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

Abstract

The present invention provides a garbage loading, bag breaking and paving system and method based on an artificial intelligence bionic manipulator, comprising a loading device, a paving device and an intelligent identification and path optimization device: the loading device comprises a conical hopper, a truss located above the conical hopper and a manipulator slidably connected to the truss via a bendable and telescopic arm; the paving device comprises an operating 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 to the image acquisition device; the manipulator comprises a bionic structure having equal fingers on both hands, and the corresponding fingers of both hands can be crossed; force feedback sensors are installed on the surface of the fingers to monitor the tension of the garbage bag in real time; each finger has a multi-joint structure for bionic movements. The present invention replaces the traditional garbage pit + chain conveyor + manual assisted paving + bag breaker, which are multiple devices that are decentralized and independently processed, and is the first to integrate loading, bag breaking and paving.

Description

Technical Field

[0001] The present invention relates to the field of solid waste resource processing technology, specifically a system and method for automatically loading, intelligently breaking, and spreading waste using an artificial intelligence-controlled biomimetic robotic arm. This system and method, based on an artificial intelligence-controlled biomimetic robotic arm, can be widely used in waste sorting, resource recovery, and waste incineration pretreatment. Background Art

[0002] As the amount of municipal solid waste continues to rise, waste management is gradually shifting from end-of-life incineration and landfill to front-end resource utilization and waste reduction. Large amounts of mixed waste are being sent directly to incineration systems without effective pretreatment, reducing the efficiency of recyclables and increasing processing costs and carbon emissions.

[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, its processing efficiency, degree of automation, and environmental adaptability directly affect the operation quality of the entire line. However, the current industry widely adopts the combination of "manual feeding + bag breaking machine + chain conveyor belt paving", which has the following significant shortcomings:

[0004] 1. Decentralized processes and low collaborative efficiency

[0005] Traditional systems rely heavily on serial equipment for segmented operations (e.g., manual loading, gear bag breaking, and chain conveyor paving). These devices occupy a large area, require numerous maintenance points, and lack real-time coordination and adjustment. This can easily lead to problems like material accumulation, bag jamming, and bag spillage, creating process bottlenecks. These decentralized processes also present the following challenges:

[0006] 1) Problems with chain conveyors: high energy consumption and high operating costs; easy to clog, garbage bags accumulate on the chain conveyor, affecting processing efficiency; manual paving is required, which 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, they cannot intelligently identify the characteristics of garbage bags, and the bag breaking method is not accurate; they have high requirements for uniform feeding and are prone to accumulation and jamming; they have poor adaptability and the bag breaking effect is uneven 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) Material pit problem: Traditional material pits cannot automatically complete automatic loading and require continuous manual intervention to push the garbage onto the chain conveyor. The garbage cannot be stacked to prevent the chain conveyor from getting stuck and affecting the process operation.

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

[0011] In the wet garbage and mixed domestic waste treatment scenarios, due to the complex material composition and different garbage bag structures, most bag breaking and spreading operations currently require manual assistance, which is labor-intensive, poses great health risks, and is difficult to ensure stability.

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

[0013] The back-end optical selection, air selection, and intelligent sorting equipment have high requirements for the thickness of the material layer and the uniformity of particle size distribution, while traditional paving systems are difficult to achieve precise control, which often results in an increase in the misselection rate and a decrease in sorting efficiency.

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

[0015] Currently, the industry's application of AI in loading and bag-breaking processes is still in its infancy, with no control mechanisms that can continuously learn and adapt to material differences. Rigid grasping strategies and single-path planning can lead to errors and downtime in situations like high-mix or compacted waste.

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

[0017] Although some existing automated systems for loading, bag breaking and spreading are designed with robotic arms, there is no integrated equipment for loading, bag breaking and spreading. Moreover, they cannot intelligently adapt to the shape of garbage bags, resulting in failure to grab or damage to garbage bags, and they are also unable to identify, grab and remove foreign objects.

[0018] As the cost of waste incineration rises, waste pretreatment systems have become a critical component of intelligent solid waste sorting. Existing systems rely on manual labor or multiple devices in tandem for bag breaking and spreading, resulting in high energy consumption, heavy pollution, and low efficiency. This makes it difficult to meet the industrial needs of resource sorting. New technologies are urgently needed that are efficient, simple, pollution-free, and can revolutionize existing treatment methods while also meeting resource sorting requirements. Summary of the Invention

[0019] The present invention solves the technical problem that the existing garbage loading, bag breaking and paving processes cannot be better automated due to various defects such as scattered processes, low coordination efficiency, high energy consumption, high cost, easy blockage, high degree of manual dependence, harsh working environment, poor bag breaking adaptability, inability to accurately adapt to the rhythm of the sorting system, uneven paving, and machine jamming. It provides a garbage loading, bag breaking and paving system and method based on an artificial intelligence bionic manipulator, which effectively breaks the technical bottleneck of the traditional pretreatment link and provides new equipment support and intelligent paths for the urban solid waste resource industry. It can overcome the above-mentioned defects well and realize the automation of loading, bag breaking and paving without on-site manual intervention. It can complete automated loading, precise bag breaking and efficient paving operations in a closed and clean environment, and provide a stable and reliable front-end processing foundation for subsequent resource extraction and intelligent sorting systems. Improve the intelligence and automation level of garbage treatment, reduce labor costs, and improve garbage classification accuracy and resource recovery efficiency. The technical solution of the present invention is as follows:

[0020] A garbage loading and bag breaking paving system using an artificial intelligence bionic manipulator, comprising:

[0021] The loading device comprises a conical hopper with a large top opening for dumping garbage, a truss located above the conical hopper, and a first manipulator slidably connected to the truss via a bendable and telescopic arm;

[0022] The paving device comprises an operating paving platform located outside the conical hopper, a conveying transmission structure and a vibrating screen device respectively connected to the operating paving platform, the truss extending to above the operating paving platform, the operating paving platform comprising an operating table and a paving surface above the operating table, the paving surface being a wide conveyor belt for conveying garbage, the conveying transmission structure driving the paving surface to move horizontally, the vibrating screen device driving the paving surface to vibrate, and the conveying transmission structure and the vibrating screen device both being located below the paving surface;

[0023] A second manipulator provided on the side of the work platform of the working paving platform;

[0024] An intelligent recognition and path optimization device includes an image acquisition device and an AI intelligent control module electrically connected to the image acquisition device. The AI ​​intelligent control module includes a recognition and analysis module, a path and motion optimization module connected to the recognition and analysis module, a manipulator control module connected to the path and motion optimization module, and a paving surface movement control module electrically connected to the recognition and analysis module. The image acquisition device includes a camera, a 3D laser radar and / or a near-infrared detector. The camera, the 3D laser radar and / or the near-infrared detector are arranged above the conical hopper and the working paving platform, and transmit the three-dimensional information of garbage in the conical hopper and on the working paving platform, including garbage accumulation conditions, garbage types, etc., and the motion information of the first and second manipulators collected in real time to the AI ​​intelligent control module; the recognition and analysis module includes an AI visual processing unit and an intelligent analysis unit. The AI ​​visual processing unit receives, recognizes and processes the three-dimensional information and the motion information in real time, and transmits the processing results to the intelligent analysis unit. The intelligent analysis unit performs intelligent analysis on the garbage and the movements of the first and second manipulators. The intelligent analysis includes detecting anomalies such as foreign objects, large items, working errors, etc. based on the three-dimensional information of the garbage, extracting garbage bag feature information such as material information, volume information and density information of the garbage bag, and sending the analysis results to the path and motion optimization module. The path and motion optimization module includes a deep learning unit, which calculates the grasping point, grasping force, grasping method, clamping force, movement on the truss, movement distance, bag breaking method and bag breaking action and optimized path of the first manipulator through deep learning, and the tearing, sorting and paving actions of the second manipulator, and transmits them to the manipulator control module to control the grasping, 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 transfer the paved garbage according to the garbage status on the working paving platform analyzed by the recognition and analysis module, and also optimizes the manipulator path through the intelligent obstacle avoidance algorithm to avoid collisions between manipulators.

