Quartering hammer control method, device and equipment applied to unmanned crushing vehicle

Through deep learning, cluster analysis and fuzzy PID control and other technologies, the accurate identification of unmanned crusher vehicles, path planning and dynamic adjustment of hammers are achieved, and the problems of low efficiency and insufficient accuracy of unmanned crusher vehicles are solved, and the problems of low efficiency and insufficient accuracy of unmanned crusher vehicles are improved. Work efficiency and safety.

CN119065231APending Publication Date: 2024-12-03SHANDONG PROMOTE MECHANICAL & ELECTRICAL TECHNOLOGY CO LTD

Patent Information

Application Number
CN202411444502.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-16
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing driverless crushing vehicles have problems such as low crushing efficiency, insufficient precision, and poor adaptability to complex environments during operation. This is mainly due to the limited intelligence of traditional control systems, which cannot achieve precise control and dynamic adjustment.

Method used

Deep learning algorithms are used to identify and locate broken targets, combine clustering algorithms and intelligent decision algorithms to plan broken paths and strategies, and accurately control and dynamic adjustment of the broken hammer through fuzzy PID control method.

Benefits of technology

It improves the working efficiency and operating accuracy of driverless crushing vehicles, reduces manual operation risks, is suitable for complex and changeable operating environments, and reduces labor costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119065231A_ABST
    Figure CN119065231A_ABST
Patent Text Reader

Abstract

The embodiment of the invention provides a breaking hammer control method, device and equipment applied to an unmanned breaking vehicle, which are applied to the technical field of engineering machinery, and the method comprises the following steps: acquiring real-time working environment data and preprocessing the real-time working environment data; on the basis of a deep learning algorithm, according to the preprocessed real-time working environment data, identifying and positioning a broken target; the recognition and positioning results are clustered through a clustering algorithm, clustering analysis is carried out according to the clustering result, and the recognition and positioning results are adjusted; planning a crushing path and a crushing strategy based on an intelligent decision algorithm according to the preprocessed real-time operation environment data and the adjusted identification and positioning result; and based on a fuzzy PID control method, according to the crushing path and the crushing strategy, the crushing hammer is precisely controlled and dynamically adjusted, so that the crushing hammer carries out crushing operation. In this way, the automation degree and the intelligent level of crushing operation can be improved, and the crushing efficiency and quality are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the technical field of engineering machinery, and further to the technical field of intelligent control of engineering machinery, and in particular to a breaker control method, device and equipment applied to an unmanned breaker truck. Background Art

[0002] As an important type of engineering machinery, crushing trucks play a key role in rock crushing, road construction and other operational scenarios. With the advancement of science and technology and the continuous improvement of engineering needs, unmanned crushing trucks have gradually become a trend in the development of the industry; however, existing unmanned crushing trucks have problems such as low crushing efficiency, insufficient precision, and poor adaptability to complex environments during operation. This is because the traditional control system has a limited degree of intelligence, and the control strategy is single when dealing with changing working conditions. It is impossible to achieve precise control and dynamic adjustment, and it is difficult to meet the needs of efficient, precise, and intelligent operations. Summary of the invention

[0003] In view of this, the present disclosure provides a breaker hammer control method, device and equipment applied to an unmanned breaker truck.

[0004] According to a first aspect of the present disclosure, a breaker hammer control method applied to an unmanned breaker truck is provided, the method comprising: Acquire real-time operating environment data and pre-process the real-time operating environment data; Based on the deep learning algorithm, the crushing target is identified and located according to the pre-processed real-time operating environment data; Using clustering algorithms, clustering the recognition and positioning results, performing cluster analysis based on the clustering results, and adjusting the recognition and positioning results based on the cluster analysis results; Based on the intelligent decision-making algorithm, the crushing path and crushing strategy are planned according to the pre-processed real-time operating environment data and the adjusted identification and positioning results; Based on the fuzzy PID control method, the breaker hammer is precisely controlled and dynamically adjusted according to the crushing path and crushing strategy to enable the breaker hammer to perform crushing operations.

[0005] In some implementations of the first aspect, acquiring real-time operating environment data and preprocessing the real-time operating environment data includes: Through data collection, the real-time operating environment data of the sensor network is obtained and the data types are divided into image data and non-image data; De-noise, enhance, crop, and normalize image data; Perform data cleaning, mode conversion, and normalization on non-image data.

[0006] In some implementations of the first aspect, based on a deep learning algorithm, the crushing target is identified and located according to the preprocessed real-time operating environment data, including: Based on the deep learning algorithm, feature extraction is performed on the pre-processed real-time operating environment data; According to the extracted features, the broken targets are identified through the target detection algorithm, and the broken targets are roughly located; Based on coarse positioning, the precise positioning algorithm is used to accurately locate the crushed target.

[0007] In some implementations of the first aspect, clustering the recognition and positioning results using a clustering algorithm, performing cluster analysis based on the clustering results, and adjusting the recognition and positioning results based on the cluster analysis results include: Using clustering algorithms, the recognition and positioning results are clustered, similar results are clustered into different clusters, and cluster analysis is performed; cluster analysis includes: checking the compactness and separation of clusters, whether there are noise points, and whether there are outliers that are incorrectly clustered; Adjust the recognition and positioning results according to the cluster analysis results.

[0008] In some implementations of the first aspect, based on an intelligent decision-making algorithm, a crushing path and a crushing strategy are planned according to preprocessed real-time operating environment data and adjusted identification and positioning results, including: Based on the intelligent decision-making algorithm, the pre-processed real-time operating environment data and the adjusted recognition and positioning results are integrated into the environmental model; the intelligent decision-making algorithm includes: intelligent path planning algorithm and intelligent strategy planning algorithm; the recognition and positioning results include: crushing target, crushing target characteristics and crushing target positioning; According to the location of the crushing target, the intelligent path planning algorithm is used to plan the best path to each crushing target in the environmental model; According to the characteristics of the crushing target, combined with the optimal path to each crushing target and the current crushing vehicle performance, the intelligent strategy planning algorithm is used to plan a crushing strategy that includes each crushing target in the environmental model; the crushing strategy includes: crushing method, crushing parameters, and crushing sequence.

[0009] In some implementations of the first aspect, based on the fuzzy PID control method, according to the crushing path and crushing strategy, the crushing hammer is precisely controlled and dynamically adjusted so that the crushing hammer performs crushing operations, including: According to the actual needs of the crushing operation, fuzzy control rules are formulated to clarify the PID parameter adjustment corresponding to different deviations and deviation change rates; The status information and operation effect of the breaker hammer are obtained in real time through sensors. Combined with fuzzy control rules, the proportional, integral and differential parameters of the PID controller are optimized to achieve precise control and dynamic adjustment of the breaker hammer, so that the breaker hammer can perform crushing operations according to the crushing path and crushing strategy.

[0010] According to a second aspect of the present disclosure, there is provided a breaker hammer control device applied to an unmanned breaker truck, the device comprising: A first processing module, used for acquiring real-time operating environment data and preprocessing the real-time operating environment data; The second processing module is used to identify and locate the crushing target based on the pre-processed real-time operating environment data based on the deep learning algorithm; The third processing module is used to cluster the recognition and positioning results using a clustering algorithm, perform cluster analysis based on the clustering results, and adjust the recognition and positioning results based on the cluster analysis results; The fourth processing module is used to plan the crushing path and crushing strategy based on the intelligent decision-making algorithm according to the pre-processed real-time operating environment data and the adjusted identification and positioning results; The fifth processing module is used to accurately control and dynamically adjust the breaker hammer based on the fuzzy PID control method according to the crushing path and crushing strategy, so that the breaker hammer can perform crushing operations.

