CN113374021B - Excavator working condition identification method based on pilot control signal of operating handle - Google Patents
Excavator working condition identification method based on pilot control signal of operating handle Download PDFInfo
- Publication number
- CN113374021B CN113374021B CN202110703743.5A CN202110703743A CN113374021B CN 113374021 B CN113374021 B CN 113374021B CN 202110703743 A CN202110703743 A CN 202110703743A CN 113374021 B CN113374021 B CN 113374021B
- Authority
- CN
- China
- Prior art keywords
- excavator
- stage
- control signal
- pilot control
- working
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02F—DREDGING; SOIL-SHIFTING
- E02F9/00—Component parts of dredgers or soil-shifting machines, not restricted to one of the kinds covered by groups E02F3/00 - E02F7/00
- E02F9/20—Drives; Control devices
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02F—DREDGING; SOIL-SHIFTING
- E02F9/00—Component parts of dredgers or soil-shifting machines, not restricted to one of the kinds covered by groups E02F3/00 - E02F7/00
- E02F9/20—Drives; Control devices
- E02F9/2004—Control mechanisms, e.g. control levers
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02F—DREDGING; SOIL-SHIFTING
- E02F9/00—Component parts of dredgers or soil-shifting machines, not restricted to one of the kinds covered by groups E02F3/00 - E02F7/00
- E02F9/20—Drives; Control devices
- E02F9/2025—Particular purposes of control systems not otherwise provided for
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02F—DREDGING; SOIL-SHIFTING
- E02F9/00—Component parts of dredgers or soil-shifting machines, not restricted to one of the kinds covered by groups E02F3/00 - E02F7/00
- E02F9/20—Drives; Control devices
- E02F9/22—Hydraulic or pneumatic drives
Abstract
The invention provides an excavator working condition identification method based on an operating handle pilot control signal, which comprises the following steps: step 1, establishing an excavator working condition recognition deep learning model; step 2, collecting pilot control signals of an excavator operating handle; step 3, preprocessing the pilot control signal; step 4, constructing a characteristic vector according to the preprocessed pilot control signal; and 5, inputting the constructed characteristic vector into the working condition recognition deep learning model of the excavator to recognize the working cycle stage of the excavator, and outputting a recognition result by the working condition recognition deep learning model of the excavator. The invention adopts the control signal waveform of the operating handle as the identification mark, has high real-time performance, establishes the excavator working condition identification deep learning model based on the LSTM deep learning model capable of solving the long-term dependence of information to identify the working cycle stage of the excavator, has high accuracy, and can apply the identification result to the staged energy-saving control of the excavator.
Description
Technical Field
The invention relates to the technical field of engineering machinery, in particular to an excavator working condition identification method based on an operating handle pilot control signal.
Background
The excavator serving as a typical engineering machine becomes indispensable key earthwork construction equipment in projects such as city transformation, mineral resource development, track construction and the like by virtue of strong adaptability and high flexibility of the excavator under severe and variable working environments.
With the increasing demand for rapid delivery of construction projects, it is becoming a focus of project managers to reduce construction costs as much as possible while accelerating the project construction progress. On one hand, in order to guarantee the project construction progress, a project manager needs to master specific operation information of all devices and personnel on a construction site in real time, for example, in order to fully exert the operation capacity of an excavator, the working cycle stage of the excavator at each moment needs to be known, such as excavation, lifting and turning, unloading, empty bucket returning, moving, standby and the like, performance indexes such as cycle time, idle time, direct work rate and the like of the equipment are obtained through statistics and analysis according to the working cycle stage, the project manager can make project-related decisions such as resource allocation, work planning and scheduling and operator training and the like based on quantitative and visual performance indexes, and the data-driven project decision basis is the accurate analysis of the working conditions of the equipment including the excavator. On the other hand, the action rule and the load change of the actuating mechanism have obvious periodicity in the excavator operation process, and the existing step power control mode is difficult to meet the variable power requirement of the periodically changed load, so that the working point of the engine is far away from an economic working area, and a large amount of energy waste and environmental pollution are caused. In order to improve the energy utilization rate of construction equipment, excavator manufacturers build vehicle health monitoring systems, and aim of staged energy-saving control, energy conservation and emission reduction is realized by monitoring key operation parameters of the systems, wherein working condition identification of the excavator is an important basis for improving matching of an engine and a load. Therefore, the method for accurately identifying the working condition of the excavator in real time has important significance for improving project construction progress and reducing construction cost.
