Bucket wheel machine unattended intelligent material taking system and method
By adaptively constructing material point cloud data using multimodal point cloud data, dividing the material handling area and dynamically adjusting parameters, the problems of low accuracy and efficiency in unattended material handling methods of bucket wheel excavators are solved, achieving more efficient and comprehensive intelligent material handling.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUNAN CHANGZHONG MACHINERY
- Filing Date
- 2025-12-30
- Publication Date
- 2026-04-10
AI Technical Summary
In existing technologies, the unattended intelligent material handling method for bucket wheel excavators has low accuracy and efficiency because the material point cloud data obtained by a single method is not accurate under environmental noise.
By acquiring multimodal point cloud data, including sparse point clouds, dense point clouds, and temperature matrices, material point cloud data is generated using the Kalman filter algorithm. The material picking area is then divided using an artificial intelligence model, and area priorities and adjustment parameters are generated to optimize the material picking and stacking process in real time.
It improves the accuracy and efficiency of unattended material handling by bucket wheel excavators, enhances the adaptability and comprehensiveness of material handling schemes, and avoids the problem of low accuracy and efficiency caused by environmental noise.
Smart Images
Figure CN121823255A_ABST
Abstract
Description
Technical Field
[0001] This application belongs to the field of bulk material yard automation technology, specifically a bucket wheel excavator unattended intelligent material handling system and method. Background Technology
[0002] Bucket wheel excavators are highly efficient, continuous bulk material handling equipment widely used in ports, mines, and power plants. They mainly consist of a slewing mechanism, a pitching device, a traveling mechanism, and a bucket wheel assembly. They continuously excavate materials using multiple buckets on the bucket wheel and then transfer the materials to designated locations via a conveying system. They can perform both material handling and stockpiling operations, featuring high efficiency, stable operation, and strong adaptability. They are one of the key pieces of equipment in modern bulk material handling systems.
[0003] Current technologies for unattended intelligent material handling in bucket wheel excavators often rely on lidar to acquire material point cloud data and then handle the material based on the excavator's status. This overlooks the fact that material point cloud data acquired through a single method is often inaccurate under environmental noise. Using this data as a basis for material handling results in low accuracy and efficiency. Therefore, further improvements are needed for unattended intelligent material handling methods for bucket wheel excavators. Summary of the Invention
[0004] This application aims to solve at least one of the technical problems existing in the prior art; to this end, this application proposes an unattended intelligent material handling system and method for bucket wheel excavators, which solves the technical problem that the existing technology ignores the fact that the accuracy of material point cloud data obtained by a single method is often poor under environmental noise, resulting in low accuracy and efficiency of the material handling method.
[0005] To achieve the above objectives, the first aspect of this application provides an unattended intelligent material handling method for a bucket wheel excavator, comprising: Acquire environmental data, material data, and target data; the target data refers to the objectives followed by the bucket wheel excavator during material handling. Material point cloud data is generated based on material data; the material point cloud data refers to the point cloud data of the material to be retrieved. The material picking area is divided based on the material point cloud data, and the area priority is generated. Material handling adjustment parameters are generated based on the material handling area, area priority, and environmental data; these material handling adjustment parameters refer to the parameter adjustment values of relevant equipment during material handling by the bucket wheel excavator. The material handling adjustment parameters are generated based on the material handling adjustment parameters and target data; the material handling adjustment parameters refer to the parameter adjustment values of relevant equipment during the material handling process of the bucket wheel excavator.
[0006] This application, through the above steps, adaptively constructs material point cloud data using multimodal point cloud data of materials, then divides the constructed material point cloud data into regions, and dynamically adjusts the material picking area in real time to carry out material picking operations. At the same time, it dynamically optimizes the operation strategy during the material stacking process, which not only improves the accuracy and efficiency of intelligent material picking, but also enhances the overall adaptability and comprehensiveness of the material picking solution.
[0007] Furthermore, the step of generating material point cloud data based on material data includes: Acquire material data and environmental data; the material data includes sparse point cloud, dense point cloud, and temperature matrix. Modal weights are generated based on environmental data; The material data is spatially aligned using a coordinate system to obtain aligned material data. Material point cloud data is obtained by combining modal weights with aligned material data and then applying a Kalman filter algorithm.