[0011] According to a third aspect of the present disclosure, an electronic device is provided. The electronic device includes: at least one processor; and a memory connected to the at least one processor in communication; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described above.

[0012] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, where the computer instructions are used to cause a computer to execute the method described above.

[0013] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program, and when the computer program is executed by a processor, the computer program implements the method described above.

[0014] In the present disclosure, by real-time acquisition and preprocessing of operating environment data, combined with deep learning, cluster analysis, intelligent decision-making and fuzzy PID control, etc., accurate identification, positioning, path planning and crushing strategy formulation of crushing targets, as well as accurate control and dynamic adjustment of crushing hammers are achieved, which effectively improves the working efficiency and operating accuracy of unmanned crushing vehicles, avoids personnel personally operating crushing vehicles in complex and dangerous environments, reduces the risk of accidents, is suitable for complex and changeable operating environments, and at the same time, reduces labor costs, and has broad application prospects.

[0015] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, among which: Figure 1 A flowchart of a breaker hammer control method applied to an unmanned breaker truck provided by an embodiment of the present disclosure is shown; Figure 2 A block diagram of a breaker hammer control device applied to an unmanned breaker vehicle according to an embodiment of the present disclosure is shown; Figure 3 A block diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0018] In addition, the term "and / or" in this article is only a description of the association relationship between the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0019] In response to the problems mentioned in the background technology, the present disclosure provides a breaker hammer control method, device and equipment applied to an unmanned breaker truck.

[0020] Specifically, the real-time operating environment data is acquired and preprocessed; based on the deep learning algorithm, the crushing target is identified and located according to the preprocessed real-time operating environment data; the recognition and positioning results are clustered using the clustering algorithm, and cluster analysis is performed based on the clustering results, and the recognition and positioning results are adjusted according to the clustering analysis results; based on the intelligent decision-making algorithm, the crushing path and crushing strategy are planned according to the preprocessed real-time operating environment data and the adjusted recognition and positioning results; based on the fuzzy PID control method, the crushing hammer is precisely controlled and dynamically adjusted according to the crushing path and crushing strategy, so that the crushing hammer can perform crushing operations.

[0021] In this way, the working efficiency and operation accuracy of unmanned crushing trucks can be improved, avoiding personnel from personally operating crushing trucks in complex and dangerous environments, reducing the risk of accidents and at the same time reducing labor costs.

[0022] The following is a detailed description of the breaker hammer control method, device and equipment for an unmanned breaker truck provided by the present disclosure in conjunction with the accompanying drawings and specific embodiments.

[0023] Figure 1 FIG. 1 is a flow chart showing a method for controlling a breaker hammer applied to an unmanned breaker truck provided by an embodiment of the present disclosure; Figure 1 As shown, the breaker control method 100 applied to an unmanned breaker vehicle may include: S110, acquiring real-time operating environment data and preprocessing the real-time operating environment data.

[0024] Specifically, the real-time operating environment data of the sensor network is acquired through data collection and the data types are divided into image data and non-image data; the image data is subjected to denoising, enhancement, cropping, and normalization processing; and the non-image data is subjected to data cleaning, mode conversion, and normalization processing. Among them: Sensor networks include but are not limited to: cameras, lidar, radar, GPS, inertial measurement units, etc.

[0025] Preprocessing of image data may include: Denoising, using filters to remove noise from images, improve image quality, and reduce interference in subsequent processing.

[0026] Enhancement: Through image enhancement techniques such as contrast adjustment and sharpening, the visual effect of the image is improved to make the target features more prominent.

[0027] Cropping: according to actual needs, crop out the image area containing key information to reduce the amount of data processed and improve processing efficiency.

[0028] Normalization maps the pixel values ​​of image data to a uniform range (such as 0-1 or -1-1) so that the consistency of the data can be maintained during subsequent algorithm processing.

[0029] Preprocessing of non-image data may include: Data cleaning: Check and correct errors, missing values ​​or outliers in the data to ensure the completeness and accuracy of the data.

[0030] Mode conversion, converting non-image data into a format or type suitable for subsequent algorithm processing, for example, converting radar data from polar coordinates to Cartesian coordinates, or converting GPS data to relative position with respect to the work area.

[0031] Normalization, similar to the normalization processing of image data, maps the various physical quantities of non-image data into a unified numerical range in order to eliminate dimensional differences and improve the convergence speed and stability of the algorithm.

[0032] More specifically, dividing the real-time working environment data into image data and non-image data includes: dividing the collected real-time working environment data into image data and non-image data according to its characteristics and uses; among them, image data mainly comes from visual sensors such as cameras, and contains rich visual information, such as color, texture, shape, etc.; non-image data includes physical quantity data such as distance, speed, position, acceleration, etc. collected by sensors such as lidar, radar, and GPS.

[0033] According to the embodiments of the present disclosure: by subdividing data types and performing targeted preprocessing, data quality is improved, providing accurate and reliable input for subsequent algorithm processing.

[0034] S120, based on deep learning algorithm, identifies and locates the crushing target according to the pre-processed real-time working environment data.

[0035] Specifically, based on the deep learning algorithm, the pre-processed real-time operating environment data is subjected to feature extraction; based on the extracted features, the target detection algorithm is used to identify the crushing target and roughly locate the crushing target; based on the rough positioning, the precise positioning algorithm is used to accurately locate the crushing target. More specifically: Select a suitable deep learning model based on the characteristics of real-time operating environment data and task requirements.

[0036] For image data, convolutional neural networks (CNN) and their efficient variants (such as ResNet, lightweight versions of VGG, MobileNet, etc.) are preferred to extract feature information from image data. The network structure is flexibly adjusted according to the size of the data set and computing resources, and the preprocessed image data is input into the image data processing deep learning model. The model automatically learns and extracts high-level features in the image through multi-layer convolutional layers, pooling layers and other structures. These features include not only low-level features such as color and texture, but also more complex semantic features such as shape, edges, and contours.

[0037] For non-image data (such as sound, vibration, GPS coordinates, temperature, humidity, etc.), since it contains time series data (such as sound, vibration) or spatial sequence data (such as GPS trajectory), it is preferred to use recurrent neural networks (RNN), especially LSTM or GRU, to capture long-term dependencies in the data to extract temporal or spatial sequence features.

[0038] The features obtained from image data are effectively fused with the features obtained from non-image data as the input of the target detection algorithm. The fusion can adopt feature vector concatenation, weighted summation or more complex attention mechanism to enhance the target detection capability.

[0039] Based on feature extraction, the target detection algorithm (such as Faster R-CNN, YOLO, etc.) is used to detect the broken targets in the image. The target detection algorithm can identify the target object in the complex background and give the bounding box of the target to achieve the preliminary positioning of the broken target, that is, coarse positioning.

[0040] The coarse positioning results can be quantitatively evaluated to check whether the bounding box accurately covers the fragmented target and whether there are any omissions or false detections. According to the evaluation results, the algorithm parameters can be adjusted to improve the accuracy of coarse positioning. The evaluation can be carried out using methods such as IoU, precision, and recall.

[0041] On the basis of coarse positioning, in order to further improve the positioning accuracy, precise positioning algorithms are introduced, including but not limited to: key point detection algorithms, algorithms based on shape matching, positioning algorithms based on image segmentation, or fusion positioning methods combining multi-sensor data.

[0042] Based on the bounding box information provided by coarse positioning, the precise positioning algorithm is used to further analyze the image content within the bounding box and extract more detailed feature information (such as the target's outline, corner points, etc.). At the same time, combined with the data from other sensors such as lidar and GPS, comprehensive analysis and processing are performed to improve the accuracy and stability of positioning.