At present, a machine vision and multi-sensor information fusion technology is a main method for realizing the identification of the working condition of the excavator. The excavator working condition identification result precision based on machine vision is greatly influenced by factors such as viewpoint and scale deviation, ambient illumination and the like; the algorithm is complex and the calculation amount is large, so that the calculation in the travelling controller of the excavator is not facilitated; the method of collecting pictures through a camera preset on site-the remote computer completes recognition cannot reflect quantitative and visual work efficiency indexes to the improvement of the energy utilization rate of the excavator, namely cannot be used for real-time adjustment of an engine working point and meets the staged energy-saving control requirement of the excavator. The excavator working condition identification method based on the multi-sensor information fusion technology ignores the inherent hysteresis of the response of a hydraulic system, and inevitable delay errors can be generated during the working cycle phase conversion when the excavator working condition identification is carried out according to the motion information of an actuating mechanism or the performance parameters of the hydraulic system, so that the compliance of the whole machine operation and the response speed and effect of the staged energy-saving control are influenced.
In order to complete five working cycle stages in a complete working cycle in sequence, the excavator needs to control the actions of each actuating mechanism according to a certain sequence. The combinations of actuators acting in each working cycle phase are different, and the action of a key actuator often means the beginning of a working condition phase, for example, the lifting of a boom means the beginning of a lifting and turning phase. The original signal of the real-time reaction actuating mechanism is the control signal of the operating handle. Therefore, the control signal of the operating handle can be used as a basis for identifying the working cycle stage of the excavator.
Disclosure of Invention
The invention provides an excavator working condition identification method based on an operating handle pilot control signal, and aims to solve the problems that the traditional excavator working condition identification method based on machine vision and multi-sensor information fusion technology neglects error identification caused by response hysteresis of a hydraulic system, and identification results cannot be applied to staged energy-saving control of an excavator.
In order to achieve the above object, an embodiment of the present invention provides an excavator working condition identification method based on an operating handle pilot control signal, including:
step 2, collecting pilot control signals of an excavator operating handle;
step 3, preprocessing the pilot control signal;
and 7, when the identification process is finished, counting and displaying the identification result output by the excavator working condition identification deep learning model.
Wherein, the step 1 specifically comprises:
step 11, taking the pilot control signal waveform of the operating handle in the excavating preparation stage, the pilot control signal waveform of the operating handle in the excavating stage, the pilot control signal waveform of the operating handle in the lifting and turning stage, the pilot control signal waveform of the operating handle in the unloading stage and the pilot control signal waveform of the operating handle in the empty bucket returning stage as segmentation marks, segmenting the operating cycle of the excavator to obtain segmentation results, wherein the segmentation results comprise an excavating preparation stage, an excavating stage, a lifting and turning stage, an unloading stage and an empty bucket returning stage;
step 12, establishing an excavator working cycle stage recognition deep learning model based on an LSTM deep learning model, taking a segmentation mark as the input of the excavator working condition recognition deep learning model, taking a segmentation result as the output of the excavator working condition recognition deep learning model, and taking a mapping relation between the segmentation mark and the segmentation result as a digging preparation stage corresponding to the waveform of an operation handle pilot control signal of the digging preparation stage; the pilot control signal waveform of the operating handle corresponds to the excavation stage in the excavation stage; the lifting and turning stage is corresponding to the lifting and turning stage by the waveform of the pilot control signal of the operating handle; the unloading stage is corresponding to the unloading stage by the waveform of the pilot control signal of the operating handle; and the empty bucket return stage corresponds to the empty bucket return stage by operating the pilot control signal waveform of the handle.
Wherein, the step 2 specifically comprises:
when the pilot system is an electric control pilot system, a current sensor is adopted to collect pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link of the electric control pilot system; when the pilot system is a hydraulic control pilot system, a pressure sensor is adopted to collect pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link of the hydraulic control pilot system; and inputting the acquired pilot control signal into the locomotive controller.
Wherein, the step 3 specifically comprises:
step 31, noise interference in the pilot control signal is removed by adopting smooth filtering in the driving controller, as follows:
wherein the content of the first and second substances,
representing the filtered signal values, y (n-2), y (n-1), y (n +1) and y (n +2) being the original signal values;
step 32, reducing the frequency of the filtered pilot control signal by adopting a system sampling mode;
step 33, normalizing the pilot control signal with the reduced frequency as follows:
wherein x' represents the normalized signal value, x represents the signal value before normalization, x max Representing the maximum value, x, of a certain signal value in the feature vector min Representing the minimum value of a certain signal value in the feature vector.