[0008] Furthermore, the generation of modal weights based on environmental data includes: Acquire environmental and material data; the material data includes sparse and dense point clouds; the environmental data includes dust concentration (FN), rainfall intensity (JYQ), and temperature fluctuation (WB). Extract the sparse signal-to-noise ratio (XSNR) corresponding to the sparse point cloud, and extract the dense point cloud density (CDM) and consistency coefficient (CYX) corresponding to the dense point cloud; Through several influencing parameters YC corresponding to the material data i,j The nonlinear relationship between the modal weights corresponding to the material data is used to construct the weight calculation function QJF. i (YC i,j ); where i represents the number corresponding to the material data, which includes sparse point cloud, dense point cloud and temperature matrix; j represents the number of several influencing parameters that affect the modal weights corresponding to the material data; the influencing parameters refer to the parameters that affect the modal weights of the material data; Several modal weights are calculated by substituting the influence parameters corresponding to the material data into the weight calculation function.
[0009] Furthermore, the step of dividing the material picking area and generating area priorities based on the material point cloud data includes: Acquire material point cloud data; Material point cloud data is input into a region generation model to obtain several material picking regions and their corresponding region stability QWX; the region generation model is constructed using an artificial intelligence model; Alarm signals are generated based on regional stability; Obtain the moving distance JL between the bucket wheel excavator and the material collection area; The priority calculation function YJF(QWX, JL) is constructed based on the nonlinear relationship between regional stability, movement distance, and regional priority. Substituting the regional stability and movement distance into the priority calculation function, we can calculate the regional priority corresponding to several material picking areas.
[0010] Furthermore, the region generation model is constructed using an artificial intelligence model, including: Acquire several historical material point cloud data and their corresponding historical material extraction areas and historical area stability; Several historical material point cloud data and their corresponding historical material extraction areas and historical area stability are divided into training data, validation data and test data; the training data, validation data and test data are preprocessed to obtain training set, validation set and test set. Choose an artificial intelligence model as the base model; The base model is trained on each training set, and the learning rate and other hyperparameters are adjusted on the validation set to obtain the pre-trained model. By validating the pre-trained model on the test set, the input material point cloud data is finally obtained, and the output is a region generation model of several material extraction areas and their corresponding regional stability.
[0011] Furthermore, the generation of alarm signals based on regional stability includes: Obtain the regional stability corresponding to several material extraction areas; Determine whether the region's stability is greater than the stability threshold; Yes, do nothing. No, determine whether the region's stability is greater than D times the stability threshold; Yes, generate an alarm signal for unstable material in the material taking area; no, generate an alarm signal for potential material collapse in the material taking area; where D is a proportionality coefficient, D∈(0,1).
[0012] This application monitors the regional stability of several material collection areas. Once the regional stability of a material collection area falls below the stability threshold, a corresponding alarm signal is issued to avoid unnecessary losses caused by the continued operation of the bucket wheel excavator, which could lead to low material collection efficiency.
[0013] Furthermore, the step of generating material handling adjustment parameters based on the material handling area, area priority, and environmental data includes: Acquire material picking area, area priority, environmental data, and adjustment tags; the adjustment tags include material picking parameter tags and material stacking parameter tags; Select the material collection areas corresponding to the highest area priority values in sequence; Extract the material collection point cloud data and material parameters corresponding to the material collection area; Integrate material collection point cloud data, material parameters, and environmental data into material collection analysis data; The material handling parameter labels and material handling analysis data are input into the equipment parameter adjustment model to obtain several material handling adjustment parameters; the equipment parameter adjustment model is obtained by constructing a multi-task branch model; the material handling adjustment parameters include the bucket wheel speed, the bucket wheel pitch angle and the bucket wheel machine forward speed.
[0014] Furthermore, the step of generating stockpile adjustment parameters based on material handling adjustment parameters and target data includes: Acquire target data, bucket wheel excavator stockpile location, environmental data, stockpile parameter labels, and material handling adjustment parameters; Generate the target stacking area based on the bucket wheel excavator's stacking location and target data; Extract the target area location corresponding to the stacking target area; The data on the bucket wheel excavator's stockpiling location, environmental data, target area location, and material reclaiming adjustment parameters are integrated into stockpiling analysis data. The stockpiling parameter labels and stockpiling analysis data are input into the equipment parameter adjustment model to obtain several stockpiling adjustment parameters; the equipment parameter adjustment model is obtained by constructing a multi-task branch model; the stockpiling adjustment parameters include boom elevation angle, conveyor belt speed and conveying direction.
[0015] This application obtains material handling and stacking adjustment parameters through a pre-trained equipment parameter adjustment model. It can adjust equipment parameters in real time based on influencing factors such as material data, target data, and environmental data, enabling the bucket wheel excavator to perform material handling and stacking operations in the optimal state. This improves the accuracy and efficiency of the intelligent material handling method of the bucket wheel excavator when unattended.