[0043] A specific embodiment is provided below to further illustrate the above content: First, model selection and model setting are performed for image data and non-image data.

[0044] For image data: ResNet-50 is selected as the deep learning model for image data processing because it performs well in image classification and object detection tasks and has a certain computational efficiency.

[0045] Input data: preprocessed image data, size is 224x224 pixels, RGB three channels.

[0046] Network structure: ResNet-50 contains multiple residual blocks, each of which consists of a convolutional layer, a batch normalization layer, a ReLU activation function, and a possible convolutional layer. The network gradually extracts high-level features of the image through multi-layer convolution and pooling operations.

[0047] Feature output: The output of the last convolutional layer or global average pooling layer is used as a feature vector with a dimension of 2048.

[0048] For non-image data: For time series data (such as sound and vibration), choose LSTM network; for spatial sequence data (such as GPS trajectory), use GRU network.

[0049] Network structure: LSTM and GRU networks contain multiple recurrent layers, each of which contains multiple memory cells to capture long-term dependencies in the data.

[0050] Feature output: The hidden state of the last time step or the sequence of hidden states of all time steps as a feature vector.

[0051] The above model is used to extract features from the preprocessed real-time operating environment data.

[0052] Furthermore, based on the extracted features, the target detection algorithm is used to identify the broken target and roughly locate the broken target: Before coarse positioning, feature fusion must be performed first. The feature vector of the 2048-dimensional image data is concatenated with the feature vector of the 128-dimensional non-image data (assuming 128 dimensions) to obtain a 2176-dimensional fused feature vector.

[0053] Use the YOLOv5 algorithm for target detection. YOLOv5 is a target detection algorithm based on single-stage detection, which can predict the category and location information of the target in real time.

[0054] Input data: The fused feature vector is used as the input of the YOLOv5 algorithm.

[0055] Output: The algorithm outputs the category and bounding box coordinates (x min ,y min , x max ,y max ) and confidence, and based on the bounding box coordinates output by YOLOv5, the preliminary positioning of the broken target is achieved.

[0056] Furthermore, the intersection over union (IoU) is used to evaluate the accuracy of the bounding box. The IoU calculation formula can be as follows: ; Among them, Area of ​​Overlap is the intersection area of ​​the predicted bounding box and the true bounding box; Area of ​​Union is the union area of ​​the predicted bounding box and the true bounding box.

[0057] According to the IoU evaluation results, the parameters of the YOLOv5 algorithm (such as learning rate, anchor box size, etc.) are adjusted to improve the accuracy of coarse positioning.

[0058] Based on the rough positioning, the precise positioning algorithm is used to accurately locate the crushed target.

[0059] Taking the shape matching algorithm as an example, you can choose contour matching or feature point matching.

[0060] The image content within the bounding box provided by coarse positioning, as well as data from other sensors such as lidar and GPS are used as input data to extract fine features of the image within the bounding box, such as contours and corner points. The extracted features are matched with the predefined fragmented target template using a shape matching algorithm. Multi-sensor data fusion is performed by combining lidar and GPS data to further improve positioning accuracy and stability.

[0061] Finally, the precise position coordinates (x, y, z) and posture information of the crushed target are output.

[0062] Another specific embodiment is provided below to further illustrate: Assume that the unmanned crushing vehicle is equipped with sensors such as high-definition cameras (resolution 1920x1080), lidar (scanning frequency 10Hz, ranging range 0-200m), and GPS receivers (positioning accuracy <2m).

[0063] Acquire real-time data and perform preprocessing, and perform feature extraction on the preprocessed real-time operating environment data.

[0064] For image data: The model selects the YOLOv5 model for image feature extraction and target detection.

[0065] Data input: Real-time image data is obtained from a high-definition camera. The image size is 1920x1080 pixels and is scaled to 640x640 pixels required by the model input.

[0066] Feature extraction: The YOLOv5 model automatically learns image features through its internal convolutional layer CSPDarknet53 to generate feature maps.

[0067] Each grid point in the feature map is responsible for predicting the target in the area, and each grid point predicts multiple bounding boxes and their corresponding confidence and category probabilities.

[0068] For non-image data: For time series data (such as sound and vibration), choose LSTM network; for spatial sequence data (such as GPS trajectory), use GRU network.

[0069] Network structure: LSTM and GRU networks contain multiple recurrent layers, each of which contains multiple memory cells to capture long-term dependencies in the data.

[0070] Feature output: The hidden state of the last time step or the sequence of hidden states of all time steps as a feature vector.

[0071] Furthermore, based on the extracted features, the target detection algorithm is used to identify the broken target and roughly locate the broken target: Before coarse positioning, feature fusion must be performed first to concatenate the feature vectors of image data with the feature vectors of non-image data to obtain a fused feature vector map.

[0072] The YOLOv5 algorithm is used for target detection. The fused feature vector map is used as input. The YOLOv5 model performs target detection on the feature map and outputs the bounding box (x min ,y min , x max ,y max ), where (x min ,y min ) is the coordinate of the upper left corner of the bounding box, (x max ,y max ) are the coordinates of the lower right corner of the bounding box.

[0073] Assume that the bounding box of a broken object is detected as (100, 50, 300, 250), which indicates the position of the object in the image.

[0074] The accuracy of coarse localization is evaluated by calculating the intersection over union (IoU) between the bounding box and the true target, where IoU = intersection area / union area.

[0075] Assuming the bounding box of the real target is (90, 45, 310, 260), we have: Coordinates of the upper left corner of the intersection: (max(100, 90), max(50, 45)) = (100, 50); Coordinates of the lower right corner of the intersection: (min(300, 310), min(250, 260)) = (300, 250); intersection width = 300-100 = 200; intersection height = 250-50 = 200; Area of ​​intersection = 200x200 = 40000.

[0076] The coordinates of the upper left corner of the union: (min(100, 90), min(50, 45)) = (90, 45); The coordinates of the lower right corner of the union: (max(300, 310), max(250, 260)) = (310, 260); Union width = 310-90 = 220; Union height = 260-45 = 215; Area of ​​union = 220x215 = 47300.

[0077] ≈0.846, indicating that there is a good overlap between the detection box and the ground-truth box.

[0078] Based on the bounding box (100, 50, 300, 250) obtained by coarse positioning, a method based on image segmentation and shape matching is used for precise positioning, and the image within the bounding box is segmented more finely to extract the precise outline of the broken target.

[0079] Assume that through shape matching, the exact center point coordinates of the broken target are obtained as (200, 150), and the width and height of the target are 200 pixels and 150 pixels respectively.

[0080] Combined with the laser radar data, the point cloud data can be used to correct the image positioning results to further improve the positioning accuracy. Furthermore, by assuming that the center point coordinates of the same target detected by the laser radar are (X lidar , Y lidar , Z lidar ), and then merged with the image positioning result on the two-dimensional plane to obtain the final precise positioning coordinates (X final , Y final ).

[0081] It is known that the precise center point coordinates of the broken target obtained through shape matching are (200, 150), and the width and height of the target are 200 pixels and 150 pixels respectively. Assuming that the center point coordinates of the same target detected by the lidar are (190, 145, 50), the image positioning result and the lidar data are fused on the two-dimensional plane, and the center point coordinates of the image positioning (200, 150) and the two-dimensional plane coordinates (190, 145) detected by the lidar are taken for fusion.

