbage, paper, plastic bags, expanded polystyrene. S23. Calculate the optimal grasping points and paths through deep learning (CNN + Transformer), and optimize the manipulator path through an intelligent obstacle avoidance algorithm to avoid collisions between manipulators. S24. Grasp the garbage in sequence according to the grasping sequence, grasping mode, optimal grasping points, and optimal grasping paths. S25. Judge the toughness of the garbage bag according to the pressure sensor and compare it with the preset toughness value. The bag breaking in S2 includes the following steps: S26. Adopt corresponding bag breaking schemes according to the toughness of the garbage bag. When the pressure sensor judges that the toughness of the garbage bag is greater than the preset toughness value, adopt high-frequency vibration bag breaking through micro-vibration cutting; when the pressure sensor judges that the toughness of the garbage bag is less than the preset toughness value, grasp both ends of the garbage bag with both hands and rotate the two ends in opposite directions to pull and tear, forming a torsional tear to complete the bag breaking operation. The paving in S3 also includes the following steps: S31. Turn over and pave the bag-broken garbage bag on the conveyor belt. S32. Automatically adjust the feeding angle and strength for uniform paving to prevent garbage accumulation and improve the screening efficiency. S33. Vibrate the conveyor belt at a vibration frequency of 50 - 200 Hz to prevent the garbage from forming clusters and improve the paving uniformity. S34. Sort the recyclable garbage after paving. It also includes the following steps: S4: Optimize the task allocation of multiple manipulators through an intelligent scheduling algorithm to improve the garbage processing throughput. S5. A leachate diversion slope with a slight inclination (2° - 15°) is arranged at the bottom of the conical hopper to direct the leachate to a designated drainage channel; a metal grid + polymer permeable filter layer is installed at the bottom of the conical hopper to intercept solid garbage; an intelligent liquid level sensor is adopted. When the leachate reaches the set value, the extraction pump is automatically started to discharge the leachate into the sewage treatment system; the leachate pipeline adopts an anti-blocking spiral sewage pipe.

Images