Cooperative control planning method of mechanical arm and conveying belt for intelligent garbage sorting

Abstract

The application relates to the technical field of intelligent garbage treatment and automatic sorting, and discloses a mechanical arm and conveying belt cooperative control planning method for intelligent garbage sorting, which comprises the following steps: controlling the conveying belt to convey garbage to a garbage bag breaking device at a set speed for bag breaking treatment, controlling the conveying belt to slow down or stop until the garbage is completely subjected to the bag breaking treatment, controlling the conveying belt to convey the garbage to a metal detection device for metal detection, identifying the garbage type, determining a garbage grabbing mode and a path, grabbing the garbage, and classifying and conveying the garbage into a designated garbage can. The technical scheme of the mechanical arm and the conveying belt cooperative control achieves efficient and accurate garbage sorting effect; compared with the sorting system relying on the mechanical arm or the conveying belt alone in the prior art, the application realizes the cooperative action of the mechanical arm and the conveying belt, so that the garbage sorting process is more smooth, and the problem of low processing efficiency caused by asynchronization of the system is solved.

Description

Technical Field

[0001] This invention relates to the field of intelligent waste treatment and automated sorting technology, specifically to a collaborative control planning method for robotic arms and conveyor belts used in intelligent waste sorting. Background Technology

[0002] With the acceleration of urbanization, waste disposal has become a crucial issue in environmental protection and resource recycling. Traditional waste sorting methods rely on manual operation or simple mechanical sorting equipment, which are not only inefficient but also prone to errors. In recent years, intelligent waste sorting technology has gradually emerged, especially in large-scale industrial waste recycling, where automated and intelligent waste sorting systems have demonstrated enormous potential.

[0003] Existing intelligent waste sorting systems primarily rely on visual recognition technology combined with robotic arm grasping technology to achieve waste classification and processing. The visual recognition module uses cameras or sensors to photograph the waste and analyze the images to identify the type of waste. Subsequently, the robotic arm grasps the waste based on the recognition results and places it into the designated recycling unit.

[0004] However, most existing intelligent waste sorting technologies rely on conveyor belts with fixed speeds, making it difficult to dynamically adjust according to the actual weight and size of the waste. This easily leads to lag in sorting heavy items and missorting of light items. Furthermore, their reliance on two-dimensional visual recognition technology makes it difficult to accurately acquire the three-dimensional spatial information of the waste, which makes the robotic arm prone to deviations during the grasping process. Moreover, the robotic arm and conveyor belt often operate independently, making it difficult for the system to adapt and adjust in a timely manner when the type of waste changes. Therefore, this invention provides a collaborative control planning method for robotic arms and conveyor belts in intelligent waste sorting to address the shortcomings of existing technologies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a collaborative control planning method for robotic arms and conveyor belts in intelligent waste sorting. This method solves the problem that existing waste sorting robotic arms and conveyor belts often operate independently, making it difficult to adapt and adjust in a timely manner when the type of waste changes.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a collaborative control planning method for a robotic arm and conveyor belt for intelligent waste sorting, comprising the following steps:

[0007] When the sensor detects that the initial garbage has been dumped onto the conveyor belt, the conveyor belt is controlled to transport the garbage to the garbage bag breaking device at a set speed for bag breaking.

[0008] When the garbage is transported to the bag breaking device, control the conveyor belt to slow down or stop until the garbage is completely broken;

[0009] After the garbage bags are broken open and the garbage is spread out, the control conveyor belt will transport the garbage to the metal detection device for metal detection;

[0010] After metal detection is completed, the control conveyor belt will transport the garbage to the visual recognition area. The 3D camera will be used to reconstruct the garbage in three dimensions, identify the type of garbage, and determine the garbage grabbing method and path.

[0011] The robotic arm is controlled to grab the garbage according to the garbage grabbing method and path determined by the recognition results, and then sort and transport the garbage to the designated garbage bin.

[0012] Preferably, the speed control of the conveyor belt includes the following steps:

[0013] The weight of the waste dumped on the conveyor belt is detected in real time by a weight sensor;

[0014] Determine if the weight exceeds the set threshold. If it does, reduce the conveyor belt speed to the set slow speed. If it does not exceed the threshold, run at the base conveyor belt speed.

[0015] During the waste conveying process, the conveyor belt speed is dynamically adjusted by combining the speed sensor data at both ends of the conveyor belt.

[0016] Preferably, the deceleration or pausing control of the conveyor belt includes the following steps;

[0017] The conveyor belt uses sensors in the bag-breaking device to monitor whether the waste has reached the bag-breaking area;

[0018] Once the waste reaches the bag-breaking area, the conveyor belt automatically switches to deceleration mode until the bag-breaking device detects that the waste has been broken open.

[0019] After the garbage bags are broken open, the conveyor belt returns to the set speed and continues to transport the garbage.

[0020] Preferably, the three-dimensional reconstruction of the waste using a 3D camera includes the following steps:

[0021] The garbage is captured from multiple perspectives using a 3D camera, and the three-dimensional feature points of the garbage are extracted using a multi-view image reconstruction algorithm.