Wherein, the step 4 specifically comprises:
reading the preprocessed pilot control signal, intercepting pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link included in a PWM signal period of a running vehicle controller when the pilot system is an electric control pilot system, and constructing a characteristic vector according to the sequence of the bucket link, the bucket rod link, the movable arm link and the rotation link; when the pilot system is a hydraulic control pilot system, intercepting pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link of the hydraulic control pilot system within 0.5 second, and constructing a characteristic vector according to the sequence of the bucket link, the bucket rod link, the movable arm link and the rotation link;
the feature vector is constructed as follows: :
(x a1 ,...,x ak ,x b1 ,...,x bl ,x c1 ,...,x cm ,x d1 ,...,x dn ) (3)
wherein x is a1 ,...,x ak The normalized bucket linkage pilot control signal discrete data is obtained; x is the number of b1 ,...,x bl The normalized discrete data of the pilot control signal of the bucket rod coupling is obtained; x is the number of c1 ,...,x cm Discrete data of the normalized boom linkage pilot control signal; x is a radical of a fluorine atom d1 ,...,x dn The normalized discrete data of the rotary joint pilot control signal is obtained.
Wherein, the step 5 specifically comprises:
inputting the constructed feature vector into an excavator working condition recognition deep learning model, and judging the working cycle stage of the excavator at the current moment according to the serial number of the maximum probability value in the excavator working condition recognition deep learning model output result; when the sequence number of the maximum probability value is 1, the working cycle stage of the excavator at the current moment is an excavating preparation stage; when the sequence number of the maximum probability value is 2, the working cycle stage of the excavator at the current moment is an excavating stage; when the sequence number of the maximum probability value is 3, the working cycle stage of the excavator at the current moment is a lifting rotation stage; when the serial number of the maximum probability value is 4, the working cycle stage of the excavator at the current moment is an unloading stage; and when the sequence number of the maximum probability value is 5, the working cycle stage of the excavator at the current moment is an empty bucket returning stage.
Wherein, the step 6 specifically comprises:
and when the identification process is not finished, removing the group of data read in earliest in the intercepted pilot control signals, adding the next group of pilot control signal data after the data read in latest, keeping the total data unchanged, repeatedly executing the steps 3 to 6 until no pilot control signal data are read in, and finishing the working condition identification of the excavator.
Wherein, the step 7 specifically comprises:
step 71, counting the work efficiency indexes of the excavator according to the obtained work cycle stages of the excavator at the current moment, wherein the work efficiency indexes comprise the number of times of completing each work cycle stage, the average consumed time of one complete work cycle, and the average consumed time and the occupation ratio of each stage in one complete work cycle, and when the work cycle stage of the excavator at the last moment is k, the cycle number of each work cycle stage is as follows:
wherein N is i The method comprises the steps of representing the cycle number of each working cycle phase, N representing the cycle number, i representing the working cycle phase, and k being 1,2,3,4 and 5 representing the stages of digging preparation, digging, lifting and turning, unloading and empty bucket returning respectively;
wherein T represents the average consumed time for completing one working cycle, and T represents the total consumed time for completing all complete working cycles of the excavator;
wherein, t i Represents the average time consumption of each working cycle phase in one working cycle, M i Representing the total times of occurrence of the value of each working cycle phase in the identification result sequence of the working cycle phases of the excavator, wherein delta t represents the sampling time interval of the driving controller;
wherein, W i Indicating one-off workThe average time consumption of each working cycle phase in the working cycle is proportional.
Wherein, the step 7 further comprises:
step 72, transmitting the excavator working condition identification result and the counted working efficiency index to a liquid crystal display screen through a CAN bus;
step 73, displaying the working cycle stage of the excavator in a curve form on a liquid crystal display screen, wherein the abscissa of the curve is the current moment, and the ordinate of the curve is the working cycle stage of the excavator at the current moment; when the ordinate value of the curve is 1, the working cycle phase of the excavator at the current moment is a digging preparation phase, when the ordinate value of the curve is 2, the working cycle phase of the excavator at the current moment is a digging phase, when the ordinate value of the curve is 3, the working cycle phase of the excavator at the current moment is a lifting rotation phase, when the ordinate value of the curve is 4, the working cycle phase of the excavator at the current moment is an unloading phase, when the ordinate value of the curve is 5, the working cycle phase of the excavator at the current moment is an empty bucket return phase;
step 74, displaying the excavator working efficiency index on the liquid crystal display screen in a form of a table;
and step 75, externally connecting a liquid crystal display screen with a controller, and uploading the working cycle stage of the excavator at the current moment and the counted working efficiency index to the excavator Internet of things platform through the controller to carry out remote monitoring on the working state of the excavator.
The scheme of the invention has the following beneficial effects:
according to the excavator working condition identification method based on the pilot control signal of the operating handle, the waveform of the operating handle control signal capable of reflecting the operating state of the actuating mechanism in real time is used as the identification mark, the instantaneity is high, the LSTM classifier capable of effectively solving the long-term dependence of information is adopted, the method is more suitable for the sequence data classification problem of the working cycle stage of the excavator, and the accuracy is high.