[0025] Both the first and second manipulators include a bionic structure with equal fingers on both hands, and the corresponding fingers of both hands can be crossed. The hands of the first manipulator are also respectively connected to arms that can be bent, extended and moved on a truss. Force feedback sensors are installed on the surface of the fingers 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 arm is attached with a multi-joint motion and rotation control structure to control the fingers and arms to perform bionic movements. 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 to transmit the sensor signal to the recognition and analysis module. 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 multi-joint motion and rotation control structure to operate the motion of the first and second manipulators.

[0026] The path and optimization module also includes an intelligent grasping mode unit, which stores different grasping modes of the first manipulator, and determines that the first manipulator adopts different intelligent grasping modes based on 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 kitchen waste bags from being damaged; and an adsorption grasping mode for lightweight garbage, paper, plastic bags, expanded polystyrene, etc. The finger end of the first manipulator is equipped with a high-frequency vibrating blade, and the multi-joint motion and rotation control structure includes a vibration drive structure, and the motion and rotation control module includes a vibration drive control of the vibration drive structure; the path and optimization module also includes a bag breaking method unit, which stores different bag breaking methods of the first manipulator, and determines that the first manipulator adopts different bag breaking methods based on the recognition and analysis of the garbage by the recognition and analysis module. The manipulator control module transmits the bag breaking method control instruction 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 and breaking bags, high-frequency vibration and twisting and breaking bags.

[0027] The cross section of the conical hopper is a polyhedron, a circle or an ellipse, and the angle range of the cone is 15°-90° with the vertical line. The inner side wall is coated with wear-resistant and corrosion-resistant material, and the side wall area of ​​the truss spanning the conical hopper, that is, the second side wall, is lower than the other side wall part, that is, the first side wall; it also includes a third manipulator, which is arranged on the inner side of the conical hopper to remove the foreign matter, large pieces and other abnormalities. The third manipulator includes a bionic structure with equal fingers on both hands, and the corresponding fingers of both hands can be crossed. A force feedback sensor is installed on the surface of the finger, and each finger is a multi-joint structure. Each finger is attached There is a multi-joint motion and rotation control structure to control the bionic motion of the fingers. The multi-joint motion and rotation control structure is integrated with a motion and rotation control module. The force feedback sensor monitors the tension of the garbage bag in real time and transmits the sensor signal to the identification and analysis module. The motion and rotation control module is connected to the identification and analysis module and the manipulator control module. The manipulator control module transmits the motion and 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 third manipulator to remove foreign objects and large items.

[0028] A fourth manipulator is also provided on the truss on the working paving platform. The fourth manipulator includes a bionic structure with equal fingers on both hands and corresponding fingers on both hands that can be crossed. A force feedback sensor is installed on the surface of the fingers. Each finger is a multi-joint structure. Each finger is attached with a multi-joint action and rotation control structure to control the fingers to perform bionic actions. The multi-joint action and rotation control structure is integrated with a motion and rotation control module. The force feedback sensor monitors the garbage tension in real time and transmits the sensor signal to the identification and analysis module. The motion and rotation control module is connected to the identification and analysis module and the manipulator control module respectively. The manipulator control module transmits the plucking 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 action and rotation control structure to operate the plucking 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 based on the recognition and analysis of the broken bag status 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 motion optimization module to the motion and rotation control module, and controls the first manipulator to flip 180° after breaking the bag through the multi-joint motion and rotation control structure, so as to evenly spread the garbage on the conveyor belt.

[0030] The first, second, third and fourth manipulators are one or more, and the path and optimization module also includes a manipulator coordination unit to coordinate the grabbing, clamping, moving, bag breaking, tearing, sorting, spreading, rejecting and plucking actions of each manipulator; the manipulator coordination unit adopts an intelligent reinforcement learning (RL) load balancing algorithm to dynamically allocate grabbing tasks according to garbage flow and optimize throughput; the manipulator coordination unit also includes a synchronous collaboration mode, which includes multiple first manipulators pulling and breaking bags at the same time and different manipulators taking charge of different tasks to collaborate synchronously, and automatically selecting the optimal bag breaking mode by detecting the toughness of the garbage bag through the force feedback sensor.

[0031] The vibration frequency of the high-frequency vibrating blade is 30-50Hz, and the vibrating screen device adopts a vibration frequency of 50-200Hz; the feeding device, paving device, second manipulator, intelligent identification 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 conical garbage hopper, and 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 is inclined 2°-15° with the bottom of the conical hopper. The diversion slope is connected to the drainage channel, and the drainage channel is an anti-clogging spiral sewage pipe. A metal grid is installed at the bottom of the conical hopper, and the metal grid is coated with a polymer permeable filter layer. The automatic extraction system includes an intelligent liquid level sensor arranged at the bottom of the conical hopper, an extraction pump connected to 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] AI intelligent control robot garbage feeding bag breaking and paving method, including:

[0034] S1. Intelligent loading 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 feature information such as garbage bag material information, volume information and density information; S2. Efficient bag grabbing and bag breaking step of the manipulator: The AI ​​artificial intelligence automatically adjusts the bionic structure of the manipulator based on the three-dimensional information of the garbage, crosses the corresponding fingers of the bionic hands to grab the garbage bag, and moves the garbage bag to the conveyor belt for bag breaking. The manipulator includes a bionic structure with two equal fingers and the corresponding fingers of the two hands can be crossed. 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, so as to adjust the grabbing and bag breaking methods and strategies in real time;

[0035] S3. Robotic arm uniform spreading step: Use a bionic structure robot to evenly spread the broken garbage bags.

[0036] The cross section of the conical hopper S1 is multifaceted, circular or elliptical, the angle of the cone is in the range of 15°-90° to the vertical line, and the inner wall is coated with wear-resistant and corrosion-resistant material;

[0037] The manipulator S2 is a first manipulator, and the hands of the first manipulator further include bendable and retractable arms connected to the hands respectively, and the other end of the arm is controllably connected to the truss provided on the conical hopper. The first manipulator is a plurality of collaborative operations to optimize task allocation; a third manipulator is also provided in the conical hopper to handle abnormalities such as foreign objects, large parts, and working errors discovered according to AI intelligent analysis;

[0038] S3: The manipulator is a second manipulator, the conveyor belt is a wide paving surface arranged on the working paving platform outside the conical hopper, the truss extends above the paving surface, and the second manipulator is a plurality of collaborative operations to optimize task allocation;

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

[0040] The crawling of S2 includes the following steps:

[0041] S21. Adjusting the grabbing order of the robot arm based on the volume information and density information in the garbage bag feature information;

[0042] S22. Adjust the gripping mode of the manipulator based on the material information in the garbage bag feature information. The gripping mode is for rigid garbage bags, hard plastic bags, etc.; the flexible gripping mode is for kitchen waste bags to prevent damage; and the adsorption gripping mode is for lightweight garbage, such as paper, plastic bags, and expanded polystyrene.