[0016] Furthermore, the step of generating the target stacking area based on the bucket wheel excavator's stacking location and target data includes the following steps: Step 1: Obtain the material stacking location and target data of the bucket wheel excavator; the target data includes the stacking area ID, stacking location, and target stacking quantity; Step 2: Calculate the positional distance WJ between the stacking location corresponding to several stacking area IDs and the stacking location of the bucket wheel excavator. k Where k represents the number corresponding to the stacking area ID; for WJ k Sort the regions in ascending order to obtain a list of regions; Step 3: Select the first stacking region ID in the region list; extract the actual stacking quantity of the stacking region ID; determine whether the actual stacking quantity corresponding to the stacking region ID is less than the target stacking quantity; If yes, then the stacking area ID will be used as the stacking target area; No, remove the stacking area ID from the area list and proceed to step three.
[0017] Furthermore, the device parameter adjustment model is obtained by constructing a multi-task branch model, including: Obtain historical material requisition analysis data and historical material requisition adjustment parameters corresponding to the material requisition parameter tags, as well as historical stockpiling analysis data and historical stockpiling adjustment parameters corresponding to the stockpiling parameter tags; The historical material taking analysis data and historical material taking adjustment parameters corresponding to the material taking parameter labels are divided into training data, validation data and test data corresponding to the material taking parameter labels; the training data, validation data and test data are preprocessed to obtain the training set, validation set and test set; The historical stockpiling analysis data and historical stockpiling adjustment parameters corresponding to the stockpiling parameter labels are divided into training data, validation data and test data corresponding to the stockpiling parameter labels; the training data, validation data and test data are preprocessed to obtain the training set, validation set and test set. Choose two machine learning models as the base models for the multi-task branching; Each model is trained on its own training set and its own pre-trained model is obtained by adjusting the learning rate and other hyperparameters on its own validation set. By validating their respective pre-trained models on their respective test sets, the final result is the input material handling parameter labels and their corresponding material handling analysis data or the material stacking parameter labels and their corresponding material stacking analysis data, and the output is the equipment parameter adjustment model of material handling adjustment parameters or material stacking adjustment parameters.
[0018] Another aspect of the present invention provides an unattended intelligent material handling system for bucket wheel excavators, comprising: a data acquisition module, a data analysis module, and an early warning module; the data acquisition module is connected to the data analysis module; the data analysis module and the early warning module are connected to each other. The data acquisition module acquires environmental data, material data, and target data through data acquisition equipment. The data analysis module: generates material point cloud data based on material data; divides material picking areas and generates area priorities based on material point cloud data; generates material picking adjustment parameters based on material picking areas, area priorities, and environmental data; and generates stockpiling adjustment parameters based on material picking adjustment parameters and target data. The early warning module provides prompts based on alarm signals.
[0019] Compared with the prior art, the beneficial effects of this application are: 1. This application generates material point cloud data based on material data; divides material picking areas and generates area priorities based on the material point cloud data; generates material picking adjustment parameters based on the material picking areas, area priorities, and environmental data; and generates stacking adjustment parameters based on the material picking adjustment parameters and target data. By adaptively constructing material point cloud data through multimodal point cloud data of materials, and then dividing the material point cloud data into areas and dynamically adjusting the material picking and stacking process based on the picking areas, the intelligent material picking method becomes more accurate and efficient, and improves the comprehensiveness of the material picking method.
[0020] 2. This application analyzes several influencing parameters corresponding to material data to obtain modal weights corresponding to several material data, so as to better allocate the degree of influence between several material data when constructing comprehensive material point cloud data, making the material point cloud data more accurate and providing accurate data support for subsequent intelligent material handling and risk warning.
[0021] 3. This application first divides the material point cloud data into regions using a pre-trained region generation model to obtain several material picking regions and their corresponding regional stability. Then, considering the impact of the moving distance and the stability of each region on the order of material picking when the bucket wheel excavator is moving to pick up materials, it quantifies the impact of the moving distance and the stability of each region on the order of material picking as region priority, providing a solid data guarantee for subsequent intelligent material picking.
[0022] 4. This application generates a stacking target area by using the stacking position and target data of the bucket wheel excavator. The shortest distance between the stacking position of the bucket wheel excavator and the stacking position corresponding to the target area ID is the main influencing factor. The application also analyzes in real time whether the stacking amount at the stacking position reaches the target stacking amount to select the stacking target area. This allows the bucket wheel excavator to stack materials more quickly and improves the comprehensiveness and efficiency of the intelligent material handling method. Attached Figure Description
[0023] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0024] Figure 1 This is a flowchart of an unattended intelligent material handling method for a bucket wheel excavator according to this application; Figure 2 A flowchart for generating alarm signals in this application; Figure 3 A flowchart is generated for the stacking target area of this application; Figure 4This is a schematic diagram illustrating the principle of an unattended intelligent material handling system for a bucket wheel excavator according to this application. Detailed Implementation
[0025] The technical solutions of this application will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.