Drawings
FIG. 1 is a flow chart embodying the present invention;
FIG. 2 is a flow chart of the present invention;
fig. 3 is a schematic diagram of the LSTM classifier of the present invention.
Detailed Description
In order to make the technical problems, technical solutions and advantages of the present invention more apparent, the following detailed description is given with reference to the accompanying drawings and specific embodiments.
The invention provides an excavator working condition identification method based on an operating handle pilot control signal, which aims at the problems that the existing excavator working condition identification method based on machine vision and multi-sensor information fusion technology ignores error identification caused by response hysteresis of a hydraulic system and the identification result cannot be applied to the staged energy-saving control of an excavator.
As shown in fig. 1 to 3, an embodiment of the present invention provides an excavator working condition identification method based on an operation handle pilot control signal, including: step 1, establishing an excavator working condition recognition deep learning model; step 2, collecting pilot control signals of an excavator operating handle; step 3, preprocessing the pilot control signal; step 4, constructing a characteristic vector according to the preprocessed pilot control signal; step 5, inputting the constructed characteristic vector into an excavator working condition recognition deep learning model to recognize the working cycle stage of the excavator, and outputting a recognition result by the excavator working condition recognition deep learning model; step 6, judging whether the excavator working condition identification process is finished or not, and when the identification process is not finished, skipping to the step 3 to continue the identification of the working cycle stage of the excavator at the next moment; and 7, when the identification process is finished, counting and displaying the identification result output by the excavator working condition identification deep learning model.
Wherein, the step 1 specifically comprises: step 11, taking the pilot control signal waveform of the operating handle in the digging preparation stage, the pilot control signal waveform of the operating handle in the digging stage, the pilot control signal waveform of the operating handle in the lifting and turning stage, the pilot control signal waveform of the operating handle in the unloading stage and the pilot control signal waveform of the operating handle in the empty bucket returning stage as segmentation marks, segmenting the operating cycle of the digging machine to obtain segmentation results, wherein the segmentation results comprise the digging preparation stage, the digging stage, the lifting and turning stage, the unloading stage and the empty bucket returning stage; step 12, establishing an excavator working cycle stage recognition deep learning model based on an LSTM deep learning model, taking a segmentation mark as the input of the excavator working condition recognition deep learning model, taking a segmentation result as the output of the excavator working condition recognition deep learning model, and taking a mapping relation between the segmentation mark and the segmentation result as a digging preparation stage corresponding to the waveform of an operation handle pilot control signal of the digging preparation stage; the pilot control signal waveform of the operating handle corresponds to the excavation stage in the excavation stage; the lifting and turning stage is corresponding to the lifting and turning stage by the waveform of the pilot control signal of the operating handle; the unloading stage is corresponding to the unloading stage by the waveform of the pilot control signal of the operating handle; and the empty bucket return stage corresponds to the empty bucket return stage by operating the pilot control signal waveform of the handle.
Wherein, the step 2 specifically comprises: when the pilot system is an electric control pilot system, a current sensor is adopted to collect pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link of the electric control pilot system; when the pilot system is a hydraulic control pilot system, a pressure sensor is adopted to collect pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link of the hydraulic control pilot system; and inputting the acquired pilot control signal into the locomotive controller.
Wherein, the step 3 specifically comprises: step 31, smooth filtering is adopted in the running vehicle controller to remove noise interference in the pilot control signal, and the following steps are carried out:
wherein the content of the first and second substances,
representing the filtered signal values, y (n-2), y (n-1), y (n +1) and y (n +2) being the original signal values;
step 32, reducing the frequency of the filtered pilot control signal by adopting a system sampling mode;
step 33, normalizing the pilot control signal with the reduced frequency as follows:
wherein x' represents the signal value after normalization, x represents the signal value before normalization, x max Representing the maximum value, x, of a signal value in the feature vector min Representing the minimum value of a certain signal value in the feature vector.
According to the excavator working condition identification method based on the pilot control signal of the operating handle, the frequency of the pilot control signal is reduced in a system sampling mode according to the data processing capacity of the excavator controller, 5 frequencies including 100Hz, 50Hz, 25Hz, 20Hz and 10Hz are set, and the real-time performance of the identification working cycle stage is guaranteed.