[0043] S23, calculate the optimal grasping point and path through deep learning (CNN+Transformer), optimize the robot path through intelligent obstacle avoidance algorithm, and avoid collisions between robots;

[0044] S24, grabbing garbage in sequence according to the grabbing order, grabbing mode, optimal grabbing point and optimal grabbing path;

[0045] S25, determining the toughness of the garbage bag according to the pressure sensor and comparing it with a preset toughness value;

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

[0047] S26. Adopt a corresponding bag-breaking scheme based on the toughness of the garbage bag. When the pressure sensor determines that the toughness of the garbage bag is greater than a preset toughness value, high-frequency vibration is used to break the bag, through micro-vibration cutting. When the pressure sensor determines that the toughness of the garbage bag is less than the preset toughness value, the garbage bag is grasped at both ends with both hands and rotated in opposite directions to tear the ends, forming a twisting tear to complete the bag-breaking operation.

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

[0049] S31, turning over the broken garbage bags and spreading them on the conveyor belt;

[0050] S32, automatically adjusts the feeding angle and force to spread the material evenly, prevents garbage accumulation and improves screening efficiency;

[0051] S33, using a 50-200Hz vibration conveyor belt to prevent garbage from clumping and improve paving uniformity;

[0052] S34, sorting the recyclable waste after paving; further comprising the following steps:

[0053] S4: Optimize the task allocation of multiple robots through intelligent scheduling algorithms to improve waste processing throughput;

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

[0055] Technical effects of the present invention:

[0056] The method of the present invention sets up a bionic structure manipulator, equipped with AI vision to automatically identify the three-dimensional information of the dumped garbage, and uses AI artificial intelligence to automatically adjust the manipulator's fingers to cross according to the three-dimensional information of the garbage, grab the garbage bag in the garbage pile, move it to the conveyor belt to break the bag and spread it evenly, that is, the bionic manipulator can be controlled by artificial intelligence to automatically load, break the bag and spread it. It replaces the traditional garbage loading system of garbage pit + chain conveyor + manual assisted paving + bag breaker and other multiple devices to work independently and separately to process garbage, and creates the first integrated loading, bag breaking and paving system. The equipment eliminates the need for a bag breaker and labor in all processes, greatly reducing energy consumption and costs. By adjusting the operation of the robot through artificial intelligence, various defects such as easy blockage of the conveyor belt, poor adaptability and accuracy of bag breaking, uneven paving, and machine jamming are eliminated, and the loading, bag breaking and paving are automated. It can complete automatic loading, precise bag breaking and efficient paving operations in a closed and clean environment, providing a stable and reliable front-end processing foundation for subsequent resource extraction and intelligent sorting systems, improving the intelligence and automation level of garbage disposal, reducing labor costs, and improving garbage classification accuracy and resource recovery efficiency.The system of the present invention includes a feeding device, a paving device, and an intelligent identification and path optimization device. The feeding device includes three components: a conical hopper, a truss, and a first manipulator. Their respective structures, positional relationships, and connection relationships enable the flexible and retractable first manipulator to penetrate into the conical hopper, grab the garbage dumped into it, and move it out of the conical hopper through the truss. The opening at the top of the conical hopper is relatively large, which firstly indicates that it is a cone with a larger top and a smaller bottom. Secondly, the garbage truck dumps the garbage from top to bottom, so the opening is relatively large, and it is convenient for the first manipulator to drop from the truss into the conical hopper to grab and dump the garbage. The paving device includes an operating paving platform, a conveying transmission structure and a vibrating screen device for paving garbage. The truss extends above the operating paving platform, so that the first manipulator can clamp the garbage bag above the operating paving platform to break the bag. The second manipulator is set on the side of the operating table of the operating paving platform to perform paving operations. 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 collects the real-time images inside the conical hopper and on the operating paving platform. The three-dimensional information of garbage on the paving platform, including the accumulation of garbage, the type of garbage, and the movement information of the first and second manipulators, is transmitted to the AI ​​intelligent control module, which is received, identified, processed, analyzed and controlled in real time by each module of the AI ​​intelligent control module. The analysis includes deep learning to calculate the grasping point, grasping force, grasping method, clamping force, movement on the truss, movement distance, bag breaking method and bag breaking action and optimized path of the first manipulator, the tearing, sorting and paving actions of the second manipulator, so as to control the relevant actions of the first and second manipulators and the garbage after paving. The first and second manipulators are simulation manipulators specially designed for the system, and their finger structures and module settings can be controlled by the above-mentioned AI intelligent control module to imitate the human hand's grasping, holding, pulling and tearing, sorting, paving, arm bending and stretching movements, thereby coordinating with other components to achieve a high degree of integration of manipulator grasping, intelligent bag breaking, dynamic paving and other operations, with a high degree of integration. The system of the present invention can replace 4 traditional equipment, reduce more than 80% of manual intervention, and increase the single-line daily garbage processing capacity by 50%.

[0057] Dumping garbage into the conical hopper, the cone can use the downward movement trend formed by the garbage's own weight to control the loading, low energy consumption, reduced operating costs, and different cone angles form different movement trends, which can be targeted at different types of garbage; the robot can accurately grab the garbage bag and move it to the conveyor belt, and can intelligently control the bag breaking and paving, thus solving the problem of accumulation on the working paving platform, improving processing efficiency, and saving a lot of energy and manpower;

[0058] Through the image acquisition device and the AI ​​intelligent control module electrically connected to the image acquisition device, AI visual analysis and path optimization are realized. It intelligently analyzes the detection of foreign objects, large items, working errors, and extracts the material information, volume information and density information of the garbage bag. The analysis results are sent to the path and motion optimization module. Then, through deep learning calculation, the robot's grasping points and optimized path are obtained to optimize the grasping order of the garbage bags. The robot's movement is controlled to automatically identify the type of garbage bag, the best grasping point, and the optimized movement path of the robot. The robot path is optimized through the intelligent obstacle avoidance algorithm to avoid collisions between robots.

[0059] Intelligent robotic gripping + dynamic force feedback: adapt to different garbage bags and prevent bags from breaking;

[0060] Other efficient bag breaking: The force feedback sensor detects the toughness of the garbage bag, combines high-frequency vibration bag breaking and torsional shear bag breaking, and automatically selects the optimal bag breaking mode;

[0061] Intelligent paving optimization: Reliability of paving work is achieved through flip paving, and spreading uniformity is improved through vibrating screens.

[0062] This invention achieves significant improvements in the following core dimensions:

[0063] 1. Improved processing efficiency: The traditional system has a manual assisted processing capacity of approximately 1.5 tons / hour per production line. However, the system of the present invention achieves a single-line processing capacity increase of 3 tons / hour through multi-manipulator collaborative operation, intelligent path planning and paving depth control, and an efficiency improvement of 100%.

[0064] 2. Enhanced automation and intelligence: The artificial intelligence of this system can use deep neural networks and image fusion algorithms to realize automatic recognition, path generation and dynamic grasping of garbage bags of different specifications and shapes. At the same time, it has the ability of self-assignment of tasks and trajectory correction, significantly reducing human intervention and having the ability to operate 24 hours a day.

[0065] 3. Strong environmental adaptability: Since there is no manual operation, a closed operation cabin design can be adopted. It can be deployed in high humidity, odor, and insect pest environments, avoiding the pollution diffusion problem caused by traditional open operations, and more in line with high-standard environmental protection site requirements.

[0066] 4. Excellent control of paving uniformity: Through the operation of the bionic manipulator, a 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%, which is suitable for subsequent air separation, optical separation, robot sorting and other process requirements.

[0067] 5. The subsequent sorting system has strong coordination capabilities: the paving path and speed are adjusted in real time by the AI ​​control system, and can be linked with photoelectric recognition, near-infrared sorting, airflow sorting and other technologies to achieve closed-loop optimization of the overall waste resource system.