[0026] Please see Figure 1 The first aspect of this application provides an unattended intelligent material handling method for a bucket wheel excavator, comprising: Acquire environmental data, material data, and target data; material data refers to the data related to the materials that the bucket wheel excavator needs to pick up; target data refers to the objectives that the bucket wheel excavator follows when picking up materials. Material point cloud data is generated based on material data; material point cloud data refers to the point cloud data of the material to be retrieved. The material picking area is divided based on the material point cloud data, and the area priority is generated. The material picking area is the area obtained by dividing the material to be picked, and the area priority refers to the evaluation of the order of picking materials in the picking area. Based on the material collection area, area priority, and environmental data, material collection adjustment parameters are generated. These parameters refer to the parameter adjustment values of relevant equipment during material collection by the bucket wheel excavator. The material handling adjustment parameters are generated based on the material handling adjustment parameters and target data. The material handling adjustment parameters refer to the parameter adjustment values of the relevant equipment of the bucket wheel excavator when stacking materials.
[0027] In this embodiment, generating material point cloud data based on material data includes: Acquire material data and environmental data; material data includes sparse point cloud, dense point cloud and temperature matrix; sparse point cloud refers to material point cloud data acquired by millimeter wave radar, dense point cloud refers to material point cloud data acquired by lidar, and temperature matrix refers to material data acquired by infrared thermal imaging equipment. Modal weights are generated based on environmental data; modal weights refer to the weight coefficients corresponding to the influence data that constitute the material point cloud data. Aligned material data is obtained by spatially aligning material data using a coordinate system. This step unifies different material data spatially, resulting in more accurate comprehensive point cloud data. Material point cloud data is obtained by combining modal weights with aligned material data and then applying a Kalman filter algorithm.
[0028] In this embodiment, generating modal weights based on environmental data includes: Acquire environmental and material data; material data includes sparse and dense point clouds; environmental data includes dust concentration (FN), rainfall intensity (JYQ), and temperature fluctuation (WB). Extract the sparse signal-to-noise ratio (XSNR) corresponding to the sparse point cloud, and extract the dense point cloud density (CDM) and consistency coefficient (CYX) corresponding to the dense point cloud; the consistency coefficient represents the difference between the measurement results of millimeter-wave radar and lidar in the same area, expressed as the standard deviation. Through several influencing parameters YC corresponding to the material data i,j The nonlinear relationship between the modal weights corresponding to the material data is used to construct the weight calculation function QJF. i (YC i,j ); where i represents the number corresponding to the material data, which includes sparse point cloud, dense point cloud and temperature matrix; j represents the number of several influencing parameters that affect the modal weights corresponding to the material data; the influencing parameters refer to the parameters that affect the modal weights of the material data; The weight calculation function satisfies the following formula: ;in, Let α be the parameter function corresponding to the i-th material data. i Represented as modal priority coefficient, α i >0, α i The specific values are set based on experience. In this embodiment, the modal priority coefficients corresponding to sparse point cloud, dense point cloud and temperature matrix are set to 1.2, 1.5 and 1.0 respectively. In this embodiment, the influencing parameters considered for sparse point clouds are dust concentration FN and sparse signal-to-noise ratio XSNR. The parameter function corresponding to sparse point clouds satisfies the following formula: ; In this embodiment, the influencing parameters considered for dense point clouds are dense point cloud density (CDM) and consistency coefficient (CYX). The parameter function corresponding to sparse point clouds satisfies the following formula: Where MY represents the point cloud density threshold, BDM represents the standard point cloud density, and BYX represents the standard consistency coefficient. Specific values are set based on experience; in this embodiment, BDM is set to 200 and BYX is set to 4. The point cloud density threshold satisfies the following formula: ; In this embodiment, the influencing parameters considered for the temperature matrix are rainfall intensity JYQ and temperature fluctuation WB, where temperature fluctuation refers to the rate of change of ambient temperature. The parameter function corresponding to the temperature matrix satisfies the following formula: ; Several modal weights are calculated by substituting the influence parameters corresponding to the material data into the weight calculation function.
[0029] This embodiment analyzes the influencing parameters related to material data to determine the modal weights corresponding to each material data. This allows for a more reasonable allocation of the influence between different data when constructing comprehensive material point cloud data, improving the accuracy of the point cloud data and providing more reliable data support for subsequent intelligent material handling operations and risk warnings.