[0082] The weighted average method can be used to calculate the final precise positioning coordinates (X final , Y final ). Assume that the weight of the image positioning result is 0.6 and the weight of the lidar positioning result is 0.4, then X final = 200×0.6 + 190×0.4 =120 + 76 = 196; Y final = 150×0.6 + 145×0.4 = 90 + 58 = 148, and the final precise positioning coordinates are (196, 148).

[0083] According to the embodiments of the present disclosure, a deep learning algorithm is used for feature extraction and target detection, combined with coarse positioning and precise positioning, to accurately identify and locate broken targets, and maintain a high recognition rate and positioning accuracy even in complex environments, thereby reducing manual intervention and improving work efficiency.

[0084] S130, clustering the recognition and positioning results using a clustering algorithm, performing cluster analysis based on the clustering results, and adjusting the recognition and positioning results based on the cluster analysis results.

[0085] Specifically, the recognition and positioning results are clustered using a clustering algorithm, similar results are clustered into different clusters, and cluster analysis is performed; wherein the cluster analysis includes: checking the compactness and separation of the clusters, whether there are noise points, and whether there are outliers that are incorrectly clustered; and adjusting the recognition and positioning results according to the cluster analysis results.

[0086] More specifically, a clustering algorithm is used to cluster the recognition and positioning results, similar recognition and positioning results are clustered into different clusters, and cluster analysis is performed; wherein the cluster analysis includes: Compactness evaluation checks the similarity or distance between members within a cluster and evaluates the compactness of the cluster.

[0087] Separability assessment analyzes the similarity or distance between different clusters to ensure that there are clear boundaries between clusters.

[0088] Noise point detection, identifying and processing those data points that do not belong to any cluster or are far away from any cluster.

[0089] Outlier handling, check whether there are outliers that are incorrectly clustered into a cluster and consider reassigning or removing these outliers.

[0090] Clustering effectiveness evaluation: internal indicators (such as silhouette coefficient, Davies-Bouldin index) or external indicators (such as Rand index, adjusted Rand index) are used to evaluate the clustering effect.

[0091] According to the results of clustering analysis, necessary adjustments are made to the original recognition and positioning results. For example, for data points that are incorrectly clustered, it may be necessary to recheck their original data or adjust the clustering parameters. For noise points and outliers, you can consider filtering them out or processing them separately to improve the accuracy and reliability of overall recognition and positioning.

[0092] A specific embodiment is provided below to better illustrate the above content: Taking the K-means algorithm to cluster multiple recognition and positioning results as an example, assuming that there are three types of fragmented targets expected, the number of clusters K is set to 3, and clustering is performed based on the coordinates of the bounding box (which can also be converted into center point coordinates and size features).

[0093] Perform cluster analysis on the clustering results, calculate the intra-cluster variance, the smaller the variance, the more compact the cluster, calculate the inter-cluster distance, set the threshold T, and regard the data points whose closest distance to the center of all clusters is greater than the threshold T as noise points; check whether there are points in each cluster that do not conform to the overall characteristics of the cluster, such as points with abnormally large or small bounding box sizes; use the silhouette coefficient S, S = (b - a) / max(a, b), where a is the average distance between a point and other points in its cluster, and b is the minimum average distance from a point to other clusters. The closer the S value is to 1, the better the clustering effect.

[0094] For example: A noise point is found with a bounding box of (50, 30, 60, 40), which is far away from all cluster centers and is decided to be removed.

[0095] There is an outlier in cluster 2, and the bounding box is abnormally large (100, 100, 500, 400). After checking the original data, it is found to be a false positive and is reassigned to another cluster or removed.

[0096] According to an embodiment of the present disclosure, a clustering algorithm is used to perform cluster analysis on the recognition and positioning results, and the recognition and positioning results are adjusted accordingly. This method effectively removes noise points and outliers, and improves the accuracy and reliability of recognition and positioning.

[0097] S140, based on intelligent decision-making algorithms, plans crushing paths and crushing strategies according to pre-processed real-time operating environment data and adjusted identification and positioning results; Specifically, based on the intelligent decision-making algorithm, the pre-processed real-time operating environment data and the adjusted recognition and positioning results are integrated into the environmental model; (wherein, the intelligent decision-making algorithm includes: an intelligent path planning algorithm and an intelligent strategy planning algorithm; the recognition and positioning results include: crushing targets, crushing target characteristics and crushing target positioning;) according to the crushing target positioning, the intelligent path planning algorithm is used to plan the best path to each crushing target in the environmental model; according to the crushing target characteristics, combined with the best path to each crushing target and the current crushing vehicle performance, the intelligent strategy planning algorithm is used to plan the crushing strategy including each crushing target in the environmental model; (the crushing strategy includes: crushing method, crushing parameters, and crushing sequence).

[0098] More specifically, the preprocessed real-time operating environment data (such as terrain data, obstacle locations, traversable areas, etc.) and the adjusted recognition and positioning results (including crushing target characteristics and crushing target positioning) can be integrated into a unified environmental model.

[0099] The environment model can be constructed using grid-based, polygon-based or point cloud methods.

[0100] In the adjusted identification and positioning results, each crushing target has its specific characteristics, such as material, hardness, size, shape, etc. These characteristics are crucial to the formulation of crushing strategies.

[0101] Detailed characteristic information of the crushing target is extracted from the recognition results, and the crushing targets are classified according to the characteristics so that different crushing strategies can be adopted later; the positioning information of each crushing target is integrated into the environmental model to ensure that the position of the target can be accurately considered during path planning; and the optimal path from the current position of the crushing vehicle to each crushing target is planned in the environmental model using an intelligent path planning algorithm (such as the A algorithm, RRT algorithm, Dijkstra algorithm combined with a heuristic function, or a machine learning algorithm).

[0102] Evaluate the best path: consider factors such as path length, safety (avoiding obstacles and impassable areas), and smoothness.

[0103] For multiple crushing targets, it is necessary to plan a path that traverses all targets, or plan multiple paths according to priority; according to the characteristics of the crushing targets and the performance of the crushing vehicle (such as crushing force, speed, energy consumption, etc.), use intelligent strategy planning algorithms (such as genetic algorithms, ant colony algorithms, and particle swarm optimization algorithms, etc.) to plan a crushing strategy that includes each crushing target in the environmental model (the crushing strategy includes: crushing method, crushing parameters, and crushing order), where: Crushing methods include but are not limited to: impact crushing, shear crushing, extrusion crushing, etc. The crushing method is determined according to the target characteristics and the capacity of the crusher.

[0104] Crushing parameters include but are not limited to: crushing force, speed, crushing depth, frequency, etc.

[0105] Furthermore, the crushing sequence can be optimized according to the priority of the crushing target, the path planning results, and the capacity of the crushing vehicle to reduce time, energy consumption, and improve efficiency.

[0106] It can be understood that during the execution process, the path and strategy can also be dynamically adjusted according to real-time feedback (such as crushing effect, energy consumption, environmental changes, etc.) to ensure the efficiency and safety of the crushing operation.

[0107] A specific embodiment is provided below to better illustrate the above content.

[0108] First, build the environment model, assuming that the input data of the model is as follows: Real-time terrain data: elevation map, with a resolution of 1m x 1m, a total of 100x100 grids.

[0109] Obstacle location: Marked as inaccessible area, with a total of 20 obstacle points, each covering a circular area with a radius of 3 meters.

[0110] Passable area: The remaining area is marked as passable.

[0111] Crushing target characteristics and positioning: There are 5 crushing targets in total, each of which includes material (such as rock, concrete), hardness (level 1-10), size (length x width x height, in meters), and shape (rectangular block).