[0022] Garbage feature images are classified by combining a deep learning model, wherein the deep learning model is either ResNet50 or MobileNet.

[0023] In the process of image feature extraction, attention mechanism is used to extract features from salient regions of garbage feature images;

[0024] Based on the identification results, the grabbing method and path for each type of waste are calculated.

[0025] Preferably, the identification of the type of waste includes the following steps:

[0026] The garbage is imaged from multiple perspectives using a 3D camera, and features are extracted from the garbage images based on deep learning algorithms to extract the geometric shape, color, texture and material properties of the garbage.

[0027] By combining convolutional neural network models and support vector machine classification methods, waste types are classified, and the grabbing points and optimal disposal locations for waste are calculated.

[0028] Based on the identified waste type, adjust the robotic arm's target grasping strategy, including selecting appropriate adsorption, clamping, or grasping modes;

[0029] After the garbage sorting is completed, the optimal path of the robotic arm is calculated, and the grasping trajectory is optimized by combining obstacle avoidance algorithm.

[0030] Preferably, the robotic arm's grasping and deployment includes the following steps:

[0031] The robotic arm automatically adjusts its gripping force based on the type of waste identified by the 3D camera;

[0032] The robotic arm's rotation angle, gripping point, and gripping force are dynamically adjusted to handle waste of different shapes and materials.

[0033] The robotic arm's grasping trajectory employs a shortest path planning method based on Dijkstra's algorithm, combined with a dynamic obstacle avoidance algorithm to optimize the grasping path.

[0034] Preferably, the Dijkstra algorithm establishes a distance map between the starting position of the robotic arm and the target grasping point, calculates the shortest path from the starting point to the ending point based on the weights between nodes, and dynamically adjusts the path; during the calculation process, the rotation angle of the robotic arm, the grasping posture, and obstacle avoidance requirements are taken into account.

[0035] Preferably, the convolutional neural network model is used to extract deep features of garbage images, including edge features, color distribution, texture patterns and geometric shapes, and to learn features through multiple convolutional, pooling and fully connected layers.

[0036] Preferably, the support vector machine classification method is used to classify waste types based on the high-dimensional features extracted by the convolutional neural network model, mapping waste to different categories based on the hyperplane classification strategy, and optimizing the classification boundary.

[0037] It also provides a robotic arm and conveyor belt collaborative control planning system for intelligent waste sorting, including the following modules:

[0038] The weight detection module is used to detect the weight information of the garbage after it has been disposed of.

[0039] The conveyor belt control module is used to dynamically adjust the conveyor belt speed based on the detected information;

[0040] The visual recognition module, including a 3D camera and recognition system, is used to photograph and reconstruct three-dimensional images of the waste on the conveyor belt, identify the type of waste, and determine the grabbing method and path.

[0041] The robotic arm control module is used to control the robotic arm to complete the garbage grabbing and sorting based on the results of the garbage recognition module;

[0042] The control module is used to receive and process data from each module to complete the coordinated control of the waste sorting process.

[0043] This invention provides a collaborative control planning method for a robotic arm and conveyor belt for intelligent waste sorting. It offers the following advantages:

[0044] 1. This invention adopts a technical solution of coordinated control of robotic arms and conveyor belts, achieving efficient and accurate waste sorting. Compared with existing sorting systems that rely solely on robotic arms or conveyor belts, this invention makes the waste sorting process smoother through the synergistic effect of the two, reducing the problem of low processing efficiency caused by system asynchrony.

[0045] 2. This invention ensures the accuracy of waste sorting and the real-time processing by using a weight detection module and a control strategy that dynamically adjusts the conveyor belt speed in real time. Compared with traditional waste treatment solutions, the system can adjust the processing speed in a timely manner according to the weight of the waste, avoiding delays for heavy items during processing and greatly improving the overall processing efficiency.

[0046] 3. By combining 3D vision recognition with intelligent robotic arm control, this invention can accurately identify and grasp waste of different shapes. Compared with the prior art, this solution has significantly improved the accuracy of recognition and the flexibility of grasping, especially in the processing of complex-shaped waste, avoiding the recognition errors and grasping difficulties existing in traditional technologies.

[0047] 4. This invention uses a control module to coordinate various subsystems in real time and automatically adjust their working status to ensure the intelligent and adaptive adjustment of the waste sorting process. Compared with traditional fixed process control schemes, the control module can optimize sorting strategies based on real-time feedback, improving the system's responsiveness and intelligence level, thereby effectively improving the overall efficiency of waste treatment. Attached Figure Description

[0048] Figure 1 This is a flowchart of the method steps of the present invention;

[0049] Figure 2 This is a system architecture diagram of the present invention. Detailed Implementation

[0050] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0051] Please see the appendix Figure 1 This invention provides a method for collaborative control planning of a robotic arm and conveyor belt for intelligent waste sorting, comprising the following steps:

[0052] S1. When the sensor detects that the initial garbage has been dumped on the conveyor belt, the conveyor belt is controlled to transport the garbage to the garbage bag breaking device at a set speed for bag breaking.