[0068] 6. High system integration, better floor space and operation and maintenance costs: This invention integrates the three major processes of loading, bag breaking, and paving into a single control architecture and structural unit, reducing space occupancy by about 30% and energy loss by 40% compared to traditional four-stage equipment, with fewer later maintenance nodes and higher system operation stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 This is a structural diagram of the garbage loading, bag breaking and paving system based on the artificial intelligence bionic manipulator of the present invention;

[0070] Figure 2 This is a flow chart of the AI ​​intelligent control manipulator garbage feeding, bag breaking and paving method of the present invention;

[0071] Figure 3 This is a flow chart of the grabbing method in the AI ​​intelligent controlled manipulator garbage feeding, bag breaking and paving method of the present invention;

[0072] Figure 4 This is a flow chart of the paving method in the AI ​​intelligent control manipulator garbage loading and bag breaking paving method provided by the present invention;

[0073] Figure 5 This is a complete flow chart of the AI ​​intelligent garbage loading and bag breaking paving method provided by the present invention;

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

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

[0076] Figure 1 The accompanying drawings are numerals as follows:

[0077] 11. Conical hopper; 12. Truss; 13. First manipulator; 21. Working paving platform; 22. Conveying transmission structure; 23. Vibrating screen device; 3. Second manipulator; 4. Intelligent identification and path optimization device; 5. Third manipulator; 6. Fourth manipulator; 71. Diversion slope; 72. Drainage channel; 73. Filtration system; 74. Intelligent liquid level sensor; 75. Extraction pump. DETAILED DESCRIPTION

[0078] The following describes the embodiments of the present invention through specific examples. Those skilled in the art will readily understand the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention.

[0079] The following is combined with Figure 1-7 The present invention is described in further detail.

[0080] The embodiment of the present invention discloses a garbage loading, bag breaking and paving system and method using an artificial intelligence bionic manipulator.

[0081] Reference Figure 1 、 Figure 6 and Figure 7 A garbage loading and bag breaking paving system using an artificial intelligence bionic manipulator includes:

[0082] The loading device comprises a conical hopper 11 with a large top opening for dumping garbage, a truss 12 located above the conical hopper 11, and a first manipulator 13 slidably connected to the truss 12 via a flexible and telescopic arm;

[0083] The paving device includes an operating paving platform 21 located outside the conical hopper 11, a conveying and transmission structure 22 and a vibrating screen device 23 respectively connected to the operating paving platform 21, the truss 12 extends above the operating paving platform 21, the operating paving platform 21 includes a working platform and a paving surface above the working platform, the paving surface is a wide conveyor belt for conveying garbage, the conveying and transmission structure 22 drives the paving surface to move horizontally, and the vibrating screen device 23 drives the paving surface to vibrate, and the conveying and transmission structure 22 and the vibrating screen device 23 are both located below the paving surface;

[0084] A second manipulator 3 is provided on the side of the working platform 21;

[0085] Intelligent identification and path optimization device 4, including image acquisition device (see Figure 6 ), an AI intelligent control module electrically connected to the image acquisition device, the AI ​​intelligent control module (see Figure 7) includes an identification and analysis module, a path and motion optimization module connected to the identification and analysis module, a manipulator control module connected to the path and motion 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 laser radar and / or a near-infrared detector, the camera, the 3D laser radar and / or the near-infrared detector are arranged above the conical hopper and the working paving platform 21, and 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 and other three-dimensional information and the motion information of the first and second manipulators 3 collected in real time are transmitted to the AI ​​intelligent control module; the identification and analysis module includes an AI visual processing unit and an intelligent analysis unit, the AI ​​visual processing unit receives, identifies and processes the three-dimensional information and the motion information in real time, and transmits the processing results to the intelligent analysis unit, the intelligent analysis unit performs intelligent analysis on the garbage and the first and second manipulators 3 movements, the intelligent analysis includes detecting abnormalities such as foreign objects, large items, working errors, etc. according to the three-dimensional information of the garbage, extracting garbage bags The garbage bag characteristic information such as material information, volume information and density information is obtained, and the analysis results are sent to the path and motion optimization module, which includes a deep learning unit. Through deep learning, the tensile strength of the garbage bag is predicted, the appropriate clamping force is selected to prevent the garbage bag from breaking, the grabbing order is dynamically adjusted, the loading efficiency is improved, and high-density garbage bags are grabbed first to prevent blockage. Specifically, the module calculates the best grabbing point, grabbing force, grabbing method, clamping force, grabbing path, moving action on the truss 12, moving distance, bag breaking method and bag breaking action and optimized path of the first manipulator 13, the tearing and paving action of the second manipulator 3, and transmits them to the manipulator control module to control the automatic grabbing, clamping, sorting, clamping movement and bag breaking action of the first manipulator 13, the tearing, sorting and paving action of the second manipulator 3, the paving surface movement control module controls the paving surface movement to transfer the paved garbage according to the garbage condition on the working paving platform 21 analyzed by the recognition and analysis module, and also optimizes the manipulator path through the intelligent obstacle avoidance algorithm to avoid collision between manipulators.

[0086] Both the first manipulator 13 and the second manipulator 3 include a bionic structure with equal fingers on both hands, and the corresponding fingers of both hands can be crossed, preferably five fingers each. The hands of the first manipulator 13 are also respectively connected to arms that can be bent, extended and moved on the truss 12. Force feedback sensors are installed on the surface of the fingers 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 arm is attached with a multi-joint motion and rotation control structure to control the fingers and arms to perform bionic movements. The multi-joint motion and rotation control structure is integrated with a 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 to transmit the sensor signal to the recognition and analysis module. 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 multi-joint motion and rotation control structure to operate the motion of the first manipulator 13 and the second manipulator 3.

[0087] In one embodiment, the top of the conical hopper 11 is designed as a polyhedron, a circle or an ellipse, with a larger opening for dumping garbage. The top diameter of the conical hopper 11 (conical barrel) can be 4-12 meters, the cone height can be 4-10 meters, the cone angle can be 35°-50°, and the bottom garbage outlet width can be 1.5-3m. The truss 12 is located above the conical hopper 11 and the span extends to the outside of the conical hopper 11. The first manipulator 13 is slidably mounted thereon, and sufficient grasping space is provided for the first manipulator 13. The truss 12 is connected to the first manipulator 13 by an arm that is bendable, retractable and movable on the truss. The retractable arm allows the manipulator to contact the bottom center of the conical hopper 11 to grasp the garbage bag. The bendable arm allows the manipulator to adjust the angle to grasp the target garbage bag, completing the grasping, clamping and clamping movement loading operations.

[0088] The working paving platform 21 is located on the outside of the conical hopper 11, and the truss 12 extends to the top of the working paving platform 21. The working paving platform 21 is used to carry the garbage after the first manipulator 13 breaks the bag. After the first manipulator 13 completes the grabbing and clamping action, it reaches the top of the working paving platform 21 through the movement of the arm on the truss 12, and then performs the bag breaking operation. After the garbage bag is broken, the internal garbage falls freely to the working paving platform 21 under the action of gravity. The paving surface on the working paving platform 21 is a wide conveyor belt. The wide surface can spread more garbage horizontally to prevent garbage accumulation. The paving surface is a working surface controlled by the conveying transmission structure 22 to move horizontally for garbage transportation. The wide surface can also prevent garbage from being scattered outside the area of ​​the working paving platform 21. The vibrating screen device 23 drives the paving surface to vibrate to avoid garbage stacking and achieve better paving effect. The vibrating screen device 23 is a vibration drive mechanism and can adopt various types of vibration drive devices in the existing technology.