[0112] The model uses a grid-based approach to construct a 100x100 two-dimensional grid map, in which obstacles and broken targets occupy the corresponding grids and their characteristics are annotated.

[0113] After the model is built, the A* algorithm is selected in combination with a heuristic function (such as Manhattan distance) for path planning. The current position of the crushing vehicle is set as the starting grid coordinate. The A* algorithm is used to search for the shortest path in the environmental model for each crushing target and the shortest path is evaluated. The path length, i.e. the total number of grids, the time to avoid obstacles, i.e. the additional cost of passing through the obstacle grid, and the smoothness, i.e. the number of turns, can be considered during the evaluation.

[0114] Then, the ant colony algorithm is used to plan the crushing strategy. On the grid map of the environmental model, the pheromone concentration is initialized for each path connecting the crushing target. Initially, the pheromone concentration of all paths can be set to the same small value, such as 1; multiple ants (assuming 50) are released, and each ant starts from the current position of the crusher. The ant selects the next crushing target to visit based on the pheromone concentration on the path and heuristic information (such as the distance from the target to the current position, the hardness of the target, etc.); when the ant selects a crushing target, the crushing method and parameters are determined according to the target characteristics. For example, for rock targets with higher hardness, the impact crushing method is selected, and a specific crushing force and speed are randomly selected according to the preset impact crushing parameter range; then continue to select the next target until all targets are traversed. The ant will leave pheromones on the path it has walked. The amount of pheromones is related to the quality of the path. The path quality can be measured according to the inverse of the crushing efficiency evaluation formula. For example, if the crushing strategy corresponding to a path has a high crushing efficiency in the simulation, then the pheromone left on the path is relatively large.

[0115] After all ants complete a traversal, they update the pheromones on the path. The pheromone update formula can be: ; in, is the pheromone concentration on the path (i, j) at the current moment, is the pheromone volatility coefficient, is the increment of pheromone on path (i, j) in this iteration.

[0116] ; Where m is the number of ants, Q is a constant, and E k is the fragmentation efficiency evaluation value of the k-th ant's traversal this time.

[0117] By repeatedly using ant colonies to build paths and update pheromones, after multiple iterations, the pheromone concentration on the path will gradually stabilize, and ants will tend to choose the optimal fragmentation path and strategy, that is, the path with the highest pheromone concentration and the corresponding fragmentation method and parameter combination.

[0118] For example: Crushing target 1: rock, hardness 8, size 2x3x4 meters, rectangular, crushed by impact.

[0119] Crushing target 2: concrete, hardness 6, size 3x2x3 meters, rectangular, crushed by shear.

[0120] … Crushed target n.

[0121] Further, the crushing parameters: Impact crushing: crushing force 50 tons, speed 2 m / s, crushing depth 0.5 m, frequency 5 times / min.

[0122] Shear crushing: crushing force 30 tons, speed 1.5 m / s, crushing depth 0.4 m, frequency 8 times / min.

[0123] Further, the breaking order: The optimal crushing order is determined to be: target 1 > target 2 > ... > target n.

[0124] During the crushing process, sensors can be used to monitor the crushing effect (such as crushing depth, crushing rate), energy consumption and environmental changes (such as new obstacles) in real time. If the crushing effect is not good, that is, the expected crushing depth is not reached, the crushing force is increased or the crushing method is adjusted. If the energy consumption exceeds expectations, the crushing parameters are adjusted or the crushing sequence is optimized to reduce energy consumption. If a new obstacle is encountered, the path is replanned to avoid the obstacle.

[0125] Among them, the crushing efficiency evaluation can be carried out by the following formula: ; Among them, Total Volume_Broken is the total volume of crushing, Time Spent is the total time, Energy Consumed is the total energy consumption.

[0126] According to the embodiments of the present disclosure, the preprocessed environmental data and the adjusted identification and positioning results are integrated into the environmental model, and the optimal path and crushing strategy are planned using an intelligent algorithm. This method can formulate accurate crushing paths and crushing strategies based on the crushing target and the performance of the crushing vehicle, thereby improving the intelligence level and operating efficiency of the crushing operation.

[0127] S150, based on the fuzzy PID control method, accurately controls and dynamically adjusts the breaker according to the crushing path and crushing strategy to enable the breaker to perform crushing operations.

[0128] Specifically, according to the actual needs of the crushing operation, fuzzy control rules are formulated to clarify the PID parameter adjustment corresponding to different deviations and deviation change rates; the status information and operation effect of the crushing hammer are obtained in real time through sensors, and the proportional, integral and differential parameters of the PID controller are optimized in combination with fuzzy control rules to achieve precise control and dynamic adjustment of the crushing hammer, so that the crushing hammer can perform crushing operations according to the crushing path and crushing strategy. More specifically: 1. Set fuzzy control rules 1.1 Determine the input variables Error (E): Indicates the difference between the current position or state of the breaker and the expected position or state, such as breaking force, breaking speed, position deviation, etc.

[0129] Error Change Rate (EC): It indicates the rate of change of the error over time, reflecting the trend of the breaker state change.

[0130] 1.2 Fuzzification The exact values ​​of E and EC are mapped to fuzzy sets, such as {negative big (NB), negative middle (NM), negative small (NS), zero (ZE), positive small (PS), positive middle (PM), positive big (PB)}.

[0131] 1.3 Formulating fuzzy control rules According to actual needs, design the rules of crushing operation, for example, "If E is PB and EC is PS, then the PID parameter adjustment amount ΔK p Increase, ΔK i Reduction, ΔK d Remain unchanged".

[0132] These rules are designed to dynamically adjust the ratio (K) of the PID controller based on the current deviation and the rate of change of the deviation. p )、Integral(K i ) and differential (K d ) parameters for more precise control.

[0133] 2. Real-time status monitoring and feedback 2.1 Sensor Data Collection Use various sensors (such as force sensors, acceleration sensors, position sensors, etc.) to collect real-time information on the working status of the breaker, including breaking force, vibration frequency, position coordinates, etc.

[0134] At the same time, the crushing path and crushing effect are monitored through technologies such as image recognition or laser ranging, and the progress and effect of the operation are evaluated.

[0135] 2.2 Data processing and feedback The collected data is pre-processed by filtering, denoising, etc. to improve data accuracy.

[0136] The processed data is input into the fuzzy controller, and the adjustment amount of PID parameters is calculated according to the fuzzy control rules.

[0137] 3. PID parameter optimization and adjustment 3.1 PID controller parameter adjustment According to the output of the fuzzy controller (i.e. the adjustment amount of the PID parameter ΔK p , ΔK i , ΔKd ), adjust the parameters of the PID controller in real time.

[0138] For example, if the system is responding too slowly, you may need to increase K p or K d ; If the system overshoots or oscillates, it may be necessary to reduce K p Or increase K i .

[0139] 3.2 Dynamic Adjustment and Feedback Loop A closed-loop control system is formed, which continuously monitors the status and operation effect of the breaker and dynamically adjusts the PID parameters until the breaker completes the crushing operation according to the predetermined crushing path and strategy.

[0140] During the entire process, the system can automatically adapt to changes in the external environment and changes in internal system characteristics to maintain efficient and stable crushing operations.

[0141] 4. Crushing strategy and path planning Based on the specific requirements of the crushing operation (such as crushing efficiency, energy consumption, crushing quality, etc.), formulate detailed crushing strategies and path planning.

[0142] The fuzzy PID control method is used to ensure that the breaker can respond flexibly in complex and changing working environments, achieving precise control and efficient operation.

[0143] According to the embodiments of the present disclosure, the breaker hammer precision control and dynamic adjustment system based on the fuzzy PID control method can effectively improve the efficiency and quality of the crushing operation and reduce energy consumption and cost.