Wherein, the step 4 specifically comprises: reading the preprocessed pilot control signal, intercepting pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link included in a PWM signal period of a running vehicle controller when the pilot system is an electric control pilot system, and constructing a characteristic vector according to the sequence of the bucket link, the bucket rod link, the movable arm link and the rotation link; when the pilot system is a hydraulic control pilot system, intercepting pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link of the hydraulic control pilot system within 0.5 second, and constructing a characteristic vector according to the sequence of the bucket link, the bucket rod link, the movable arm link and the rotation link;
the feature vector is constructed as follows:
(x a1 ,...,x ak ,x b1 ,...,x bl ,x c1 ,...,x cm ,x d1 ,...,x dn ) (3)
wherein x is a1 ,...,x ak The normalized data is the discrete data of the pilot control signal of the bucket link; x is the number of b1 ,...,x bl The normalized discrete data of the pilot control signal of the bucket rod coupling is obtained; x is the number of c1 ,...,x cm Discrete data of the normalized movable arm linkage pilot control signal are obtained; x is a radical of a fluorine atom d1 ,...,x dn For normalized rotary joint pilot control signalDiscrete data.
Wherein, the step 5 specifically comprises: inputting the constructed feature vector into an excavator working condition recognition deep learning model, and judging the working cycle stage of the excavator at the current moment according to the serial number of the maximum probability value in the excavator working condition recognition deep learning model output result; when the sequence number of the maximum probability value is 1, the working cycle stage of the excavator at the current moment is an excavating preparation stage; when the sequence number of the maximum probability value is 2, the working cycle stage of the excavator at the current moment is an excavating stage; when the sequence number of the maximum probability value is 3, the working cycle stage of the excavator at the current moment is a lifting rotation stage; when the serial number of the maximum probability value is 4, the working cycle stage of the excavator at the current moment is an unloading stage; and when the sequence number of the maximum probability value is 5, the working cycle stage of the excavator at the current moment is an empty bucket returning stage.
Wherein, the step 6 specifically comprises: and when the identification process is not finished, removing the group of data read in earliest in the intercepted pilot control signals, adding the next group of pilot control signal data after the data read in latest, keeping the total data unchanged, repeatedly executing the steps 3 to 6 until no pilot control signal data are read in, and finishing the working condition identification of the excavator.
Wherein, the step 7 specifically comprises: step 71, counting the work efficiency indexes of the excavator according to the obtained work cycle stages of the excavator at the current moment, wherein the work efficiency indexes comprise the number of times of completing each work cycle stage, the average consumed time of one complete work cycle, and the average consumed time and the occupation ratio of each stage in one complete work cycle, and when the work cycle stage of the excavator at the last moment is k, the cycle number of each work cycle stage is as follows:
wherein N is i Denotes the number of cycles per working cycle phase, N denotes the number of cycles, i denotes the working cycle phase, k is 1,2,3,4,5 denotes the digging preparation and digging respectivelyLifting and rotating, unloading and returning the empty bucket;
wherein T represents the average consumed time for completing one working cycle, and T represents the total consumed time for completing all complete working cycles of the excavator;
wherein, t i Represents the average time consumption of each working cycle phase in one working cycle, M i Representing the total times of occurrence of the value of each working cycle phase in the identification result sequence of the working cycle phases of the excavator, wherein delta t represents the sampling time interval of the traveling crane controller;
wherein, W i The average time consumption ratio of each working cycle phase in one working cycle is shown.
Wherein, the step 7 further comprises: step 72, transmitting the excavator working condition identification result and the counted working efficiency index to a liquid crystal display screen through a CAN bus; step 73, displaying the working cycle stage of the excavator in a curve form on a liquid crystal display screen, wherein the abscissa of the curve is the current moment, and the ordinate of the curve is the working cycle stage of the excavator at the current moment; when the ordinate value of the curve is 1, the working cycle phase of the excavator at the current moment is a digging preparation phase, when the ordinate value of the curve is 2, the working cycle phase of the excavator at the current moment is a digging phase, when the ordinate value of the curve is 3, the working cycle phase of the excavator at the current moment is a lifting rotation phase, when the ordinate value of the curve is 4, the working cycle phase of the excavator at the current moment is an unloading phase, when the ordinate value of the curve is 5, the working cycle phase of the excavator at the current moment is an empty bucket return phase; step 74, displaying the excavator working efficiency index on a liquid crystal display screen in a form of a table; and 75, externally connecting the liquid crystal display screen with a controller, and uploading the working cycle stage of the excavator at the current moment and the counted working efficiency index to the excavator Internet of things platform through the controller to carry out remote monitoring on the working state of the excavator.
According to the excavator working condition identification method based on the pilot control signal of the operating handle, the LSTM deep learning model adopts a 5-layer deep learning network structure and comprises an input layer, an RNN layer, a first full-connection layer, a second full-connection layer and an output layer; the input layer nodes are the number of elements contained in the feature vector, the RNN layer is an LSTM deep learning model containing 320 nodes, the first full-connection layer is provided with 80 nodes, the second full-connection layer is provided with 20 nodes, and the output layer is provided with 5 nodes matched with the number of the excavator activity types; a Sigmoid activation function is adopted between the input layer and the RNN layer, a tanh activation function is adopted between the RNN layer and the first full-connection layer, a ReLU activation function is adopted between the first full-connection layer and the second full-connection layer, and a softmax activation function is adopted between the second full-connection layer and the output layer.