[0089] Figure 1 The second manipulators 3 are fixedly arranged at intervals on both sides of the working platform of the working paving platform 21, and are used to sort recyclable garbage on the one hand, and to tear the garbage bags that are not completely broken to optimize the paving effect on the other hand;

[0090] Figure 6 In the figure, 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, wherein the image acquisition device includes a camera, a 3D laser radar and / or a near-infrared detector. The camera is used to capture 2D images through visible light and record visual information such as color and texture. The 3D laser radar is used to emit laser pulses, measure distance through reflection time, generate high-precision 3D point clouds, and then obtain spatial coordinates and distance. The near-infrared detector is used to detect object characteristics, such as thermal radiation, material reflection, etc., and output grayscale images or specific spectral data. The camera acquires 2D images, and the laser radar and / or near-infrared detector acquires 3D images. Combining them can obtain more comprehensive image information. The camera, 3D laser radar and / or near-infrared detector are arranged above the conical hopper 11 and the working paving platform 21, and transmits the three-dimensional information of garbage in the conical hopper and on the working paving platform 21, including garbage accumulation conditions, garbage types, etc., and the action information of the first and second manipulators 13 and 3, which are collected in real time to the AI ​​intelligent control module.

[0091] Figure 7In the embodiment, the AI ​​intelligent control module includes a recognition and analysis module, a path and motion optimization module connected to the recognition and analysis module, a manipulator control module connected to the path and motion optimization module, and a paving surface movement control module electrically connected to the recognition and analysis module. The recognition and analysis module includes an AI visual processing unit and an intelligent analysis unit. The AI ​​visual processing unit receives, recognizes and processes the three-dimensional information and the motion information in real time, and transmits the processing results to the intelligent analysis unit. The intelligent analysis unit performs intelligent analysis on the garbage and the first and second manipulators 13 and 3. The intelligent analysis includes discovering abnormalities such as foreign matter, large items, working errors, etc. based on the three-dimensional information of the garbage, extracting garbage bag feature information such as material information, volume information and density information of the garbage bag, and sending the analysis results to the path and motion optimization module. The path and motion optimization module includes a deep learning unit, which calculates the best grasping point, grasping force, grasping method, grasping path, moving action on the truss 12, and moving distance of the first manipulator 13 through deep learning. , bag breaking method and bag breaking action and optimized path, the tearing and paving action of the second manipulator 3, and transmit it to the manipulator control module to control the grabbing, clamping, clamping movement and bag breaking action of the first manipulator 13, the tearing, sorting and paving action of the second manipulator 3, the paving surface movement control module controls the movement of the paving surface to transfer the paved garbage according to the garbage status on the working paving platform 21 analyzed by the recognition and analysis module, and obtains the garbage bag feature information through analysis of the collected image information, and then calculates the corresponding manipulator action, motion path and operation mode through deep learning. The modules and units are closely connected and cooperate with each other to carry out grabbing, clamping, clamping movement, bag breaking action sorting, and paving action in sequence to achieve reliable feeding, sorting and paving operations, and also optimize the manipulator path through the intelligent obstacle avoidance algorithm to avoid collisions between manipulators, thereby improving the fault tolerance of the system and solving the problems in the existing technology such as garbage bag accumulation affecting processing efficiency, single bag breaking method, uneven bag breaking effect for different plastic garbage bags, manual assisted paving, high labor intensity and harsh working environment.

[0092] Figure 1 In the embodiment, the first manipulator 13 and the second manipulator 3 both include a bionic structure with two equal fingers, and the corresponding fingers of the two hands can be crossed, preferably five fingers each, and a force feedback sensor is installed on the surface of the finger to monitor the tension of the garbage bag in real time and transmit the signal to the motion and rotation control module. Each finger is a multi-joint structure, and each finger and arm is attached with a multi-joint motion and rotation control structure to control the fingers and arms to perform bionic movements. The multi-joint motion and rotation control structure is integrated with the motion and rotation control module (see Figure 7); the motion and rotation control module is connected to the recognition and analysis module and the manipulator control module respectively, and transmits the sensor signal to the recognition and analysis module. 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 multi-joint motion and rotation control structure to operate the first and second manipulators 13 and 3, which can cope with various environments and improve the flexibility of operation.

[0093] See also Figure 1 、 Figure 6 and Figure 7 The path and optimization module also includes an intelligent grabbing mode unit, which stores different grabbing modes of the first manipulator 13, and determines that the first manipulator 13 adopts different intelligent grabbing modes according to the identification and analysis of the garbage by the identification and analysis module. The intelligent grabbing modes include a clamping mode for rigid garbage bags, hard plastic bags, etc.; a flexible grabbing mode for preventing kitchen waste bags from being damaged; an adsorption grabbing mode for light garbage, paper, plastic bags, expanded polystyrene, etc.; the fingertips of the first manipulator 13 are equipped with high-frequency vibration blades, and the multi-joint action and rotation control structure includes a vibration drive structure (see Figure 7 ), the motion and rotation control module includes a vibration drive control for the vibration drive structure, and the path and optimization module also includes a bag breaking method unit ( Figure 7 ), stores different bag breaking methods of the first manipulator 13, and judges that the first manipulator 13 adopts different bag breaking methods according to the identification and analysis of the garbage by the identification and analysis module, thereby improving the flexibility of the bag breaking work. The manipulator control module transmits the bag breaking method control instruction 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 and breaking bags, high-frequency vibration and twisting and breaking bags.

[0094] exist Figure 7 In the example, the intelligent grabbing mode unit stores different grabbing modes of the first manipulator 13, and determines whether the first manipulator 13 adopts different intelligent grabbing modes according to the identification and analysis of the garbage by the identification and analysis module, thereby improving the flexibility of the grabbing work.

[0095] See also Figure 1 、 Figure 6 and Figure 7The cross section of the conical hopper 11 is multifaceted, circular or elliptical, and the angle of the cone is 15°-90° with the vertical line. The inner wall is coated with wear-resistant and corrosion-resistant material, and the truss 12 spans the side wall area of ​​the conical hopper, that is, the second side wall is lower than the other side wall parts, that is, the first side wall; it also includes a third manipulator 5, which is arranged on the inside of the conical hopper to remove the foreign matter, large pieces and other abnormalities. The third manipulator 5 includes a bionic structure with equal fingers on both hands, and the corresponding fingers of both hands can be crossed, preferably with five fingers each. A force feedback sensor is installed on the surface of the finger, and each finger is a multi-joint structure. Each finger is attached with a multi-joint action and rotation control structure. Figure 7 To control the bionic movements of the fingers, a motion and rotation control module is integrated on the multi-joint motion and rotation control structure. The force feedback sensor monitors the tension of the garbage bag in real time and transmits the sensor signal to the recognition and analysis module. The motion and rotation control module is connected to the recognition and analysis module and the manipulator control module. The manipulator control module transmits the motion and 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 third manipulator 5 to remove foreign objects and large items.

[0096] exist Figure 1 The cross-section of the conical hopper 11 is circular or elliptical to avoid dead angles in the grabbing position and improve the reliability of the system. The inclined conical design of the side wall allows the garbage bag to roll freely into the bottom center of the conical hopper 11. The angle range of the cone is 15°-90° to the vertical line. The inner wall is coated with wear-resistant and corrosion-resistant material, which increases durability on the one hand and improves the smoothness of the inner wall on the other hand. The third manipulator 5 is arranged on the inside of the conical hopper to remove foreign matter, large pieces and other abnormalities. The startup sequence of the third manipulator 5 takes precedence over the first manipulator 13 and the second manipulator 3. Foreign matter and large pieces are removed before the loading operation, further improving the reliability of the system.