[0144] A specific embodiment is provided below to illustrate the above content in more detail.

[0145] 1. Fuzzy control rule design Input variable definition: Deviation (E): Taking the crushing force of a breaker as an example, assuming that the expected crushing force is 100 tons, the actual crushing force is measured in real time.

[0146] Deviation change rate (EC): It indicates the rate of change of the crushing force deviation in continuous time.

[0147] Fuzzy Sets: The fuzzy sets of E and EC are defined as {NB, NM, NS, ZE, PS, PM, PB}, corresponding to negative large, negative medium, negative small, zero, positive small, positive medium, and positive large, respectively.

[0148] Fuzzy control rules: If E=PB and EC=PS, then ΔK p=+0.1,ΔK i =-0.05, ΔK d =0 (indicates that the proportional coefficient increases by 0.1, the integral coefficient decreases by 0.05, and the differential coefficient remains unchanged).

[0149] 2. Real-time status monitoring and feedback Sensor data collection: Use a force sensor to measure the crushing force of the breaker in real time. Assume that the current measured value is 90 tons.

[0150] The vibration and position information of the breaker are monitored using acceleration sensors and position sensors, but the main focus here is on the breaking force.

[0151] Data processing: Calculated deviation E = 100 tons (expected) - 90 tons (actual) = 10 tons.

[0152] The deviation change rate EC = +2 tons / second is calculated through time difference, indicating that the crushing force is increasing.

[0153] 3. PID parameter optimization and adjustment Assume that the initial parameters of the PID controller are K p =2, K i =0.5, K d =0.1.

[0154] According to the above fuzzy control rules E=PB, EC=PS, the adjusted PID parameters are: K p_new =K p +ΔKp=2+0.1=2.1; K i_new =K i +ΔK i =0.5-0.05=0.45; K d_new =K d +ΔK d =0.1+0=0.1.

[0155] Assuming the control signal is U and the error signal is E(t), the output of the PID controller can be expressed as: ; 4. Crushing strategy and path planning Assuming that the crushing path has been planned as a series of points, the breaker needs to reach these points in sequence and perform the crushing operation.

[0156] At each point, the fuzzy PID control method is used to adjust the crushing force, speed and other parameters of the crushing hammer according to the crushing target characteristics (such as material and hardness) of the point and the current state of the crushing vehicle.

[0157] 5. Real-time feedback and dynamic adjustment During the crushing process, the status and operating results of the breaker are continuously monitored by sensors.

[0158] If it is found that the crushing force deviates greatly from the expected value or the crushing effect is not good, the PID parameter adjustment amount is recalculated according to the new deviation and the deviation change rate, and the parameters of the PID controller are updated.

[0159] This closed-loop control method ensures that the breaker can perform efficient and stable crushing operations according to the predetermined crushing path and strategy.

[0160] A more detailed specific embodiment is provided below to further illustrate the above content.

[0161] 1. Input variable definition Deviation (E): Taking the crushing force of a breaker as an example, assuming that the expected crushing force Pexpected = 80 tons, the actual crushing force is measured in real time by a high-precision force sensor with a sampling frequency of 10 times per second. For example, at a certain moment t1, the actual crushing force measured by the force sensor is Pactual (t1) = 75 tons.

[0162] Deviation change rate (EC): The deviation change rate is obtained by calculating the difference between two adjacent deviations and dividing it by the sampling time interval. Assuming that the deviation at the previous moment t0 is E(t0), the deviation at the current moment t1 is E(t1), and the sampling time interval is 0.1 seconds, the deviation change rate is: .

[0163] For example, E(t0)=6 tons (P 期望 =80 tons, the actual crushing force at the previous moment was 74 tons), E(t1)=5 tons (the current actual crushing force P 实际 (t1)=75 tons), then EC(t1)=-10 tons / second, indicating that the crushing force deviation is decreasing.

[0164] 2. Fuzzy Sets The fuzzy sets of E and EC are defined as {NB, NM, NS, ZE, PS, PM, PB}, corresponding to negative large, negative medium, negative small, zero, positive small, positive medium, and positive large, respectively. In order to more accurately describe the fuzziness of the variable, the membership function is defined for each fuzzy set. For example, for the deviation E, the triangular membership function is used. When E≤-15 tons, the membership of NB is 1. As E increases, the membership of NB gradually decreases to 0. When -15 < E≤-5 tons, the membership of NM increases from 0 to 1 and then decreases to 0, and so on. For the deviation change rate EC, a similar triangular membership function is also used, which is defined according to its actual range of change. Assuming that the range of EC is -20 tons / second to 20 tons / second, when EC≤-15 tons / second, the membership of NB is 1. As EC increases, the membership of NB gradually decreases to 0, etc.

[0165] 3. Fuzzy control rules Design the following rules: If E = PB and EC = PS, then ∆K p =+0.15,∆K i =-0.08,∆K d =0 (indicates that the proportional coefficient increases by 0.15, the integral coefficient decreases by 0.08, and the differential coefficient remains unchanged).

[0166] If E = NM and EC = NB, then ∆K p =-0.1,∆K i = +0.06, ∆K d =+0.05 (indicates that the proportional coefficient decreases by 0.1, the integral coefficient increases by 0.06, and the differential coefficient increases by 0.05).

[0167] If E = ZE and EC = ZE, then ∆K p = 0, ∆K i =0,∆K d =0 (indicates that the PID parameters remain unchanged).

[0168] 4. Sensor Data Collection The force sensor is used to measure the crushing force of the breaker in real time, with a measurement accuracy of ±0.5 tons. In addition to the force sensor, an acceleration sensor (with a measurement range of ±5g and an accuracy of ±0.05g) and a position sensor (with a positioning accuracy of ±0.01 meters) are also used to monitor the vibration and position information of the breaker in real time. For example, during the crushing operation, the acceleration sensor detects that the vibration acceleration of the breaker is 2g (g is the acceleration of gravity), and the position sensor monitors that the current position coordinates of the breaker are (x, y, z) = (10.5, 20.3, 5.8) meters.

[0169] At the same time, image recognition technology (using a high-definition camera with a resolution of 1920×1080, taking 30 frames of images per second) and laser ranging technology (measuring range of 0.1 meters to 100 meters, accuracy of ±0.05 meters) are used to monitor the crushing path and crushing effect. Through the image analysis algorithm, the size and shape of the crushing area, as well as whether there are large pieces of unbroken materials and other information can be identified; the laser rangefinder is used to measure the distance between the breaker and the surrounding obstacles to ensure the safety of the crushing operation.

[0170] 5. Data processing The collected force sensor data is filtered and the mean filter algorithm is used to average the data of 5 consecutive sampling points to reduce noise interference.

[0171] Calculate the deviation E = P 期望 -P 实际 For example, the expected crushing force is 80 tons, and the actual crushing force after filtering is 75.1 tons, then the deviation E = 80 - 75.1 = 4.9 tons.

[0172] The deviation change rate EC is calculated by time difference. Assuming that the deviation calculated last time is E(t0)=5.2 tons (the last sampling time is t0), and the deviation calculated currently is E(t1)=4.9 tons (the current sampling time is t1, and the sampling time interval is ∆t = 0.1 seconds), the deviation change rate EC=(4.9 - 5.2) / (0.1)=-3 tons / second, indicating that the crushing force deviation is decreasing, and the speed of decrease is 3 tons per second.