[0030] In this embodiment, the process of dividing the material picking area based on the material point cloud data and generating area priorities includes: Acquire material point cloud data; Material point cloud data is input into the region generation model to obtain several material picking regions and their corresponding region stability QWX; region stability refers to the stability of the material when picking up materials in the region. The higher the region stability, the safer it is to pick up materials. QWX∈[0,1]; the region generation model is constructed using an artificial intelligence model. Alarm signals are generated based on regional stability; Obtain the moving distance JL between the bucket wheel excavator and the material collection area; the moving distance refers to the distance that the bucket wheel excavator needs to move when collecting material from the material collection area; The priority calculation function YJF(QWX, JL) is constructed based on the nonlinear relationship between regional stability, movement distance, and regional priority. The priority calculation function satisfies the following formula: Where JL is the moving distance after maximum-minimum normalization, JL∈[0,1]; β and t represent the stability enhancement index and distance decay coefficient, respectively, both of which are greater than 0. The specific values are set according to experience. In this embodiment, β and t are set to 1.5 and 0.5, respectively. Substituting the regional stability and movement distance into the priority calculation function, we can calculate the regional priority corresponding to several material picking areas.
[0031] This embodiment first uses a pre-trained region generation model to divide the material point cloud data, obtaining multiple material collection areas and their corresponding region stability indices. Then, during the material collection process of the bucket wheel excavator, the influence of the moving distance and the stability of each region on the material collection order is quantified, and a comprehensive region priority is generated to provide solid data support and more efficient operational basis for subsequent intelligent material collection decisions.
[0032] The region generation model in this embodiment is constructed using an artificial intelligence model, including: Acquire several historical material point cloud data and their corresponding historical material extraction areas and historical area stability; Several historical material point cloud data and their corresponding historical material extraction areas and historical area stability are divided into training data, validation data, and test data; the training data, validation data, and test data are preprocessed to obtain training set, validation set, and test set; the ratio between training set, test set, and validation set is 7:2:1; Choose an artificial intelligence model as the base model; artificial intelligence models include PointNet++, etc. The base model is trained on each training set, and the learning rate and other hyperparameters are adjusted on the validation set to obtain the pre-trained model. By validating the pre-trained model on the test set, the input material point cloud data is finally obtained, and the output is a region generation model of several material extraction areas and their corresponding regional stability.
[0033] Please see Figure 2 In this embodiment, generating an alarm signal based on regional stability includes: Obtain the regional stability corresponding to several material extraction areas; Determine if the region's stability exceeds a stability threshold; the stability threshold is set based on experience. Yes, do nothing. No, determine whether the region's stability is greater than D times the stability threshold; Yes, generate an alarm signal for unstable material in the material taking area; no, generate an alarm signal for potential collapse of material in the material taking area; where D is a proportionality coefficient, D∈(0,1), and the specific value is set according to experience. In this embodiment, D is set to 0.6, that is, the stability of the area is divided into three levels, and corresponding alarm signals are generated for each level.
[0034] This embodiment monitors the stability of each material taking area in real time. When the stability of a material taking area is detected to be lower than the set threshold, the system will immediately issue a corresponding warning signal to remind the operator to take countermeasures in a timely manner, thereby effectively avoiding the safety risks and efficiency losses caused by the bucket wheel excavator continuing to operate in an unstable area.
[0035] In this embodiment, the generation of material handling adjustment parameters based on the material handling area, area priority, and environmental data includes: Acquire material picking area, area priority, environmental data, and adjust labels; adjustment labels include material picking parameter labels and material stacking parameter labels; adjustment labels refer to the equipment parameter labels that need to be adjusted, where material picking parameter labels are labels related to the material picking position of the bucket wheel excavator, and material stacking parameter labels are labels related to the material stacking position of the bucket wheel excavator; The material collection area corresponding to the highest regional priority value is selected sequentially; that is, the material collection operation is performed on the material collection area corresponding to the highest regional priority value each time. When the material collection of a material collection area is completed, the regional priority is updated and a new material collection area is selected for material collection. Extract the material point cloud data and material parameters corresponding to the material picking area; the material point cloud data refers to the material point cloud data corresponding to the material picking area. The material handling point cloud data, material parameters, and environmental data are integrated into material handling analysis data; material handling analysis data refers to the data required to adjust the material handling-related equipment. The material handling parameter labels and material handling analysis data are input into the equipment parameter adjustment model to obtain several material handling adjustment parameters; the equipment parameter adjustment model is obtained by constructing a multi-task branch model; the material handling adjustment parameters include bucket wheel speed, bucket wheel pitch angle and bucket wheel machine forward speed, etc.
[0036] In this embodiment, generating stockpile adjustment parameters based on material handling adjustment parameters and target data includes: Acquire target data, bucket wheel excavator stockpile location, environmental data, stockpile parameter labels, and material handling adjustment parameters; The stacking target area is generated based on the stacking location and target data of the bucket wheel excavator. The stacking target area refers to the specific area where the material needs to be stacked during the stacking process. Extract the target area location corresponding to the stacking target area; in this embodiment, the target area location refers to the center location of the stacking target area; The data on the bucket wheel excavator's stockpiling location, environmental data, target area location, and material handling adjustment parameters are integrated into stockpiling analysis data; stockpiling analysis data refers to the data required for adjusting stockpiling-related equipment. Several stockpiling adjustment parameters are obtained by inputting stockpiling parameter labels and stockpiling analysis data into the equipment parameter adjustment model; the equipment parameter adjustment model is obtained by constructing a multi-task branch model; the stockpiling adjustment parameters include boom elevation angle, conveyor belt speed and conveying direction, etc.
[0037] Please see Figure 3 In this embodiment, generating the target stacking area based on the stacking location of the bucket wheel excavator and target data includes the following steps: Step 1: Obtain the material stacking location and target data of the bucket wheel excavator; the target data includes the stacking area ID, stacking location, and target stacking quantity; the material stacking location of the bucket wheel excavator refers to the position where the bucket wheel excavator will stack the material. Step 2: Calculate the positional distance WJ between the stacking location corresponding to several stacking area IDs and the stacking location of the bucket wheel excavator. k Where k represents the number corresponding to the stacking area ID; for WJ k Sort the regions in ascending order to obtain a list of regions; Step 3: Select the first stacking area ID in the area list; extract the actual stacking quantity of the stacking area ID; determine whether the actual stacking quantity corresponding to the stacking area ID is less than the target stacking quantity; in this embodiment, the actual stacking quantity can be calculated by constructing the stacking material point cloud data of the material of the stacking area ID, converting the point cloud into a continuous surface model through surface reconstruction technology, and then calculating the actual stacking quantity through numerical integration method. If yes, then the stacking region ID will be used as the stacking target region; No, remove the stacking region ID from the region list and proceed to step three.
[0038] This embodiment uses the shortest distance between each stacking location and its corresponding target area ID as the main decision factor through the above steps, and analyzes in real time whether the current stacking amount has reached the set target stacking amount, thereby dynamically selecting the optimal stacking area, enabling the bucket wheel excavator to complete the material stacking process more quickly and efficiently, and further improving the comprehensiveness and operating efficiency of the intelligent material handling system.
[0039] The device parameter adjustment model in this embodiment is obtained by constructing a multi-task branch model, including: Obtain historical material requisition analysis data and historical material requisition adjustment parameters corresponding to the material requisition parameter tags, as well as historical stockpiling analysis data and historical stockpiling adjustment parameters corresponding to the stockpiling parameter tags; The historical material taking analysis data and historical material taking adjustment parameters corresponding to the material taking parameter labels are divided into training data, validation data and test data corresponding to the material taking parameter labels; the training data, validation data and test data are preprocessed to obtain the training set, validation set and test set; the ratio between the training set, test set and validation set is 7:2:1; Several historical stockpiling analysis data and historical stockpiling adjustment parameters corresponding to the stockpiling parameter labels are divided into training data, validation data, and test data corresponding to the stockpiling parameter labels; the training data, validation data, and test data are preprocessed to obtain the training set, validation set, and test set; the ratio between the training set, test set, and validation set is 7:2:1; Two machine learning models are selected as the base models for the multi-task branch; both machine learning models are neural network models. Each model is trained on its own training set and its own pre-trained model is obtained by adjusting the learning rate and other hyperparameters on its own validation set. By validating their respective pre-trained models on their respective test sets, the final result is the input material handling parameter labels and their corresponding material handling analysis data or the material stacking parameter labels and their corresponding material stacking analysis data, and the output is the equipment parameter adjustment model of material handling adjustment parameters or material stacking adjustment parameters.
[0040] Please see Figure 4 Another embodiment of this application provides an unattended intelligent material handling system for a bucket wheel excavator, including: a data acquisition module, a data analysis module, and an early warning module; the data acquisition module and the data analysis module are electrically and / or communicatively connected; the data analysis module and the early warning module are electrically and / or communicatively connected. Data acquisition module: Acquires environmental, material, and target data through data acquisition devices, including millimeter-wave radar, lidar, infrared thermal imaging equipment, and various sensors. Data analysis module: Generates material point cloud data based on material data; divides material picking areas and generates area priorities based on material point cloud data; generates material picking adjustment parameters based on picking areas, area priorities, and environmental data; generates stockpiling adjustment parameters based on material picking adjustment parameters and target data; Early warning module: Provides prompts based on alarm signals, including alarm signals indicating unstable materials in the material handling area and alarm signals indicating a risk of material collapse in the material handling area.
[0041] Some of the data in the above formula are calculated by removing dimensions and taking their numerical values. The formula is the closest to the real situation obtained by software simulation of a large amount of collected data. The preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.
[0042] The working principle of this application is as follows: Environmental data, material data, and target data are acquired; material point cloud data is generated based on the material data; material point cloud data is used to divide the material collection area and generate area priorities; material collection adjustment parameters are generated based on the material collection area, area priorities, and environmental data; and stacking adjustment parameters are generated based on the material collection adjustment parameters and target data. Material point cloud data is adaptively constructed using multimodal point cloud data. Subsequently, the material point cloud data is divided into regions, and material collection and stacking processes are dynamically adjusted based on the dynamic material collection areas. This makes the intelligent material collection method more accurate and efficient, and improves the comprehensiveness of the material collection method. It avoids the problem in existing technologies where material point cloud data acquired through a single method often suffers from poor accuracy under environmental noise, leading to low accuracy and efficiency in the material collection method.
[0043] The above embodiments are only used to illustrate the technical methods of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of this application without departing from the spirit and scope of the technical methods of this application.
Claims
1. A method for unattended intelligent material handling of a bucket wheel excavator, characterized in that, include: Acquire environmental data, material data, and target data; the target data refers to the objectives followed by the bucket wheel excavator during material handling. Material point cloud data is generated based on material data; the material point cloud data refers to the point cloud data of the material to be retrieved. The material picking area is divided based on the material point cloud data, and the area priority is generated. Material handling adjustment parameters are generated based on the material handling area, area priority, and environmental data; these material handling adjustment parameters refer to the parameter adjustment values of relevant equipment during material handling by the bucket wheel excavator. The material handling adjustment parameters are generated based on the material handling adjustment parameters and target data; the material handling adjustment parameters refer to the parameter adjustment values of relevant equipment during the material handling process of the bucket wheel excavator.
2. The unattended intelligent material handling method for a bucket wheel excavator according to claim 1, characterized in that, The step of generating material point cloud data based on material data includes: Acquire material data and environmental data; the material data includes sparse point cloud, dense point cloud, and temperature matrix. Modal weights are generated based on environmental data; The material data is spatially aligned using a coordinate system to obtain aligned material data. Material point cloud data is obtained by combining modal weights with aligned material data and then applying a Kalman filter algorithm.
3. The unattended intelligent material handling method for a bucket wheel excavator according to claim 2, characterized in that, The generation of modal weights based on environmental data includes: The environmental data includes dust concentration FN, rainfall intensity JYQ, and temperature fluctuation WB; Extract the sparse signal-to-noise ratio (XSNR) corresponding to the sparse point cloud, and extract the dense point cloud density (CDM) and consistency coefficient (CYX) corresponding to the dense point cloud; Through several influencing parameters YC corresponding to the material data i,j The nonlinear relationship between the modal weights corresponding to the material data is used to construct the weight calculation function QJF. i (YC i,j ); where i represents the number corresponding to the material data, which includes sparse point cloud, dense point cloud and temperature matrix; j represents the number of several influencing parameters that affect the modal weights corresponding to the material data; the influencing parameters refer to the parameters that affect the modal weights of the material data; Several modal weights are calculated by substituting the influence parameters corresponding to the material data into the weight calculation function.
4. The unattended intelligent material handling method for a bucket wheel excavator according to claim 1, characterized in that, The step of dividing the material picking area and generating area priorities based on material point cloud data includes: Acquire material point cloud data; Material point cloud data is input into a region generation model to obtain several material picking regions and their corresponding region stability QWX; the region generation model is constructed using an artificial intelligence model; Alarm signals are generated based on regional stability; Obtain the moving distance JL between the bucket wheel excavator and the material collection area; The priority calculation function YJF(QWX, JL) is constructed based on the nonlinear relationship between regional stability, movement distance, and regional priority. Substituting the regional stability and movement distance into the priority calculation function, we can calculate the regional priority corresponding to several material picking areas.
5. The unattended intelligent material handling method for a bucket wheel excavator according to claim 4, characterized in that, The region generation model is constructed using an artificial intelligence model, including: Acquire several historical material point cloud data and their corresponding historical material extraction areas and historical area stability; Several historical material point cloud data and their corresponding historical material extraction areas and historical area stability are divided into training data, validation data and test data; the training data, validation data and test data are preprocessed to obtain training set, validation set and test set. Choose an artificial intelligence model as the base model; The base model is trained on each training set, and the learning rate and other hyperparameters are adjusted on the validation set to obtain the pre-trained model. By validating the pre-trained model on the test set, the input material point cloud data is finally obtained, and the output is a region generation model of several material extraction areas and their corresponding regional stability.