[0097] See also Figure 1 、 Figure 6 and Figure 7 The truss 12 on the working paving platform 21 is further provided with a fourth manipulator 6, which includes a bionic structure with equal fingers on both hands and crossable corresponding fingers on both hands, preferably 5 fingers each, with force feedback sensors installed on the surface of the fingers, each finger is a multi-joint structure, and each finger is attached with a multi-joint action and rotation control structure, see Figure 7, to control the fingers to perform bionic movements, the multi-joint movement and rotation control structure is integrated with a movement and rotation control module, the force feedback sensor monitors the garbage tension in real time and transmits the sensor signal to the recognition and analysis module, the movement and rotation control module is connected to the recognition and analysis module and the manipulator control module respectively, the manipulator control module transmits the plucking movement and optimized path control instructions optimized by the path and movement optimization module to the movement and rotation control module to control the multi-joint movement and rotation control structure to operate the plucking movement of the fourth manipulator 6.

[0098] exist Figure 1 In the embodiment, the plucking action of the fourth manipulator 6 can realize fast sorting operation, 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 also Figure 1 、 Figure 6 and Figure 7 , the path and optimization module also includes an intelligent flip paving unit, see Figure 7 The unit stores the flipping and paving mode of the first manipulator 13, and starts the flipping and paving of the first manipulator 13 based on the recognition and analysis of the broken bag status 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 motion optimization module to the motion and rotation control module, and controls the first manipulator 13 to flip 180° after breaking the bag through the multi-joint motion and rotation control structure, so as to evenly spread the garbage on the conveyor belt.

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

[0101] Reference Figure 1 、 Figure 6 and Figure 7 The first, second, third and fourth manipulators 6 are one or more, and the path and optimization module also includes a manipulator coordination unit to coordinate the grabbing, clamping, moving, bag breaking, tearing, sorting, spreading, rejecting and plucking actions of each manipulator; the manipulator coordination unit adopts an intelligent reinforcement learning (RL) load balancing algorithm to dynamically allocate grabbing tasks according to garbage flow and optimize throughput; the manipulator coordination unit also includes a synchronous collaboration mode, which includes multiple first manipulators 13 pulling and breaking bags at the same time and different manipulators taking charge of different tasks to collaborate synchronously.

[0102] In this embodiment, the robot coordination unit utilizes an intelligent reinforcement learning (RL) load balancing algorithm, which can sense system state changes in real time. It dynamically allocates grabbing tasks and adjusts strategies based on garbage flow, adapting to complex environments such as load fluctuations and heterogeneous resources, offering significant advantages. It can also employ the A algorithm combined with the Dijkstra algorithm to calculate the shortest path, reduce motion paths, and minimize robot conflicts.

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

[0104] The feeding device, paving device, various manipulators, and intelligent identification and path optimization device 4 of the present invention are all arranged in a closed chamber and are located in the same space. Environmental problems such as secondary dust, odor diffusion, and mosquito breeding are avoided, and the cleanliness of operations and environmental protection standards are significantly improved, meeting high-demand scenarios at home and abroad (such as underground transfer, wet garbage treatment centers, etc.). The modular partitioning and functional integration of the closed structure can constitute an independent structural protection point. The vibrating blade adopts a low-frequency vibration mode of 30-50Hz to ensure the reliability of bag breaking with high impact force and adaptability; the low-frequency vibration cycle is longer, allowing the blade to fully reset between two vibrations, ensuring the continuity and stability of the cutting action. The garbage bag material is diverse, and low-frequency vibration can be adapted 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. Low-frequency vibration has lower power requirements for the drive motor, reducing energy consumption; at the same time, the vibration inertia is small, reducing fatigue loss of the structure and extending the service life;

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

[0106] See also Figure 1A garbage leachate collection system is connected to the bottom of the conical garbage hopper 11, and 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 with an inclination of 2°-15° to the bottom of the conical hopper 11. The diversion slope 71 is connected to the drainage channel 72, and 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 the metal grid is coated with a polymer permeable filter layer. The automatic extraction system includes an intelligent liquid level sensor 74 arranged at the bottom of the conical hopper 11, an extraction pump 75 connected to the drainage channel 72, and a sewage treatment system connected to the end of the drainage channel 72. The leachate flows from the diversion slope 71 to the drainage channel 72.

[0107] exist Figure 1 In the embodiment, a diversion slope 71 with a 2°-15° inclination can be provided at the bottom of the conical hopper 11. The 2°-15° is a gentle slope, and the garbage bag will not change its stacking shape, 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 the metal grid is coated with a polymer permeable filter layer 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 clean the garbage leachate.

[0108] The intelligent identification and path optimization device 4 of the present invention can adopt an autonomously trained intelligent path planning model, combined with image recognition, grasping point prediction and behavior evaluation mechanism, to realize autonomous classification judgment and dynamic paving strategy for irregular garbage materials. The AI ​​algorithm not only ensures the accuracy of paving uniformity and sequence, but also has the function of self-correction of faults and optimization of working behavior. It can constitute behavior control logic protection, parameter control range protection and AI model training framework protection. Each bionic manipulator has the ability of multi-degree-of-freedom linkage, simulating the actions of human hands such as grasping, rotating, tearing, and unfolding; and is equipped with force feedback and dynamic grasping adjustment modules to ensure that the bag breaking action is both efficient and avoids damaging the recyclables inside. It has the dual advantages of structure and motion control, and has high stability and adaptability to industrial implementation.

[0109] See also Figure 2 , AI intelligent control robot loading bag breaking and paving methods include:

[0110] S1. Intelligent loading step: Using AI vision to automatically identify the three-dimensional information of the garbage dumped into the conical hopper, including the accumulation 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. Efficient bag grabbing and bag breaking by the manipulator: The AI ​​automatically adjusts the bionic structure of the manipulator based on the three-dimensional information of the garbage, crosses the ten fingers of the manipulator's hands to grab the garbage bag, 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 of both hands can be crossed. Preferably, there are five fingers on each hand. A force feedback sensor is installed on the surface of each finger to monitor the tension of the garbage bag in real time and transmit a signal to the AI ​​to adjust the method and strategy of grabbing and bag breaking in real time;

[0112] S3. Robotic arm uniform spreading step: Use a bionic structure robot to evenly spread the broken garbage bags.

[0113] S1 The dumping means that the garbage truck dumps the garbage into a conical hopper 11 with a larger top opening. The cross section of the conical hopper 11 is polygonal, circular or elliptical, and the angle of the cone ranges from 15° to 90° with the vertical line. The inner wall is coated with wear-resistant and anti-corrosion material;

[0114] The manipulator S2 is a first manipulator 13. The hands of the first manipulator 13 also include bendable and retractable arms connected to the hands respectively. The other end of the arm is connected to the truss 12 provided on the conical hopper 11 for controllable movement. The first manipulator 13 is a plurality of collaborative operations to optimize task allocation. A third manipulator 5 is also provided in the conical hopper 11 to handle abnormalities such as foreign objects, large parts, and working errors discovered by AI intelligent analysis.

[0115] The manipulator in S3 is the second manipulator 3, the conveyor belt is a wide paving surface provided on the working paving platform 21 outside the conical hopper 11, the truss 12 extends above the paving surface, and the second manipulator 3 is a plurality of collaborative operations to optimize task allocation;

[0116] The paving described in S3 also includes automatically adjusting the angle and force of the bionic structure first manipulator 13 feeding the material to the conveyor belt after breaking the bag through AI artificial intelligence, as well as the collaborative operation of multiple first and second manipulators 3.

[0117] Reference Figure 3 、 Figure 4 , AI intelligent control robot loading bag breaking paving method,

[0118] The crawling of S2 includes the following steps:

[0119] S21. Adjusting the grabbing order of the robot arm based on the volume information and density information in the garbage bag feature information;

[0120] S22. Adjust the gripping mode of the manipulator based on the material information in the garbage bag feature information. The gripping mode is for rigid garbage bags, hard plastic bags, etc.; the flexible gripping mode is for kitchen waste bags to prevent damage; and the adsorption gripping mode is for lightweight garbage, such as paper, plastic bags, and expanded polystyrene.

[0121] S23, calculate the optimal grasping point and path through deep learning (CNN+Transformer), optimize the robot path through intelligent obstacle avoidance algorithm, and avoid collisions between robots;

[0122] S24, grabbing garbage in sequence according to the grabbing order, grabbing mode, optimal grabbing point and optimal grabbing path;

[0123] S25, determining the toughness of the garbage bag according to the pressure sensor and comparing it with a preset toughness value;

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

[0125] S26. Adopt a corresponding bag-breaking scheme based on the toughness of the garbage bag. When the pressure sensor determines that the toughness of the garbage bag is greater than a preset toughness value, high-frequency vibration is used to break the bag, through micro-vibration cutting. When the pressure sensor determines that the toughness of the garbage bag is less than the preset toughness value, the garbage bag is grasped at both ends with both hands and rotated in opposite directions to tear the ends, forming a twisting tear to complete the bag-breaking operation.

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

[0127] S31, turning over the broken garbage bags and spreading them on the conveyor belt;

[0128] S32, automatically adjusts the feeding angle and force to spread the material evenly, prevents garbage accumulation and improves screening efficiency;

[0129] S33, using a 50-200Hz vibration conveyor belt to prevent garbage from clumping and improve paving uniformity;

[0130] S34. Sorting recyclable garbage after spreading.

[0131] Reference Figure 5 , further comprising the following steps:

[0132] S4: Optimize the task allocation of multiple robots through intelligent scheduling algorithms to improve waste processing throughput;

[0133] S5. A slightly inclined (2°-15°) leachate diversion slope 71 is installed at the bottom of the conical hopper 11 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 to automatically start the extraction pump 75 when the leachate reaches the set value to discharge the leachate into the sewage treatment system. The leachate pipeline uses an anti-clogging spiral sewage pipe.

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

[0135] Control Target Identification accuracy / control error range Trash bag grabbing point identification deviation ≤±5cm Crawl success rate ≥95.8％ Tear path control deviation ≤±10° Paving path trajectory offset error ≤3cm Control latency (end-to-end processing time) ≤80ms

[0136] Table 1

[0137] Therefore, the system of the present invention still has high recognition rate, high motion accuracy and stable operation capability in a highly complex material environment, which constitutes the core advantage and technical barrier of the present invention compared with traditional equipment.

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

[0140]

[0141] Table 2

[0142] In summary, compared with the existing traditional garbage bag spreading system, the present invention has shown significant improvements in processing efficiency, degree of automation, energy consumption control and recycling effect, and has significant technological advancement and industrial application prospects.

Claims

1. A garbage loading and bag breaking paving system based on an artificial intelligence bionic manipulator, comprising: The loading device comprises a conical hopper with a large top opening for dumping garbage, a truss located above the conical hopper, and a first manipulator slidably connected to the truss via a bendable and telescopic arm; The paving device comprises an operating paving platform located outside the conical hopper, a conveying transmission structure and a vibrating screen device respectively connected to the operating paving platform, the truss extending to above the operating paving platform, the operating paving platform comprising an operating table and a paving surface above the operating table, the paving surface being a wide conveyor belt for conveying garbage, the conveying transmission structure driving the paving surface to move horizontally, the vibrating screen device driving the paving surface to vibrate, and the conveying transmission structure and the vibrating screen device both being located below the paving surface; A second manipulator provided on the side of the work platform of the working paving platform; An intelligent recognition and path optimization device includes an image acquisition device and an AI intelligent control module electrically connected to the image acquisition device. The AI ​​intelligent control module includes a recognition and analysis module, a path and motion optimization module connected to the recognition and analysis module, a manipulator control module connected to the path and motion optimization module, and a paving surface movement control module electrically connected to the recognition and analysis module. The image acquisition device includes a camera, a 3D laser radar and / or a near-infrared detector. The camera, the 3D laser radar and / or the near-infrared detector are arranged above the conical hopper and the working paving platform, and transmit the three-dimensional information of garbage in the conical hopper and on the working paving platform, including garbage accumulation conditions, garbage types, etc., and the motion information of the first and second manipulators collected in real time to the AI ​​intelligent control module; the recognition and analysis module includes an AI visual processing unit and an intelligent analysis unit. The AI ​​visual processing unit receives, recognizes and processes the three-dimensional information and the motion information in real time, and transmits the processing results to the intelligent analysis unit. The intelligent analysis unit performs intelligent analysis on the garbage and the movements of the first and second manipulators. The intelligent analysis includes detecting anomalies such as foreign objects, large items, and working errors based on the three-dimensional information of the garbage, extracting characteristic information of the garbage bag such as material information, volume information, and density information of the garbage bag, and sending the analysis results to the path and motion optimization module. The path and motion optimization module includes a deep learning unit, which calculates the grasping point, grasping force, grasping method, clamping force, movement on the truss, movement distance, bag breaking method and bag breaking action and optimized path of the first manipulator through deep learning, and the tearing, sorting, and paving actions of the second manipulator, and transmits them to the manipulator control module to control the grasping, 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 transfer the paved garbage according to the garbage status on the working paving platform analyzed by the recognition and analysis module, and also optimizes the manipulator path through the intelligent obstacle avoidance algorithm to avoid collisions between manipulators. The first manipulator and the second manipulator each include a bionic structure having two equal fingers, and the corresponding fingers of the two hands can be crossed. The hands of the first manipulator are also respectively connected to arms that can bend, stretch and move on a truss. Force feedback sensors are installed on the surface of the fingers 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. Each finger and arm is attached with a multi-joint motion and rotation control structure to control the fingers and arms to perform bionic movements. The motion and rotation control module is integrated into the multi-joint motion and rotation control structure. The motion and rotation control module is connected to the recognition and analysis module and the manipulator control module respectively, and transmits the sensor signal to the recognition and analysis module. 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 multi-joint motion and rotation control structure to operate the action of the first and second manipulators.

2. The garbage loading and bag breaking and paving system based on artificial intelligence bionic manipulator according to claim 1 is characterized in that: The path and motion optimization module also includes an intelligent grasping mode unit, which stores different grasping modes of the first manipulator, and determines that the first manipulator adopts different intelligent grasping modes based on 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 kitchen waste bags from being damaged; and an adsorption grasping mode for lightweight garbage, paper, plastic bags, expanded polystyrene, etc. The finger end of the first manipulator is equipped with a high-frequency vibrating blade, and the multi-joint motion and rotation control structure includes a vibration drive structure, and the motion and rotation control module includes a vibration drive control of the vibration drive structure; the path and motion optimization module also includes a bag breaking method unit, which stores different bag breaking methods of the first manipulator, and determines that the first manipulator adopts different bag breaking methods based on the recognition and analysis of the garbage by the recognition and analysis module. The manipulator control module transmits the bag breaking method control instruction 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 and breaking bags, high-frequency vibration and twisting and breaking bags.