According to the excavator working condition identification method based on the pilot control signal of the operating handle, the pilot control signal in the actual operation process of the excavator is collected, 28000 feature vectors are extracted from 50 groups of effective and complete excavation test data, and a sample space is formed together with corresponding output; the sample space is divided into 8: 2, dividing the ratio into a training set and a verification set; the deep learning model for identifying the working conditions of the excavator adopts an Adam algorithm with an initial learning rate of 0.001 to optimize and complete iterative updating of trainable parameters of the model, and the maximum iterative step number is set to be 50 steps; embedding the trained excavator working condition recognition deep learning model into a driving controller.
According to the excavator working condition identification method based on the pilot control signal of the operating handle, the waveform of the operating handle control signal which reflects the operation state of the actuating mechanism in real time is used as the identification mark, the instantaneity is high, the problem of sequence data classification of the working cycle stage of the excavator by establishing the excavator working condition identification deep learning model based on the LSTM deep learning model capable of solving the long-term dependence of information is solved, the accuracy is high, and the identification result can be applied to the staged energy-saving control of the excavator.
While the foregoing is directed to the preferred embodiment of the present invention, it will be understood by those skilled in the art that various changes and modifications may be made without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (6)
1. An excavator working condition identification method based on an operating handle pilot control signal is characterized by comprising the following steps:
step 1, establishing an excavator working condition recognition deep learning model, wherein the step 1 specifically comprises the following steps:
step 11, taking the pilot control signal waveform of the operating handle in the excavating preparation stage, the pilot control signal waveform of the operating handle in the excavating stage, the pilot control signal waveform of the operating handle in the lifting and turning stage, the pilot control signal waveform of the operating handle in the unloading stage and the pilot control signal waveform of the operating handle in the empty bucket returning stage as segmentation marks, segmenting the operating cycle of the excavator to obtain segmentation results, wherein the segmentation results comprise an excavating preparation stage, an excavating stage, a lifting and turning stage, an unloading stage and an empty bucket returning stage;
step 12, establishing an excavator working cycle stage recognition deep learning model based on an LSTM deep learning model, taking a segmentation mark as the input of the excavator working condition recognition deep learning model, taking a segmentation result as the output of the excavator working condition recognition deep learning model, and taking a mapping relation between the segmentation mark and the segmentation result as a digging preparation stage corresponding to the waveform of an operation handle pilot control signal of the digging preparation stage; the pilot control signal waveform of the operating handle corresponds to the excavation stage in the excavation stage; the lifting and turning stage is corresponding to the lifting and turning stage by the waveform of the pilot control signal of the operating handle; the unloading stage is corresponding to the unloading stage by the waveform of the pilot control signal of the operating handle; the empty bucket return stage is corresponding to the empty bucket return stage by the waveform of the pilot control signal of the operating handle;
step 2, collecting pilot control signals of an operating handle of the excavator, wherein the step 2 specifically comprises the following steps: when the pilot system is an electric control pilot system, a current sensor is adopted to collect pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link of the electric control pilot system; when the pilot system is a hydraulic control pilot system, a pressure sensor is adopted to collect pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link of the hydraulic control pilot system; inputting the acquired pilot control signal into a locomotive controller;
step 3, preprocessing the pilot control signal, wherein the step 3 specifically comprises:
step 31, noise interference in the pilot control signal is removed by adopting smooth filtering in the driving controller, as follows:
wherein the content of the first and second substances,
representing the filtered signal values, y (n-2), y (n-1), y (n +1) and y (n +2) being the original signal values;
step 32, reducing the frequency of the filtered pilot control signal by adopting a system sampling mode;
step 33, normalizing the pilot control signal with the reduced frequency, as follows:
wherein x' represents the signal value after normalization, x represents the signal value before normalization, x max Representing the maximum value, x, of a signal value in the feature vector min Representing the minimum value of a certain signal value in the feature vector;
step 4, constructing a characteristic vector according to the preprocessed pilot control signal;
step 5, inputting the constructed characteristic vector into an excavator working condition recognition deep learning model to recognize the working cycle stage of the excavator, and outputting a recognition result by the excavator working condition recognition deep learning model;
step 6, judging whether the excavator working condition identification process is finished or not, and when the identification process is not finished, skipping to the step 3 to continue the identification of the working cycle stage of the excavator at the next moment;
and 7, when the identification process is finished, counting and displaying the identification result output by the excavator working condition identification deep learning model.