[0053] S2. When the garbage is transported to the bag breaking device, control the conveyor belt to slow down or stop until the garbage bag breaking process is completed.

[0054] S3. After the garbage bags are broken and the garbage is spread out, control the conveyor belt to transport the garbage to the metal detection device for metal detection, and control the speed of the conveyor belt according to the working status of the metal detection device and the current speed of the conveyor belt.

[0055] S4. After completing the metal detection, control the conveyor belt to transport the garbage to the visual recognition area, use a 3D camera to perform three-dimensional reconstruction of the garbage, identify the type of garbage and determine the garbage grabbing method and path;

[0056] S5. Control the robotic arm to grab the garbage according to the garbage grabbing method and path determined by the recognition results, and transfer the garbage to the designated garbage bin.

[0057] In step S1 of this embodiment, during the intelligent waste sorting process, the system needs to effectively detect the initial state of the waste and control the conveyor belt to transport the waste to the bag-breaking device for pre-processing at a set speed. This process not only affects the smooth progress of the waste sorting process but also the accuracy of subsequent identification and grasping. Generally, the initial state of the waste is quite complex, possibly appearing in bagged, loose, or piled-up form. Therefore, the operating parameters of the conveyor belt need to be reasonably adjusted for different types of waste states to ensure the effectiveness of the bag-breaking process. In one possible implementation, the system uses multi-sensor fusion technology to detect the characteristics of the initial waste and, combined with the conveyor belt dynamics model, optimizes the conveying speed so that the waste can be fully processed in the bag-breaking area.

[0058] In this embodiment, the system first uses a sensor array to detect the arrival of waste. Alternatively, a weight sensor (such as a strain gauge sensor) can be used to make a preliminary judgment on the waste's mass. Specifically, when waste is dumped onto the conveyor belt, the instantaneous mass change measured by the sensors can be used to determine the waste's density distribution. Let the initial mass carried by the conveyor belt be... The total change in mass after dumping the garbage is:

[0059] ;

[0060] in, This represents the change in the mass of the conveyor belt. This represents the total mass on the conveyor belt at the current moment. This represents the initial mass of the conveyor belt. Generally, the system allows setting a mass threshold. ,like If the signal is received, it is determined that the garbage has been successfully disposed of, triggering the conveyor belt start signal.

[0061] In one possible implementation, in addition to quality inspection, the system also utilizes optical sensors and an infrared ranging module to identify the geometric dimensions of the waste. Specifically, the system uses LiDAR to scan the three-dimensional shape of the waste to obtain its height. and area The waste filling degree is determined based on the following calculation formula:

[0062] ;

[0063] in, For the density of garbage, For the area of ​​the garbage, For the height of the garbage, This is the effective conveying area of ​​the conveyor belt. If... If the density of the waste is high, the conveying speed needs to be reduced appropriately to ensure the processing effect of the bag-breaking device.

[0064] In some embodiments, the system can also incorporate an RGB-Depth (RGB-D) camera for further morphological analysis. Alternatively, a deep learning model (such as MobileNet) can be used to classify the shape features of the waste, distinguishing between bagged and loose waste. When bagged waste is detected, the system can introduce a speed regulation mechanism into the conveyor belt control algorithm, employing an exponentially decaying speed control strategy, i.e.:

[0065] ;

[0066] in, Current time Transmission speed at that time For the maximum initial transmission speed, The attenuation coefficient is... This is the base of the natural logarithm. This allows the speed to be gradually reduced as the waste approaches the bag-breaking device, improving the bag-breaking efficiency.

[0067] In another possible implementation, to ensure that waste does not roll or accumulate during transport, the conveyor belt employs dynamic acceleration compensation control, using a PID (proportional-integral-derivative) control algorithm to achieve stable transport. Assume the desired speed of the conveyor belt is... The real-time measured transmission speed is The speed error is then defined as:

[0068] ;

[0069] PID control calculations are as follows:

[0070] ;

[0071] in, For speed error, For the desired speed, For actual speed, , and These are the proportional, integral, and differential gain parameters, respectively. The output value of the conveyor belt controller. Indicates the duration of the control process. For the time variable in integration operations, For speed error The derivative with respect to time. PID control enables the conveyor belt to dynamically adjust its operating speed according to the garbage load, avoiding uneven conveying caused by garbage accumulation.

[0072] In some embodiments, the surface material of the conveyor belt may be a high-friction coefficient anti-slip coating to improve the stability of waste conveying. Alternatively, polyurethane (PU) or rubber composite materials may be used, with a static friction coefficient of [missing information - likely a coefficient of friction]. Typically greater than 0.8, this effectively prevents waste from slipping due to inertia. Furthermore, in some scenarios, the conveyor belt can be integrated with a lateral flow guide device to correct the position of larger waste items, allowing them to enter the bag-breaking area more evenly.

[0073] In another possible implementation, to further optimize the operation of the conveyor belt, the system can use intelligent learning algorithms to predict and adaptively adjust for blockages that may occur during waste transport. Specifically, the system can record conveyor belt operation data and use a recurrent neural network (RNN) to analyze historical transport data, predict potential waste accumulation points in advance, and then adjust the conveyor speed or change the waste arrangement in the next round of transport to reduce the probability of blockages.

[0074] In some embodiments, the control unit of the conveyor belt can be linked with a host computer system, enabling the transmitted data to be updated to the central processing unit (CPU) in real time. Alternatively, the Modbus TCP / IP communication protocol can be used to remotely monitor and adjust the conveyor belt speed and waste status data, so as to maintain stable operation under complex working conditions.

[0075] For step S2, in this embodiment, during the waste transportation process, to ensure that the waste is fully processed in the bag-breaking device, the system needs to precisely control the operation of the conveyor belt, appropriately slowing down or pausing it when the waste arrives at the bag-breaking device to ensure the stability of the bag-breaking process. Generally, the efficiency of waste bag-breaking is affected by factors such as waste shape, bag material, and quantity. Therefore, the control strategy for slowing down or pausing the conveyor belt needs to incorporate real-time detection data to adjust the conveyor belt's operation to match the working rhythm of the bag-breaking device. In one possible implementation, the system determines the conveyor belt's movement mode based on waste location detection, shape analysis, and the load status of the bag-breaking device, ensuring that the waste enters the bag-breaking area under optimal conditions.

[0076] In this embodiment, the system first needs to determine whether the waste has reached the bag-breaking device. As an option, a laser rangefinder sensor can be used to monitor the movement trajectory of the waste. Specifically, a set of laser rangefinder sensors is arranged in the area of ​​the conveyor belt near the bag-breaking device to measure the leading edge position of the waste. Location of the bag-breaking device inlet distance :

[0077] ;

[0078] in, Current location of trash Location of the bag-breaking device inlet The distance between them.

[0079] In general, the system sets a critical distance. ,when At that time, the conveyor belt deceleration or pause control logic is triggered to ensure that the garbage can enter the bag breaking device stably.

[0080] In one possible implementation, the conveyor belt deceleration process can employ a staged speed control strategy. Specifically, as the waste approaches the bag-breaking device, the conveyor belt first decelerates at a linearly decreasing speed, allowing the waste to slowly enter the bag-breaking area to prevent it from scattering due to inertial impact. Let the initial speed of the conveyor belt be... The final target speed during the deceleration phase is The speed control function of the conveyor belt can then be expressed as:

[0081] ;

[0082] in, Conveyor belt in time The speed of time, This is the initial speed of the conveyor belt (the speed before deceleration). This is a deceleration parameter, the magnitude of which can be dynamically adjusted according to the density and shape of the waste and the processing capacity of the bag-breaking device. In some embodiments, if the waste bags are large or the packing density is high, the system can increase... This is to accelerate the deceleration process and prevent the garbage from shifting or deviating when it enters the bag-breaking device.

[0083] Alternatively, the system can further pause the conveyor belt once the waste has fully entered the bag-breaking device to ensure the smooth progress of the bag-breaking process.

[0084] Specifically, the system can detect the operating status of the bag-breaking device using pressure sensors or image recognition technology. For example, a contact pressure sensor can be installed at the inlet of the bag-breaking device to detect when waste exerts a stable contact force on the device. When the system determines that the garbage has entered the bag-breaking state, the conveyor belt will stop running until the bag breaking is completed.

[0085] In some embodiments, to improve the stability of the bag-breaking process, the system can also dynamically adjust the pause time of the conveyor belt based on the load status of the bag-breaking device. As one possible implementation, the system monitors the load status of the bag-breaking device through motor power monitoring; if the power consumption of the bag-breaking device... Exceeding the set threshold This indicates that the garbage is still in the process of being broken open, and the system keeps the conveyor belt paused to prevent the garbage from being transported away before it is completely broken open. When detected... And the duration exceeds the set time. When the system determines that the bag breaking is complete, it resumes the operation of the conveyor belt.

[0086] In another possible implementation, the system can further optimize the conveyor belt's pause strategy by incorporating visual recognition. Specifically, industrial cameras can be used to monitor the bag-breaking status of the waste in real time. If the system detects that the waste bag has been completely torn and the waste inside has been dispersed, the control system sends a command to the conveyor belt to resume operation. Alternatively, the system can use deep learning models (such as YOLO or ResNet) to process the image data to accurately determine the completion status of bag breaking, thereby improving the system's automation level and processing efficiency.

[0087] In some embodiments, to prevent waste from accumulating during pauses, the system can also incorporate a vibration mechanism or lateral guidance device to slightly vibrate or guide the waste during conveyor belt pauses, making the waste more evenly distributed in the working area of ​​the bag-breaking device and improving the bag-breaking effect. As one possible implementation, the vibration mechanism can be driven by an eccentric wheel motor, periodically applying low-amplitude vibrations when the conveyor belt pauses, making the waste bags easier for the bag-breaking device to handle.

[0088] In step S3, in this embodiment, after the waste enters the bag-breaking device and completes preliminary processing, the system needs to further classify and identify the waste. At this point, the waste that has already undergone preliminary sorting usually has different shapes, colors, sizes, and other characteristics, requiring the use of image recognition technology and machine vision systems, combined with artificial intelligence algorithms, to classify the waste more accurately. In the aforementioned steps, the system precisely controls the conveying and processing status of the waste through sensor data and control algorithms, ensuring that the waste stably enters the bag-breaking area and is effectively processed. In the following steps, the system further identifies and classifies the waste using a high-resolution camera or multispectral sensor, ensuring that subsequent classification operations can be executed accurately.

[0089] In this embodiment, the system uses a camera array positioned behind the bag-breaking device to capture real-time images of the waste after the bag has been broken. Alternatively, an industrial-grade CCD camera can be used in conjunction with an infrared sensor to capture multi-dimensional images of the waste's appearance. Specifically, the camera captures external information about the waste based on reflected light intensity and color differences, generating RGB and depth images. Image preprocessing techniques are then used to enhance image contrast, making the waste features in the image more prominent.

[0090] In some embodiments, the system can combine a convolutional neural network (CNN) to analyze the acquired images and automatically identify the types of waste. Specifically, the CNN model extracts feature information of waste from the image through multiple convolution and pooling operations, and combines it with a pre-trained classification model to classify the waste. For example, when different types of waste such as metal, plastic, and paper are detected in the image, the system automatically classifies and labels them. The output of the CNN model can then be represented as:

[0091] ;

[0092] in, Indicates the classification result. For the input image, This is the classification function in the CNN model.

[0093] As an alternative, the system can also incorporate deep learning transfer learning methods during the waste sorting process. This transfer learning technique allows the classification model to adaptively adjust to different scenarios. Specifically, the system can use a pre-trained model and fine-tune it based on actual waste characteristics to adapt to waste identification needs in different environments. One advantage of transfer learning is its ability to rapidly improve model accuracy, especially when waste types are diverse and change rapidly.

[0094] In another possible implementation, in addition to visual recognition technology, the system can also incorporate spectral analysis technology for further determination of waste materials. Specifically, by installing near-infrared sensors on the conveyor belt, the material of the waste can be accurately identified. For example, by measuring the absorption characteristics of waste in the near-infrared band, the system can determine whether the waste is plastic, metal, or paper, further refining the classification criteria.

[0095] In some embodiments, the system can also incorporate a two-dimensional laser scanner or a three-dimensional LiDAR (Light Detection and Ranging) to reconstruct the three-dimensional structure of the waste, based on visual recognition technology, to help determine the shape and volume of the waste. Specifically, the system scans the outer contour of the waste and combines it with known geometric features of the object to infer its possible physical properties and provide a basis for subsequent processing.

[0096] In some scenarios, the system can also utilize sensor data fusion technology to comprehensively consider input data from different sensors and further optimize the accuracy and stability of waste sorting through algorithms such as Kalman filters. For example, by fusing data from multiple sensors (such as visual sensors, infrared sensors, and ultrasonic sensors) using a Kalman filter, high-precision waste sorting and identification can be achieved in various environments.

[0097] In another possible implementation, the system can incorporate an intelligent feedback control mechanism to adjust the sorting strategy in real time based on the sorting results during the waste sorting process. For example, when the system identifies a high proportion of a certain type of waste, it can adjust the operating parameters of the sorting equipment (such as wind speed and screening speed) to optimize the sorting effect and ensure sorting efficiency and accuracy.

[0098] For step S4, in this embodiment, after visual recognition and classification, the next step is to actually process the waste according to the classification results, including placing it into appropriate recycling containers or initiating further processing procedures. This step, involving processing operations based on the classification results, is crucial for improving resource recycling efficiency and achieving effective waste sorting. To ensure the accuracy and efficiency of processing, the core task of this step is to allocate different categories of waste to corresponding processing devices or storage areas through a precise control system based on the waste classification results. Generally, sorted waste will be processed differently according to its type; for example, recyclables, non-recyclables, and hazardous materials will be sent to their respective processing units for further processing. To ensure the smooth operation of this process, the system needs to employ high-precision control algorithms and processing mechanisms.

[0099] Based on the classification information obtained in previous steps, the system uses an automated sorting device to allocate different categories of waste. Alternatively, the sorting device can employ an airflow sorting system or a robotic arm, depending on the type and characteristics of the waste. For lightweight recyclables, such as plastic bottles and paper, the system uses airflow to blow them to the corresponding recycling bins. Specifically, the system adjusts the airflow intensity through an airflow pressure control valve; when lightweight waste is detected, the force of the airflow... It can be controlled using the following formula:

[0100] ;

[0101] in, The force of airflow It is an aerodynamic constant. The cross-sectional area of ​​the object. This refers to the pressure difference between the airflow and the waste. The system adjusts... The size ensures that lightweight items are accurately distributed to the recycling bins.