[0173] 6. PID controller parameter adjustment Assume that the initial parameters of the PID controller are K p =1.8, K i =0.4, K d =0.08. According to the above fuzzy control rules, when E = PM (assuming that the value of E is 8 tons at this time, it belongs to PM according to the definition of fuzzy set) and EC = NS (assuming that the value of EC is -4 tons / second at this time, it belongs to NS according to the definition of fuzzy set), by querying the fuzzy control rule table, ∆K is obtained. p =+0.08,∆K i =-0.03,∆K d =+0.04.

[0174] The adjusted PID parameters are: K p_new =K p +∆K p =1.8+0.08=1.88; K i_new =Ki +∆K i =0.4-0.03=0.37; K d_new =K d +∆K d =0.08+0.04=0.12.

[0175] 7. Dynamic Adjustment and Feedback Loops The PID controller calculates the control signal U according to the adjusted parameters. The relationship between the control signal and the error signal E(t) can be expressed by the following control signal formula: ; The integral term is calculated using the trapezoidal integration method, and the differential term is calculated by difference approximation. For example, at the current time t1, the error signal is E(t1) = 4.9 tons, the error signal at the previous time t0 is E(t0) = 5.2 tons, and the sampling time interval is ∆t = 0.1 seconds, then the differential term is approximately: ; Assume that in the previous time period, the integral value of the error signal is: ; Then the integral value at the current time t1 is: ; K p_new =1.88, K i_new =0.37, K d_new =0.12, E(t1)=4.9 tons, , , substituting into the above control signal formula, we can get: U(t1)=1.88×4.9+0.37×10.051+0.12×(-3)=11.5708.

[0176] The breaker adjusts the parameters such as breaking force and speed according to the control signal. For example, the control signal adjusts the hydraulic pressure of the breaker through the hydraulic control system, thereby changing the striking force and frequency of the breaker. Assuming that the relationship between the control signal and the hydraulic pressure is linear, and the proportional coefficient is 2 (unit: tons / MPa), the hydraulic pressure should be adjusted to 11.5708 / 2 = 5.7854MPa.

[0177] The status and operation effect of the breaker are continuously monitored, and the above process of data collection, processing, PID parameter adjustment and control signal output is repeated every 0.1 seconds. The PID parameters are dynamically adjusted until the breaker completes the crushing operation according to the predetermined crushing path and strategy. During the whole process, the system can automatically adapt to changes in the external environment (such as sudden changes in rock hardness, the appearance of obstacles, etc.) and changes in internal system characteristics (such as performance changes caused by equipment wear) to maintain efficient and stable crushing operations.

[0178] 8. Broken path planning Assume that the crushing area is a rectangular area with a length of 20 meters and a width of 15 meters. The targets to be crushed are evenly distributed in the area. According to the on-site terrain and obstacle distribution, the A* algorithm is used to pre-plan the moving path of the crusher. The crushing area is divided into 100×100 grids, each with a side length of 0.2 meters. The A* algorithm is used to search for the shortest path from the starting position to each crushing target point, and these paths are connected to form a complete crushing path. For example, the path from the starting position (x0, y0) = (0, 0) to the first crushing target point (x1, y1) = (5, 3) passes through a series of grid points, which are: (0, 0)-(0, 1)-(0, 2)-(1, 2)-(2, 2)-(3, 2)-(4, 2)-(5, 2)-(5, 3).

[0179] 9. Fragmentation strategy development Combined with the specific requirements of the crushing operation (such as crushing efficiency, energy consumption, crushing quality, etc.), formulate corresponding crushing strategies for different crushing target characteristics; for rock areas with higher hardness, adopt greater crushing force and lower hitting frequency to ensure the crushing effect; for areas with lower hardness, the crushing force can be appropriately reduced and the hitting frequency can be increased to improve crushing efficiency and reduce energy consumption. Assuming that in a certain area, the hardness of the rock is level 8 (the hardness range is 1 to 10, and level 10 is the highest hardness), according to the pre-set strategy, the crushing force should be adjusted to 70 tons and the hitting frequency to 4 times per minute.

[0180] At each crushing point, the crushing force, speed and other parameters of the crushing hammer are adjusted in real time according to the crushing target characteristics (such as material, hardness) and the current state of the crushing vehicle using the fuzzy PID control method. For example, when the crushing hammer reaches a crushing point with a hardness of level 7, the current crushing force is 68 tons, the deviation E = 2 tons (the expected crushing force is 70 tons) monitored by the sensor, and the deviation change rate EC = -1 ton / second (indicating that the crushing force deviation is decreasing). According to the fuzzy control rules, the PID parameters are adjusted to gradually approach the expected value of the crushing force, while ensuring the stability and efficiency of the crushing process.

[0181] During the crushing process, the status and operation effect of the breaker are continuously monitored by sensors. In addition to real-time monitoring of parameters such as crushing force, vibration, and position, the crushing effect is evaluated in real time through image recognition technology. For example, the image of the crushing area is collected and analyzed every 5 seconds, and indicators such as the area increase rate of the crushing area and the average particle size of the material after crushing are calculated to determine whether the crushing effect meets expectations. If it is found that the crushing force deviates from the expected value by a large margin (for example, the deviation exceeds ±5 tons) or the crushing effect is not good (for example, the area increase rate of the crushing area is lower than 80% of the expected value, or the average particle size of the material after crushing is greater than the set upper limit), the PID parameter adjustment amount is recalculated according to the new deviation and deviation change rate, and the parameters of the PID controller are updated.

[0182] Furthermore, suppose that at a certain moment, the actual crushing force is 65 tons, the expected crushing force is 70 tons, the deviation E = 70 - 65 = 5 tons, and the deviation change rate EC = -2 tons / second. According to the fuzzy control rules, the query obtains ∆K p =+0.12,∆K i =-0.06, ∆K d =+0.03. Then the updated PID parameters are: K p_new =K p +∆K p , K p_new =1.88 + 0.12 = 2; K i_new =K i +∆K i , K i_new =0.37 - 0.06 = 0.31; K d_new =K d +∆K d , K d_new =0.12 + 0.03 = 0.15.

[0183] Recalculate the control signal according to the new PID parameters, adjust the working parameters of the breaker, make the crushing force gradually approach the expected value, ensure the efficient and stable crushing operation, and continuously make dynamic adjustments based on real-time feedback to adapt to various complex situations, so as to ensure that the breaker can perform efficient and stable crushing operations according to the predetermined crushing path and strategy.

[0184] According to the embodiments of the present disclosure, by formulating fuzzy control rules and optimizing the parameters of the PID controller in real time, the method achieves precise control and dynamic adjustment of the breaker hammer, so that the breaker hammer can operate according to established paths and strategies, improves the stability and safety of the crushing operation, and reduces energy consumption and mechanical wear.

[0185] According to the embodiments of the present disclosure, the following technical effects are achieved: By integrating advanced technologies such as deep learning, cluster analysis, intelligent decision-making and fuzzy PID control, a set of efficient, precise and intelligent unmanned crushing vehicle hammer control methods and devices have been constructed. The device can perceive the working environment in real time, accurately identify and locate the crushing target, intelligently plan the crushing path and strategy, and realize precise control and dynamic adjustment of the crushing hammer, thereby effectively improving the working efficiency, operation accuracy and safety of the unmanned crushing vehicle, reducing labor costs and operational risks, and has broad market application prospects.

[0186] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present disclosure is not limited by the order of the actions described, because according to the present disclosure, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present disclosure.

[0187] The above is an introduction to the method embodiment. The following is a further explanation of the scheme disclosed in the present invention through an apparatus embodiment.