2. The excavator working condition identification method based on the operating handle pilot control signal as claimed in claim 1, wherein the step 4 specifically comprises:
reading the preprocessed pilot control signal, intercepting pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link included in a PWM signal period of a running vehicle controller when the pilot system is an electric control pilot system, and constructing a characteristic vector according to the sequence of the bucket link, the bucket rod link, the movable arm link and the rotation link; when the pilot system is a hydraulic control pilot system, intercepting pilot control signals of a bucket link, a bucket rod link, a movable arm link and a rotation link of the hydraulic control pilot system within 0.5 second, and constructing a characteristic vector according to the sequence of the bucket link, the bucket rod link, the movable arm link and the rotation link;
the feature vector is constructed as follows:
(x a1 ,...,x ak ,x b1 ,...,x bl ,x c1 ,...,x cm ,x d1 ,...,x dn ) (3)
wherein x is a1 ,...,x ak The normalized data is the discrete data of the pilot control signal of the bucket link; x is the number of b1 ,...,x bl The normalized discrete data of the pilot control signal of the bucket rod coupling is obtained; x is a radical of a fluorine atom c1 ,...,x cm Discrete data of the normalized boom linkage pilot control signal; x is the number of d1 ,...,x dn The normalized discrete data of the rotary joint pilot control signal is obtained.
3. The excavator working condition identification method based on the operating handle pilot control signal as claimed in claim 2, wherein the step 5 specifically comprises:
inputting the constructed feature vector into an excavator working condition recognition deep learning model, and judging the working cycle stage of the excavator at the current moment according to the serial number of the maximum probability value in the excavator working condition recognition deep learning model output result; when the sequence number of the maximum probability value is 1, the working cycle stage of the excavator at the current moment is an excavating preparation stage; when the sequence number of the maximum probability value is 2, the working cycle stage of the excavator at the current moment is an excavating stage; when the sequence number of the maximum probability value is 3, the working cycle stage of the excavator at the current moment is a lifting rotation stage; when the serial number of the maximum probability value is 4, the working cycle stage of the excavator at the current moment is an unloading stage; and when the sequence number of the maximum probability value is 5, the working cycle stage of the excavator at the current moment is an empty bucket returning stage.
4. The excavator working condition identification method based on the operating handle pilot control signal as claimed in claim 3, wherein the step 6 specifically comprises:
and when the identification process is not finished, removing the group of data read in earliest in the intercepted pilot control signals, adding the next group of pilot control signal data after the data read in latest, keeping the total data unchanged, repeatedly executing the steps 3 to 6 until no pilot control signal data are read in, and finishing the working condition identification of the excavator.
5. The excavator working condition identification method based on the operating handle pilot control signal as claimed in claim 4, wherein the step 7 specifically comprises:
step 71, counting the work efficiency indexes of the excavator according to the obtained work cycle stages of the excavator at the current moment, wherein the work efficiency indexes comprise the number of times of completing each work cycle stage, the average consumed time of one complete work cycle, and the average consumed time and the occupation ratio of each stage in one complete work cycle, and when the work cycle stage of the excavator at the last moment is k, the cycle number of each work cycle stage is as follows:
wherein N is i The method comprises the steps of representing the cycle number of each working cycle phase, N representing the cycle number, i representing the working cycle phase, and k being 1,2,3,4 and 5 representing the stages of digging preparation, digging, lifting and turning, unloading and empty bucket returning respectively;
wherein T represents the average consumed time for completing one working cycle, and T represents the total consumed time for completing all complete working cycles of the excavator;
wherein, t i Represents the average time consumption of each working cycle phase in one working cycle, M i Representing the total times of occurrence of the value of each working cycle phase in the identification result sequence of the working cycle phases of the excavator, wherein delta t represents the sampling time interval of the traveling crane controller;
wherein, W i The average time consumption ratio of each working cycle phase in one working cycle is shown.