3. The garbage loading bag breaking and paving system based on artificial intelligence bionic manipulator according to claim 2 is characterized in that The cross section of the conical hopper is a polyhedron, a circle or an ellipse, and the angle range of the cone is 15°-90° with the vertical line. The inner side wall is coated with wear-resistant and corrosion-resistant material, and the side wall area of ​​the truss spanning the conical hopper, that is, the second side wall, is lower than the other side wall part, that is, the first side wall; it also includes a third manipulator, which is arranged on the inner side of the conical hopper to remove the foreign matter, large pieces and other abnormalities. The third manipulator includes a bionic structure with equal fingers on both hands, and the corresponding fingers of both hands can be crossed. A force feedback sensor is installed on the surface of the finger, and each finger is a multi-joint structure. Each finger is attached There is a multi-joint motion and rotation control structure to control the bionic motion of the fingers. The multi-joint motion and rotation control structure is integrated with a motion and rotation control module. The force feedback sensor monitors the tension of the garbage bag in real time and transmits the sensor signal to the identification and analysis module. The motion and rotation control module is connected to the identification and analysis module and the manipulator control module. The manipulator control module transmits the motion and 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 third manipulator to remove foreign objects and large items.

4. The garbage loading bag breaking and paving system based on artificial intelligence bionic manipulator according to claim 3 is characterized in that A fourth manipulator is also provided on the truss on the working paving platform. The fourth manipulator includes a bionic structure with equal fingers on both hands and corresponding fingers on both hands that can be crossed. A force feedback sensor is installed on the surface of the fingers. Each finger is a multi-joint structure. Each finger is attached with a multi-joint action and rotation control structure to control the fingers to perform bionic actions. The multi-joint action and rotation control structure is integrated with a motion and rotation control module. The force feedback sensor monitors the garbage tension in real time and transmits the sensor signal to the identification and analysis module. The motion and rotation control module is connected to the identification and analysis module and the manipulator control module respectively. The manipulator control module transmits the plucking 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 action and rotation control structure to operate the plucking action of the fourth manipulator.

5. The garbage loading and bag breaking and paving system based on artificial intelligence bionic manipulator according to claim 4 is characterized in that: 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 based on the recognition and analysis of the broken bag status 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 motion optimization module to the motion and rotation control module, and controls the first manipulator to flip 180° after breaking the bag through the multi-joint motion and rotation control structure, so as to evenly spread the garbage on the conveyor belt.

6. The garbage loading bag breaking and paving system based on artificial intelligence bionic manipulator according to claim 5 is characterized in that The first, second, third and fourth manipulators are one or more, and the path and optimization module also includes a manipulator coordination unit to coordinate the grabbing, clamping, moving, bag breaking, tearing, sorting, spreading, rejecting and plucking actions of each manipulator; the manipulator coordination unit adopts an intelligent reinforcement learning (RL) load balancing algorithm to dynamically allocate grabbing tasks according to garbage flow and optimize throughput; the manipulator coordination unit also includes a synchronous collaboration mode, which includes multiple first manipulators pulling and breaking bags at the same time and different manipulators taking charge of different tasks to collaborate synchronously, and automatically selecting the optimal bag breaking mode by detecting the toughness of the garbage bag through the force feedback sensor.

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

8. The garbage loading and bag breaking and paving system based on artificial intelligence bionic manipulator according to claim 7 is characterized in that: A garbage leachate collection system is connected to the bottom of the conical garbage hopper, and 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 is inclined 2°-15° with the bottom of the conical hopper. The diversion slope is connected to the drainage channel, and the drainage channel is an anti-clogging spiral sewage pipe. A metal grid is installed at the bottom of the conical hopper, and the metal grid is coated with a polymer permeable filter layer. The automatic extraction system includes an intelligent liquid level sensor arranged at the bottom of the conical hopper, an extraction pump connected to 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.

9. AI intelligent control manipulator garbage feeding bag breaking and paving method, which is characterized by include: S1. Intelligent loading step: Use AI vision to automatically identify the three-dimensional information of the garbage dumped into the conical hopper, including the accumulation situation, garbage type, and garbage bag characteristics such as material information, volume information, and density information; S2. Efficient bag grabbing and bag breaking by the manipulator: The AI ​​automatically adjusts the bionic structure of the manipulator based on the three-dimensional information of the garbage, crosses the corresponding fingers of the two hands to grab the garbage bag, and moves the garbage bag to the conveyor belt for bag breaking. The manipulator includes a bionic structure with two equal fingers, and the corresponding fingers of the two hands can be crossed. 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 ​​to adjust the grabbing and bag breaking methods and strategies in real time; S3. Robotic arm uniform spreading step: Use a bionic structure robot to evenly spread the broken garbage bags.

10. The AI ​​intelligent control manipulator garbage feeding bag breaking and paving method according to claim 9 is characterized by: The cross section of the conical hopper S1 is multifaceted, circular or elliptical, the angle of the cone is in the range of 15°-90° to the vertical line, and the inner wall is coated with wear-resistant and corrosion-resistant material; The manipulator S2 is a first manipulator, and the hands of the first manipulator further include bendable and retractable arms connected to the hands respectively, and the other end of the arm is controllably connected to the truss provided on the conical hopper. The first manipulator is a plurality of collaborative operations to optimize task allocation; a third manipulator is also provided in the conical hopper to handle abnormalities such as foreign objects, large parts, and working errors discovered according to AI intelligent analysis; S3: The manipulator is a second manipulator, the conveyor belt is a wide paving surface arranged on the working paving platform outside the conical hopper, the truss extends above the paving surface, and the second manipulator is a plurality of collaborative operations to optimize task allocation; The paving in S3 also includes automatically adjusting the angle and force of the bionic structure first manipulator feeding the material to the conveyor belt after breaking the bag through AI artificial intelligence, and the coordinated operation of multiple first and second manipulators; The crawling of S2 includes the following steps: S21. Adjusting the grabbing order of the robot arm based on the volume information and density information in the garbage bag feature information; S22. Adjust the gripping mode of the manipulator based on the material information in the garbage bag feature information. The gripping mode is for rigid garbage bags, hard plastic bags, etc.; the flexible gripping mode is for kitchen waste bags to prevent damage; and the adsorption gripping mode is for lightweight garbage, such as paper, plastic bags, and expanded polystyrene. S23, calculate the optimal grasping point and path through deep learning (CNN+Transformer), optimize the robot path through intelligent obstacle avoidance algorithm, and avoid collisions between robots; S24, grabbing garbage in sequence according to the grabbing order, grabbing mode, optimal grabbing point and optimal grabbing path; S25, determining the toughness of the garbage bag according to the pressure sensor and comparing it with a preset toughness value; The bag breaking step of S2 includes the following steps: S26. Adopt a corresponding bag-breaking scheme based on the toughness of the garbage bag. When the pressure sensor determines that the toughness of the garbage bag is greater than a preset toughness value, high-frequency vibration is used to break the bag, through micro-vibration cutting. When the pressure sensor determines that the toughness of the garbage bag is less than the preset toughness value, the garbage bag is grasped at both ends with both hands and rotated in opposite directions to tear the ends, forming a twisting tear to complete the bag-breaking operation. The paving of S3 also includes the following steps: S31, turning over the broken garbage bags and spreading them on the conveyor belt; S32, automatically adjusts the feeding angle and force to spread the material evenly, prevents garbage accumulation and improves screening efficiency; S33, using a 50-200Hz vibration conveyor belt to prevent garbage from clumping and improve paving uniformity; S34, sorting the recyclable waste after paving; further comprising the following steps: S4: Optimize the task allocation of multiple robots through intelligent scheduling algorithms to improve waste processing throughput; S5. A slightly inclined (2°-15°) leachate diversion slope is installed at the bottom of the conical hopper to direct the leachate to the designated drainage channel; a metal grid + polymer permeable filter layer is installed at the bottom of the conical hopper to intercept solid waste; an intelligent liquid level sensor is used to automatically start the extraction pump when the leachate reaches the set value to discharge the leachate into the sewage treatment system; the leachate pipeline uses an anti-clogging spiral sewage pipe.

Images