[0102] For heavier waste, such as metal and glass bottles, the system uses a robotic arm sorting system. Specifically, the robotic arm identifies the characteristics of the target waste through a vision recognition system and then initiates a precise gripping function. To achieve this process, the robotic arm incorporates a force feedback control system, which monitors the gripping force on the object in real time. The gripping position should be carefully controlled to avoid damaging the trash or causing it to become uncontrollable. The gripping force can be adjusted using the following formula:

[0103] ;

[0104] in, To monitor the gripping force of an object in real time, The adjustment constant for the gripping force, By adjusting the gripping force in real time to measure the change in object displacement, the system can efficiently and safely classify heavy objects and place them in appropriate storage areas.

[0105] In another possible implementation, the system can also incorporate vibrating screening technology, which is particularly suitable for crushed materials or finer waste. Specifically, the system uses an electric vibrator to cause periodic vibrations on the surface of the waste, causing items to separate naturally according to their weight or size. By precisely controlling the vibration frequency and amplitude, the system can classify waste in a shorter time, thereby improving processing efficiency.

[0106] In some embodiments, the system can also intelligently adjust the sorting method based on the characteristics of the waste, such as dynamically selecting suitable sorting equipment based on the material properties, shape characteristics, and weight distribution of the objects. In this case, the system uses deep learning algorithms to intelligently analyze the identified waste categories and automatically select appropriate sorting equipment and methods. For example, some complex-shaped waste, such as cardboard boxes, may require a combination of robotic arm gripping and airflow to complete the sorting.

[0107] In another embodiment, to achieve more efficient waste disposal, the system can also optimize the sorting process in real time through an information feedback mechanism. Specifically, the system monitors the working status and efficiency of each sorting device in real time through a sensor network and adjusts the sorting strategy based on the real-time data. When the recycling efficiency of a certain type of waste is lower than expected, the system will automatically optimize the equipment parameters or adjust the sorting process, thereby improving the overall efficiency of the entire system.

[0108] For step S5, in this embodiment, it is ensured that each type of waste is accurately processed in its corresponding processing unit to avoid confusion, misallocation, or omission. Generally, subsequent processing may include mechanical processing, thermal treatment, compression processing, or chemical treatment, depending on the type of waste. To ensure the efficiency and sustainability of this process, the system needs to provide precise control mechanisms and real-time feedback to monitor the waste processing status and adjust the processing methods accordingly.

[0109] In this embodiment, the system monitors the waste processing status in real time through a sensor network installed on each processing unit. Alternatively, the system can use various sensors, such as temperature and humidity sensors, pressure sensors, or gas detection sensors, to detect environmental parameters during the processing. Specifically, when waste enters the processing unit, the sensors measure important parameters such as temperature, humidity, and compressibility in real time to determine the waste processing progress and whether the processing method needs adjustment. At this time, the system adjusts the operating status of the processing unit based on the data collected by the sensors to ensure the stability of the processing process.

[0110] For example, when processing wet waste (such as kitchen waste), the system can determine the humidity level based on the humidity measured by a humidity sensor. To adjust the compression rate and drying time of wet waste. If Greater than a certain set value The system will automatically activate the heating device to shorten the waste drying time, thereby reducing the impact of moisture during wet waste treatment on subsequent processing stages. Specifically, the humidity threshold... Set using the following formula:

[0111] ;

[0112] in, The humidity is measured by a humidity sensor. Indicates the mass of moisture in the waste. The total mass of the waste.

[0113] In another possible implementation, if the processing involves harder materials such as metal waste or plastic, the system can use a mechanical compression device to compress the waste, further reducing the volume of waste generated after processing. At this time, a pressure sensor detects the compression force. Exceeding the preset threshold The system can automatically stop the compression process to avoid over-compression of waste or damage to the equipment. Specifically, compression force control can be achieved using the following formula:

[0114] ;

[0115] in, The compressive force detected by the pressure sensor. The stiffness constant of the compression equipment, This represents the change in the object's displacement.

[0116] Alternatively, for waste processed in a thermal treatment process, the system can use a temperature control system to ensure that the processing temperature is maintained within a suitable range. For example, during pyrolysis, the system uses temperature sensors to monitor the temperature of the waste processing unit in real time, ensuring that it fluctuates within a set temperature range. If the temperature... Exceeding the maximum safe temperature The system will automatically reduce the heat source output to prevent overheating from causing uneven processing or equipment failure.

[0117] Furthermore, to enhance the intelligence of the processing, the system can also incorporate machine learning models to adaptively adjust the waste disposal process. For example, based on historical data, the system can predict potential problems that may arise during the processing of different types of waste and optimize the processing flow in advance. Specifically, the machine learning model can be trained based on real-time data from sensors and predict the changing trends of waste disposal in the next few seconds or minutes, thereby adjusting the operating strategies of the processing equipment.

[0118] In some embodiments, the system can optimize the operation of the processing unit through a dynamic optimization control algorithm. For example, through reinforcement learning algorithms, the system can continuously adjust processing parameters based on real-time feedback from waste disposal to achieve long-term optimization of waste disposal.

[0119] The robotic arm and conveyor belt collaborative control planning system for intelligent waste sorting described below can be referred to in correspondence with the robotic arm and conveyor belt collaborative control planning method for intelligent waste sorting described above.

[0120] Please see the appendix Figure 2 The present invention also provides a collaborative control and planning system for intelligent waste sorting robotic arms and conveyor belts, comprising the following modules:

[0121] The weight detection module detects the weight of waste after it is deposited. Using weighing sensors mounted on the conveyor belt, it measures the weight of each item in real time. This weight information is crucial for determining the type of waste and the appropriate subsequent processing method. For example, heavier items may be metal or glass, while lighter items may be plastic or paper. In some embodiments, the system can also combine the relationship between weight and object size to further improve sorting accuracy. The output of this module is transmitted to the conveyor belt control module and the robotic arm control module for real-time adjustment of the sorting strategy.

[0122] The conveyor belt control module dynamically adjusts the conveyor belt speed based on detected information. Generally, if a heavy item is detected, the conveyor belt speed will be appropriately reduced to allow the robotic arm sufficient time to complete the grasping action. For lighter waste, the conveyor belt speed can be increased to improve overall processing efficiency. Alternatively, this module can use a PID control algorithm to precisely control the conveyor belt speed based on real-time data, ensuring stable system operation and accurate waste grasping.

[0123] The visual recognition module, including a 3D camera and recognition system, is used to photograph and reconstruct the 3D structure of the waste on the conveyor belt, identify the type of waste, and determine the grasping method and path. The 3D camera can acquire the three-dimensional shape and spatial position of the waste, providing the robotic arm with more accurate grasping path information. Specifically, the visual recognition module uses image processing algorithms to extract features from each item and combines them with a deep learning model to identify its category (such as plastic, paper, metal, etc.). In addition, the system also determines the posture and path required for the robotic arm to grasp the waste based on its shape and characteristics. This module can also update the spatial position of the waste in real time through a visual positioning system to ensure grasping accuracy.

[0124] The robotic arm control module controls the robotic arm to complete waste grabbing and sorting based on the results of waste identification by the vision recognition module. Specifically, the control module first selects an appropriate grabbing method based on the type and characteristics of the items, such as using grippers or suction cups. For irregularly shaped items, it selects a suitable path. The robotic arm precisely executes the grabbing task through force feedback control and path planning algorithms. In some embodiments, the robotic arm can also adjust the grabbing force based on feedback information to ensure the stability and safety of the grabbing action.

[0125] The control module receives and processes data from various modules, enabling coordinated control of the waste sorting process. It coordinates the vision recognition module, weight detection module, conveyor belt control module, and robotic arm control module to achieve effective information transmission and feedback. Through real-time optimization algorithms, the control module adjusts system operating parameters based on real-time data from each module, ensuring the efficiency and stability of the entire waste sorting process. Specifically, the control module can dynamically adjust processing strategies based on the different characteristics of the waste. For example, during waste identification, if special items (such as batteries or hazardous materials) are detected, the system can promptly change the processing flow to ensure safety.

[0126] The system in this embodiment can be used to execute the above method embodiments, and its principle and technical effect are similar, so they will not be described again here.

[0127] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A collaborative control planning method for robotic arms and conveyor belts used in intelligent waste sorting, characterized in that, Includes the following steps: When the sensor detects that the initial garbage has been dumped onto the conveyor belt, the conveyor belt is controlled to transport the garbage to the garbage bag breaking device at a set speed for bag breaking. When the garbage is transported to the bag breaking device, control the conveyor belt to slow down or stop until the garbage is completely broken; After the garbage bags are broken open and the garbage is spread out, the control conveyor belt will transport the garbage to the metal detection device for metal detection; After metal detection is completed, the control conveyor belt will transport the garbage to the visual recognition area. The 3D camera will be used to reconstruct the garbage in three dimensions, identify the type of garbage, and determine the garbage grabbing method and path. The robotic arm is controlled to grab the garbage according to the garbage grabbing method and path determined by the recognition results, and then sort and transport the garbage to the designated garbage bins. The control conveyor belt transports garbage to the garbage bag-breaking device at a set speed for bag-breaking processing, including: Record conveyor belt operation data; A recurrent neural network model is used to analyze historical transport data and predict potential waste accumulation points. In the next round of transport, the transport speed or waste arrangement can be adjusted in advance. The step of controlling the conveyor belt to slow down or stop when the garbage is transported to the bag-breaking device until the garbage bag breaking process is completed includes: Based on waste location detection, morphology analysis, and the load status of the bag-breaking device, the conveyor belt is controlled to slow down or pause. The method of controlling the conveyor belt to slow down or stop based on waste location detection, morphological analysis, and the load status of the bag-breaking device includes: A laser rangefinder is installed in the area of ​​the conveyor belt near the bag-breaking device to measure the distance between the leading edge of the waste and the inlet of the bag-breaking device. When the distance between the leading edge of the waste and the inlet of the bag-breaking device is lower than a preset critical value, the conveyor belt is controlled to decelerate. The speed control function of the conveyor belt is: ； Where V(t) is the speed of the conveyor belt at time t, V init The initial speed of the conveyor belt is V, and the final target speed of the deceleration section is V. min α is a deceleration parameter, and its value is dynamically adjusted according to the density and shape of the waste and the processing capacity of the bag-breaking device. When the waste has completely entered the bag-breaking device, the conveyor belt is paused, and the pause time is dynamically adjusted according to the load status of the bag-breaking device; when the power consumption P of the bag-breaking device is greater than a set threshold P... th At that time, the conveyor belt remains paused.

2. The method for coordinated control planning of robotic arms and conveyor belts for intelligent waste sorting according to claim 1, characterized in that, The speed control of the conveyor belt includes the following steps: The weight of the waste dumped on the conveyor belt is detected in real time by a weight sensor; Determine if the weight exceeds the set threshold. If it does, reduce the conveyor belt speed to the set slow speed. If it does not exceed the threshold, run at the base conveyor belt speed. During the waste conveying process, the conveyor belt speed is dynamically adjusted by combining the speed sensor data at both ends of the conveyor belt.

3. The collaborative control planning method for robotic arms and conveyor belts for intelligent waste sorting according to claim 1, characterized in that, The deceleration or pausing control of the conveyor belt includes the following steps; The conveyor belt uses sensors in the bag-breaking device to monitor whether the waste has reached the bag-breaking area; Once the waste reaches the bag-breaking area, the conveyor belt automatically switches to deceleration mode until the bag-breaking device detects that the waste has been broken open. After the garbage bags are broken open, the conveyor belt returns to the set speed and continues to transport the garbage.

4. The collaborative control planning method for robotic arms and conveyor belts for intelligent waste sorting according to claim 1, characterized in that, The method of using a 3D camera to reconstruct the three-dimensional structure of the waste includes the following steps: The garbage is captured from multiple perspectives using a 3D camera, and the three-dimensional feature points of the garbage are extracted using a multi-view image reconstruction algorithm. Garbage feature images are classified by combining a deep learning model, wherein the deep learning model is either ResNet50 or MobileNet. In the process of image feature extraction, attention mechanism is used to extract features from salient regions of garbage feature images; Based on the identification results, the grabbing method and path for each type of waste are calculated.

5. The collaborative control planning method for robotic arms and conveyor belts for intelligent waste sorting according to claim 1, characterized in that, The identification of the waste type includes the following steps: The garbage is imaged from multiple perspectives using a 3D camera, and features are extracted from the garbage images based on deep learning algorithms to extract the geometric shape, color, texture and material properties of the garbage. By combining convolutional neural network models and support vector machine classification methods, waste types are classified, and the grabbing points and optimal disposal locations for waste are calculated. Based on the identified waste type, adjust the robotic arm's target grasping strategy, including selecting appropriate adsorption, clamping, or grasping modes; After the garbage sorting is completed, the optimal path of the robotic arm is calculated, and the grasping trajectory is optimized by combining obstacle avoidance algorithm.

6. The collaborative control planning method for robotic arms and conveyor belts for intelligent waste sorting according to claim 1, characterized in that, The robotic arm's grasping and deployment includes the following steps: The robotic arm automatically adjusts its gripping force based on the type of waste identified by the 3D camera; The robotic arm's rotation angle, gripping point, and gripping force are dynamically adjusted to handle waste of different shapes and materials. The robotic arm's grasping trajectory employs a shortest path planning method based on Dijkstra's algorithm, combined with a dynamic obstacle avoidance algorithm to optimize the grasping path.

7. The collaborative control planning method for robotic arms and conveyor belts for intelligent waste sorting according to claim 6, characterized in that, The Dijkstra algorithm establishes a distance graph between the starting position of the robotic arm and the target grasping point, calculates the shortest path from the starting point to the ending point based on the weights between nodes, and dynamically adjusts the path. During the calculation process, the rotation angle of the robotic arm, the grasping posture, and obstacle avoidance requirements are taken into account.

8. The collaborative control planning method for robotic arms and conveyor belts for intelligent waste sorting according to claim 5, characterized in that, The convolutional neural network model is used to extract deep features from garbage images, including edge features, color distribution, texture patterns and geometric shapes, and to learn features through multiple convolutional, pooling and fully connected layers.

9. The method for coordinated control planning of robotic arms and conveyor belts for intelligent waste sorting according to claim 5, characterized in that, The support vector machine classification method is used to classify waste types based on the high-dimensional features extracted by the convolutional neural network model. It maps waste to different categories based on the hyperplane classification strategy and optimizes the classification boundary.

10. A robotic arm and conveyor belt collaborative control planning system for intelligent waste sorting, applied to the robotic arm and conveyor belt collaborative control planning method for intelligent waste sorting as described in any one of claims 1-9, characterized in that, Includes the following modules: The weight detection module is used to detect the weight information of the garbage after it has been disposed of. The conveyor belt control module is used to dynamically adjust the conveyor belt speed based on the detected information; The visual recognition module, including a 3D camera and recognition system, is used to photograph and reconstruct three-dimensional images of the waste on the conveyor belt, identify the type of waste, and determine the grabbing method and path. The robotic arm control module is used to control the robotic arm to complete the garbage grabbing and sorting based on the results of the garbage recognition module; The control module is used to receive and process data from each module to complete the coordinated control of the waste sorting process.

Images