[0188] Figure 2 FIG. 1 shows a block diagram of a breaker hammer control device applied to an unmanned breaker vehicle according to an embodiment of the present disclosure; Figure 2 As shown, the breaker hammer control device 200 applied to the unmanned breaker vehicle may include: A first processing module 210 is used to obtain real-time operating environment data and pre-process the real-time operating environment data; The second processing module 220 is used to identify and locate the crushing target based on the pre-processed real-time operating environment data based on the deep learning algorithm; The third processing module 230 is used to cluster the recognition and positioning results using a clustering algorithm, perform cluster analysis based on the clustering results, and adjust the recognition and positioning results based on the cluster analysis results; The fourth processing module 240 is used to plan a crushing path and crushing strategy based on an intelligent decision-making algorithm according to the pre-processed real-time operating environment data and the adjusted identification and positioning results; The fifth processing module 250 is used to accurately control and dynamically adjust the breaker hammer based on the fuzzy PID control method according to the crushing path and crushing strategy, so that the breaker hammer can perform crushing operations.

[0189] Understandably, Figure 2Each module in the breaker hammer control device 200 applied to the unmanned crushing vehicle shown has the function of implementing each step in the breaker hammer control method 100 applied to the unmanned crushing vehicle provided by the embodiment of the present disclosure, and can achieve its corresponding technical effect. The specific working process of the described module can refer to the corresponding process in the aforementioned method embodiment. For the convenience and conciseness of the description, it will not be repeated here.

[0190] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0191] Figure 3 A block diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure is shown.

[0192] like Figure 3 As shown, electronic device 300 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0193] The electronic device 300 includes a computing unit 301, which can perform various appropriate actions and processes according to a computer program stored in a ROM 302 or a computer program loaded from a storage unit 308 into a RAM 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 can also be stored. The computing unit 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An I / O interface 305 is also connected to the bus 304.

[0194] A number of components in the electronic device 300 are connected to the I / O interface 305, including: an input unit 306, such as a keyboard, a mouse, etc.; an output unit 307, such as various types of displays, speakers, etc.; a storage unit 308, such as a disk, an optical disk, etc.; and a communication unit 309, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 309 allows the electronic device 300 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0195] The computing unit 301 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 301 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 301 performs the various methods and processes described above, such as method 100. For example, in some embodiments, the method 100 may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 308. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 300 via the ROM 302 and / or the communication unit 309. When the computer program is loaded into the RAM 303 and executed by the computing unit 301, one or more steps of the method 100 described above may be performed. Alternatively, in other embodiments, the computing unit 301 may be configured to perform the method 100 in any other appropriate manner (e.g., by means of firmware).

[0196] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0197] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0198] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0199] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0200] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0201] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0202] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0203] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A breaker hammer control method applied to an unmanned breaker truck, characterized in that: The method comprises: Acquiring real-time operating environment data and preprocessing the real-time operating environment data; Based on the deep learning algorithm, the crushing target is identified and located according to the pre-processed real-time operating environment data; Using clustering algorithms, clustering the recognition and positioning results, performing cluster analysis based on the clustering results, and adjusting the recognition and positioning results based on the cluster analysis results; Based on the intelligent decision-making algorithm, the crushing path and crushing strategy are planned according to the pre-processed real-time operating environment data and the adjusted identification and positioning results; Based on the fuzzy PID control method, the breaker hammer is precisely controlled and dynamically adjusted according to the crushing path and crushing strategy to enable the breaker hammer to perform crushing operations.

2. The method according to claim 1, characterized in that The acquiring of real-time operating environment data and preprocessing of the real-time operating environment data includes: By data collection, real-time operating environment data of the sensor network is acquired and data types are divided, and the real-time operating environment data is divided into image data and non-image data; Performing denoising, enhancement, cropping, and normalization processing on the image data; The non-image data is subjected to data cleaning, mode conversion and normalization processing.

3. The method according to claim 1, characterized in that The deep learning algorithm is used to identify and locate the crushing target according to the pre-processed real-time operating environment data, including: Based on the deep learning algorithm, feature extraction is performed on the pre-processed real-time operating environment data; According to the extracted features, a target detection algorithm is used to identify the broken target, and the broken target is roughly located; Based on the rough positioning, the crushing target is accurately positioned using a precise positioning algorithm.

4. The method according to claim 1, characterized in that The clustering algorithm is used to cluster the recognition and positioning results, and cluster analysis is performed according to the clustering results, and the recognition and positioning results are adjusted according to the clustering analysis results, including: Using a clustering algorithm, cluster the recognition and positioning results, cluster similar results into different clusters, and perform cluster analysis; wherein the cluster analysis includes: checking the compactness and separation of the clusters, whether there are noise points, and whether there are outliers that are incorrectly clustered; Adjust the recognition and positioning results according to the cluster analysis results.

5. The method according to claim 1, characterized in that The intelligent decision-making algorithm is used to plan the crushing path and crushing strategy according to the pre-processed real-time operating environment data and the adjusted identification and positioning results, including: Based on the intelligent decision-making algorithm, the pre-processed real-time working environment data and the adjusted recognition and positioning results are integrated into the environmental model; wherein the intelligent decision-making algorithm includes: intelligent path planning algorithm and intelligent strategy planning algorithm; the recognition and positioning results include: crushing target, crushing target characteristics and crushing target positioning; According to the positioning of the crushing target, an optimal path to each crushing target is planned in the environmental model using an intelligent path planning algorithm; According to the characteristics of the crushing target, combined with the best path to reach each crushing target and the current crushing vehicle performance, an intelligent strategy planning algorithm is used to plan a crushing strategy including each crushing target in the environmental model; the crushing strategy includes: crushing method, crushing parameters, and crushing sequence.

6. The method according to claim 1, characterized in that The fuzzy PID control method is based on the crushing path and crushing strategy to accurately control and dynamically adjust the crushing hammer so that the crushing hammer can perform crushing operations, including: According to the actual needs of the crushing operation, fuzzy control rules are formulated to clarify the PID parameter adjustment corresponding to different deviations and deviation change rates; The status information and operation effect of the breaker hammer are obtained in real time through sensors. Combined with fuzzy control rules, the proportional, integral and differential parameters of the PID controller are optimized to achieve precise control and dynamic adjustment of the breaker hammer, so that the breaker hammer can perform crushing operations according to the crushing path and crushing strategy.

7. A breaker hammer control device for an unmanned breaker truck, characterized in that: The device comprises: A first processing module, used for acquiring real-time operating environment data and preprocessing the real-time operating environment data; The second processing module is used to identify and locate the crushing target based on the pre-processed real-time operating environment data based on the deep learning algorithm; The third processing module is used to cluster the recognition and positioning results using a clustering algorithm, perform cluster analysis based on the clustering results, and adjust the recognition and positioning results based on the cluster analysis results; The fourth processing module is used to plan the crushing path and crushing strategy based on the intelligent decision-making algorithm according to the pre-processed real-time operating environment data and the adjusted identification and positioning results; The fifth processing module is used to accurately control and dynamically adjust the breaker hammer based on the fuzzy PID control method according to the crushing path and crushing strategy, so that the breaker hammer can perform crushing operations.

8. An electronic device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-6.

10. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Citation Information

Patent Citations

  • Self-optimization control device and method based on crushing working conditions

    CN109778937A

  • Clustering cluster quality evaluation method and device, equipment and storage medium

    CN110516752A

  • Mining stone crusher operation system

    CN112275437A

  • Unmanned driving control method and device for engineering vehicle, engineering vehicle and electronic equipment

    CN113282090A

  • Power line foreign matter detection method based on Faster R-CNN

    CN113989209A

Cited By

  • Intelligent mining crushing robot control method based on working condition recognition

    CN120618662A