6. The unattended intelligent material handling method for a bucket wheel excavator according to claim 1, characterized in that, The process of generating material extraction adjustment parameters based on the extraction area, area priority, and environmental data includes: Acquire material picking area, area priority, environmental data, and adjustment tags; the adjustment tags include material picking parameter tags and material stacking parameter tags; Select the material collection areas corresponding to the highest area priority values in sequence; Extract the material collection point cloud data and material parameters corresponding to the material collection area; Integrate material collection point cloud data, material parameters, and environmental data into material collection analysis data; The material taking parameter labels and material taking analysis data are input into the equipment parameter adjustment model to obtain several material taking adjustment parameters; the equipment parameter adjustment model is obtained by constructing a multi-task branch model.
7. The unattended intelligent material handling method for a bucket wheel excavator according to claim 1, characterized in that, The process of generating stockpile adjustment parameters based on material handling adjustment parameters and target data includes: Acquire target data, bucket wheel excavator stockpile location, environmental data, stockpile parameter labels, and material handling adjustment parameters; Generate the target stacking area based on the bucket wheel excavator's stacking location and target data; Extract the target area location corresponding to the stacking target area; The data on the bucket wheel excavator's stockpiling location, environmental data, target area location, and material reclaiming adjustment parameters are integrated into stockpiling analysis data. Several stockpile adjustment parameters are obtained by inputting stockpile parameter labels and stockpile analysis data into the equipment parameter adjustment model; the equipment parameter adjustment model is obtained by constructing a multi-task branch model.
8. The unattended intelligent material handling method for a bucket wheel excavator according to claim 7, characterized in that, The process of generating the target stacking area based on the bucket wheel excavator's stacking location and target data includes the following steps: Step 1: Obtain the material stacking location and target data of the bucket wheel excavator; the target data includes the stacking area ID, stacking location, and target stacking quantity; Step 2: Calculate the positional distance WJ between the stacking location corresponding to several stacking area IDs and the stacking location of the bucket wheel excavator. k Where k represents the number corresponding to the stacking area ID; for WJ k Sort the regions in ascending order to obtain a list of regions; Step 3: Select the first stacking region ID in the region list; extract the actual stacking quantity of the stacking region ID; determine whether the actual stacking quantity corresponding to the stacking region ID is less than the target stacking quantity; If yes, then the stacking area ID will be used as the stacking target area; No, remove the stacking area ID from the area list and proceed to step three.
9. A method for unattended intelligent material handling of a bucket wheel excavator according to claim 6 or claim 7, characterized in that, The device parameter adjustment model is obtained by constructing a multi-task branch model, including: Obtain historical material requisition analysis data and historical material requisition adjustment parameters corresponding to the material requisition parameter tags, as well as historical stockpiling analysis data and historical stockpiling adjustment parameters corresponding to the stockpiling parameter tags; The historical material taking analysis data and historical material taking adjustment parameters corresponding to the material taking parameter labels are divided into training data, validation data and test data corresponding to the material taking parameter labels; the training data, validation data and test data are preprocessed to obtain the training set, validation set and test set; The historical stockpiling analysis data and historical stockpiling adjustment parameters corresponding to the stockpiling parameter labels are divided into training data, validation data and test data corresponding to the stockpiling parameter labels; the training data, validation data and test data are preprocessed to obtain the training set, validation set and test set. Choose two machine learning models as the base models for the multi-task branching; Each model is trained on its own training set and its own pre-trained model is obtained by adjusting the learning rate and other hyperparameters on its own validation set. By validating their respective pre-trained models on their respective test sets, the final result is the input material handling parameter labels and their corresponding material handling analysis data or the material stacking parameter labels and their corresponding material stacking analysis data, and the output is the equipment parameter adjustment model of material handling adjustment parameters or material stacking adjustment parameters.
10. An unattended intelligent material handling system for a bucket wheel excavator, applied to the unattended intelligent material handling method for a bucket wheel excavator as described in any one of claims 1-9, characterized in that, include: Data acquisition module, data analysis module, and early warning module; The data acquisition module is connected to the data analysis module; the data analysis module is connected to the early warning module. The data acquisition module acquires environmental data, material data, and target data through data acquisition equipment. The data analysis module: generates material point cloud data based on material data; divides material picking areas and generates area priorities based on material point cloud data; generates material picking adjustment parameters based on material picking areas, area priorities, and environmental data; and generates stockpiling adjustment parameters based on material picking adjustment parameters and target data. The early warning module provides prompts based on alarm signals.