6. The method for identifying the operating condition of the excavator based on the pilot control signal of the operating handle as claimed in claim 5, wherein the step 7 further comprises:
step 72, transmitting the excavator working condition identification result and the counted working efficiency index to a liquid crystal display screen through a CAN bus;
step 73, displaying the working cycle stage of the excavator in a curve form on a liquid crystal display screen, wherein the abscissa of the curve is the current moment, and the ordinate of the curve is the working cycle stage of the excavator at the current moment; when the ordinate value of the curve is 1, the working cycle phase of the excavator at the current moment is a digging preparation phase, when the ordinate value of the curve is 2, the working cycle phase of the excavator at the current moment is a digging phase, when the ordinate value of the curve is 3, the working cycle phase of the excavator at the current moment is a lifting rotation phase, when the ordinate value of the curve is 4, the working cycle phase of the excavator at the current moment is an unloading phase, when the ordinate value of the curve is 5, the working cycle phase of the excavator at the current moment is an empty bucket return phase;
step 74, displaying the excavator working efficiency index on the liquid crystal display screen in a form of a table;
and 75, externally connecting the liquid crystal display screen with a controller, and uploading the working cycle stage of the excavator at the current moment and the counted working efficiency index to the excavator Internet of things platform through the controller to carry out remote monitoring on the working state of the excavator.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110703743.5A CN113374021B (en) | 2021-06-24 | 2021-06-24 | Excavator working condition identification method based on pilot control signal of operating handle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110703743.5A CN113374021B (en) | 2021-06-24 | 2021-06-24 | Excavator working condition identification method based on pilot control signal of operating handle |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113374021A CN113374021A (en) | 2021-09-10 |
| CN113374021B true CN113374021B (en) | 2022-09-09 |
Family
ID=77578885
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110703743.5A Active CN113374021B (en) | 2021-06-24 | 2021-06-24 | Excavator working condition identification method based on pilot control signal of operating handle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113374021B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114164878B (en) * | 2021-11-10 | 2023-07-04 | 中联重科土方机械有限公司 | Method, processor, system and excavator for identifying working conditions |
| CN116451809A (en) * | 2023-06-16 | 2023-07-18 | 北谷电子股份有限公司 | Excavator working condition identification method and system based on DAGSVM algorithm |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7050893B2 (en) * | 2000-03-31 | 2006-05-23 | Hitachi Construction Machinery Co., Ltd. | Method of detection of actual operating time of machinery deployed at construction sites, data collection and management system, and base station |
| JP2003194017A (en) * | 2001-12-27 | 2003-07-09 | Sumitomo (Shi) Construction Machinery Manufacturing Co Ltd | Hydraulic pressure circuit shutoff device for construction machine |
| CN107908874B (en) * | 2017-11-15 | 2021-10-29 | 上海华兴数字科技有限公司 | Working condition identification method and device and engineering mechanical equipment |
| CN108678063B (en) * | 2018-05-28 | 2021-02-02 | 重庆工商大学 | Engine control method and device based on automatic working condition recognition of loader |
| CN108978769B (en) * | 2018-07-03 | 2021-03-16 | 柳州柳工挖掘机有限公司 | Excavator working condition identification timing method and system and excavator |
-
2021
- 2021-06-24 CN CN202110703743.5A patent/CN113374021B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113374021A (en) | 2021-09-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113374021B (en) | Excavator working condition identification method based on pilot control signal of operating handle | |
| CN110110943B (en) | Fleet energy efficiency comprehensive intelligent optimization management system and optimization method based on big data | |
| CN111230887A (en) | Industrial gluing robot running state monitoring method based on digital twin technology | |
| CN113006188B (en) | Excavator staged power matching method based on LSTM neural network | |
| Shi et al. | Intelligent identification for working-cycle stages of excavator based on main pump pressure | |
| CN108427280A (en) | A kind of overhead crane anti-swing control method based on sliding mode control theory | |
| CN111651530A (en) | Intelligent port monitoring system | |
| CN114444816A (en) | Identification and control digital twin system of large-tonnage loader | |
| CN113157732B (en) | Underground scraper fault diagnosis method based on PSO-BP neural network | |
| CN101763087A (en) | Industrial process dynamic optimization system and method based on nonlinear conjugate gradient method | |
| CN113627268A (en) | Model training method, and method and device for detecting fault of speed reducer for mine hoist | |
| CN110188939B (en) | Wind power prediction method, system, equipment and storage medium of wind power plant | |
| CN113239720A (en) | Subway vehicle running gear fault diagnosis method based on deep migration learning | |
| CN1514395A (en) | signal machine for realizing traffic control based on image signal | |
| CN116681548A (en) | Intelligent building site management cloud platform based on BIM+GIS | |
| CN116081488A (en) | Unmanned control method for scene self-adaptive single-rail lifting robot | |
| JP3400062B2 (en) | Plant control device and tunnel ventilation control device | |
| CN114880967A (en) | Urban drainage dispatching management and control system capable of improving flood prevention and drainage work | |
| CN117446670B (en) | Automatic control method and system for tower crane based on man-machine co-fusion | |
| CN112084230A (en) | Surface water quality prediction method | |
| CN115478574B (en) | Excavator load prediction method based on radial basis function neural network | |
| CN110472741A (en) | A kind of small wave width study filtering system of three-domain fuzzy and method | |
| CN111182039A (en) | Road equipment Internet of things edge calculation method | |
| CN115114983B (en) | Method for acquiring and analyzing electric quantity data based on big data equipment and computer system | |
| CN117446670A (en) | Automatic control method and system for tower crane based on man-machine co-fusion |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |