A device and method for collecting agricultural machinery operating conditions

Abstract

This invention discloses an agricultural machinery operating condition data acquisition device and method, including a data acquisition module. The data acquisition module is sequentially connected to a sensor module, a controller module, a GPS positioning system module, and a camera module. The sensor module is sequentially connected to a temperature sensor unit, a pressure sensor unit, a fuel level sensor unit, and a speed sensor unit. The controller module is sequentially connected to a stroke controller unit and an engine control unit. This invention, based on the characteristics of agricultural operations and combined with the actual operation process, uses data to assist each other to identify the working and stopping states of the tractor, obtains all data after the tractor starts working, eliminates invalid data from non-normal working states, and performs relevant algorithm analysis to obtain the actual fuel consumption data during operation. This device can accurately reflect the different fuel consumption data consumed by different operating methods in the actual operation of agricultural machinery.

Description

Technical Field

[0001] This invention belongs to the field of agricultural machinery operating condition data acquisition technology, and specifically relates to an agricultural machinery operating condition data acquisition device and method. Background Technology

[0002] With the continuous upgrading of modern agricultural technology, tractors, as an important piece of agricultural machinery, have seen data collection and analysis become an important means to improve agricultural production efficiency and optimize agricultural production management. Tractor data acquisition devices can collect tractor operating data, including working hours, fuel consumption, equipment operating status, etc., and make scientific decisions and management through data analysis, model building, and other means, thereby improving agricultural production efficiency and management level.

[0003] In existing technologies, when tractor data acquisition devices collect data, it is difficult to accurately monitor fuel consumption data based on the actual operation process. This results in a large amount of invalid data being mixed in during data collection and algorithm analysis. Consequently, the obtained fuel consumption data deviates from the actual fuel consumption data during operation and cannot provide accurate assistance in improving agricultural machinery production efficiency.

[0004] Therefore, it is necessary to invent a device and method for collecting agricultural machinery operating conditions to solve the above problems, which can be used to accurately reflect the fuel consumption consumed during the operation of agricultural tractors. Summary of the Invention

[0005] To address the aforementioned problems, this invention provides an agricultural machinery operating condition data acquisition device and method to solve the issues raised in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an agricultural machinery operating condition acquisition device and method, comprising a data acquisition module, wherein the data acquisition module is respectively connected to a sensor module, a controller module, a GPS positioning system module and a camera module, the sensor module is sequentially connected to a temperature sensor unit, a pressure sensor unit, a fuel level sensor unit and a speed sensor unit, and the controller module is sequentially connected to a stroke controller unit and an engine control unit.

[0007] The data acquisition module is connected to the cloud platform via network signal.

[0008] Preferably, the sensor module is a core component of the tractor data acquisition module, capable of collecting real-time data on the tractor's operating environment and mechanical equipment; the temperature sensor monitors the temperature of the agricultural machinery engine system in real-time; the air pressure sensor monitors the air pressure of the agricultural machinery engine system in real-time; the fuel level sensor monitors the changes in fuel level in the agricultural machinery fuel tank system in real-time; and the speed sensor monitors the speed of the agricultural machinery in real-time. The data collected by the sensor acquisition module can be used for comprehensive monitoring and data analysis of the tractor's working status.

[0009] Preferably, the controller module is another core component of the tractor data acquisition module, and has the functions of controlling, converting and processing data; the stroke controller unit controls the stroke data of the agricultural machinery, the engine controller unit controls the working data of the agricultural machinery engine, and the controller module transmits the data of the stroke control unit and the engine controller unit, so that the stroke data and engine data of the agricultural machinery can be processed and analyzed.

[0010] Preferably, the cloud platform is a new technology that combines tractor data acquisition and analysis with cloud computing.

[0011] Preferably, the GPS positioning system module provides accurate latitude and longitude data of the agricultural machinery in real time for positioning the agricultural machinery in the operating area.

[0012] Preferably, the camera module allows for remote viewing of the on-site operation.

[0013] A method for collecting agricultural machinery operating conditions, characterized by the following steps:

[0014] Step 1: Connect the temperature sensor unit and the air pressure sensor unit to the agricultural tractor engine system through the sensor module to access the relevant engine operating status and collect data. Connect the fuel level sensor unit of the sensor module to the agricultural tractor fuel tank system to access the relevant real-time fuel consumption and collect data. Connect the speed sensor unit of the sensor module to the agricultural tractor speed control system to access the real-time operating speed of the tractor speed control system, collect data, and upload it to the data acquisition module.

[0015] Step 2: Connect the stroke control unit to the agricultural tractor stroke control system through the controller module, access the relevant working status of the tractor's working stroke, collect data, and connect the engine controller unit to the agricultural tractor engine system through the controller module, access the relevant working status of the tractor engine control, collect data, and upload the data to the data acquisition module.

[0016] Step 3: The GPS positioning system module provides accurate latitude and longitude data of the agricultural machinery in real time and uploads it to the data acquisition module;

[0017] Step 4: Based on the characteristics of agricultural operations and the actual operation process, the data from the temperature sensor unit, air pressure sensor unit, fuel level sensor unit, speed sensor unit, stroke controller unit and engine control unit are used to identify the working and stopping states of the tractor in different operations, obtain all the data of the tractor after starting, eliminate invalid data of non-normal working state and perform relevant algorithm analysis to obtain the actual fuel consumption data during the operation.

[0018] Step 5: The cloud platform enables refined and centralized management of tractor data, which can accurately reflect the different fuel consumption data of different operating methods in the actual working process of agricultural tractors.

[0019] Step 6: Conduct real-time video monitoring via the camera module.

[0020] The technical effects and advantages of this invention are as follows:

[0021] 1. This invention, based on the characteristics of agricultural operations and combined with the actual operation process, uses data from temperature sensor unit, air pressure sensor unit, fuel quantity sensor unit, speed sensor unit, stroke controller unit, and engine control unit to identify the working and stopping states of the tractor under different operations, obtains all data on the tractor's operation after startup, eliminates invalid data from non-normal working states, and performs relevant algorithm analysis to obtain the actual fuel consumption data during operation. This device can truly reflect the different fuel consumption data consumed by different operating methods in the actual operation of agricultural machinery.

[0022] 2. This invention constructs a working condition recognition model based on an autoencoder and a long short-term memory network (LSTM). This method can automatically identify complex working conditions, reduce human intervention, and further optimize the accuracy of fuel consumption data by using a throttle control unit and a fuel injection controller.

[0023] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a three-dimensional schematic diagram of the present invention;

[0026] Figure 2 This is a flowchart of the working condition identification algorithm of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Example 1: The present invention provides, as follows Figure 1 The agricultural machinery operating condition acquisition device and method shown includes a data acquisition module, which is connected to a sensor module, a controller module, a GPS positioning system module and a camera module respectively. The sensor module is sequentially connected to a temperature sensor unit, a pressure sensor unit, a fuel level sensor unit and a speed sensor unit. The controller module is sequentially connected to a stroke controller unit and an engine control unit.

[0029] The data acquisition module is connected to the cloud platform via network signal.

[0030] In one specific embodiment of the present invention, the sensor module is a core component of the tractor data acquisition module, capable of collecting data information on the tractor's operating environment and mechanical equipment in real time; the temperature sensor monitors the temperature of the agricultural machinery engine system in real time; the air pressure sensor monitors the air pressure of the agricultural machinery engine system in real time; the fuel level sensor monitors the changes in the fuel level of the agricultural machinery fuel tank system in real time; and the speed sensor monitors the movement speed of the agricultural machinery in real time. The data collected by the sensor acquisition module can be used for comprehensive monitoring and data analysis of the tractor's working status.

[0031] In one specific embodiment of the present invention, the controller module is another core component of the tractor data acquisition module, possessing the functions of controlling, converting, and processing data. The stroke controller unit controls the stroke data of the agricultural machinery, and the engine controller unit controls the operating data of the agricultural machinery engine. The controller module transmits data from the stroke control unit and the engine controller unit, enabling the processing and analysis of the agricultural machinery's stroke and engine data. The aforementioned stroke controller unit and engine controller unit are conventional control components and belong to known technology. The stroke controller can collect and analyze large amounts of real-time data and information, such as soil moisture and meteorological data, through sensors and monitoring equipment.

[0032] As a specific embodiment of the present invention, the cloud platform is a novel technology that combines tractor data acquisition and analysis with cloud computing.

[0033] As a specific embodiment of the present invention, the GPS positioning system module provides accurate latitude and longitude data of agricultural machinery in real time for positioning the agricultural machinery in the operating area.

[0034] In one specific embodiment of the present invention, the camera module remotely views the on-site operation.

[0035] As a specific embodiment of the present invention, a method for collecting agricultural machinery operating conditions is provided:

[0036] The sensor module is a core component of the tractor data acquisition device. It integrates various sensors to collect real-time data on the tractor's operating environment and mechanical equipment. The sensor module connects the temperature sensor unit and the air pressure sensor unit to the agricultural tractor's engine system to access the engine's relevant operating status and collect data. The sensor module connects the fuel level sensor unit to the agricultural tractor's fuel tank system to access the real-time fuel consumption and collect data. The sensor module connects the speed sensor unit to the agricultural tractor's speed control system to access the real-time operating speed of the tractor's speed control system, collect data, and upload it to the data acquisition module.

[0037] The controller is another core component of the tractor data acquisition device. It has the functions of controlling, converting and processing data. The controller module connects the stroke control unit to the agricultural tractor stroke control system, accesses the working status related to the tractor's working stroke, and collects data. The controller module also connects the engine controller unit to the agricultural tractor engine system, accesses the working status related to the tractor engine control, collects data, and uploads it to the data acquisition module.

[0038] The GPS positioning system module provides accurate latitude and longitude data of agricultural machinery in real time, uploads it to the data acquisition module, and uses the camera module to conduct auxiliary real-time video monitoring to locate the position and working status of the tractor.

[0039] Based on the characteristics of agricultural operations and the actual operation process, data from temperature sensor unit, air pressure sensor unit, fuel level sensor unit, speed sensor unit, stroke controller unit, and engine control unit are used to identify the working and stopping states of the tractor under different conditions. All data on the tractor's operation after startup is obtained, and invalid data from non-working states is eliminated. Invalid data includes fuel consumption data generated by movement during non-working states, fuel consumption data generated during tractor maintenance, and fuel consumption data when the tractor's fuel tank is abnormally depleted. Relevant algorithm analysis is then performed to obtain the actual fuel consumption data during operation.

[0040] Tractor fuel consumption calculation formula:

[0041] The tractor's fuel consumption per 100 kilometers is calculated based on the tractor's fuel consumption data after removing invalid data and the mileage data after removing invalid mileage data. The calculation formula is as follows:

[0042] Fuel consumption per 100 kilometers = fuel consumption (liters) ÷ mileage (kilometers) × 100;

[0043] Under the same operating mode, the actual fuel consumption of a tractor can be obtained by controlling a single variable, depending on the different loads, road conditions, and ambient temperatures.

[0044] By using data from temperature sensor unit, air pressure sensor unit, fuel quantity sensor unit, speed sensor unit, stroke controller unit and engine control unit to assist each other, different operating methods in the actual working process of the tractor can be identified. Different real fuel consumption data are obtained according to different types of operations. The cloud platform calculation realizes refined and centralized management of tractor data, which can truly reflect the different fuel consumption data consumed by different operating methods in the actual working process of agricultural tractors.

[0045] Tractor fuel consumption calculation formula:

[0046] The tractor's fuel consumption per 100 kilometers is calculated based on the tractor's fuel consumption data after removing invalid data and the mileage data after removing invalid mileage data. The calculation formula is as follows:

[0047] Fuel consumption per 100 kilometers = fuel consumption (liters) ÷ mileage (kilometers) × 100;

[0048] Under the same load, road conditions, and ambient temperature, the actual fuel consumption of a tractor under different operating modes can be determined.

[0049] Example 2: By constructing a working condition recognition model based on an autoencoder and a long short-term memory network (LSTM), the algorithm can learn to extract representative and important features from multi-dimensional data (temperature, air pressure, fuel level, speed) collected by sensors. The autoencoder can extract core features from the data by compressing and reconstructing it, while the LSTM can capture long-term dependencies in the time series, thereby more accurately identifying different working conditions of agricultural tractors. This method can automatically identify complex operating conditions, reduce human intervention, and enable the throttle control unit and fuel injection controller to further optimize the accuracy of fuel consumption data.

[0050] Step 1: Data Preprocessing

[0051] First, multi-dimensional data is collected through sensor modules; then, the data from each sensor is standardized to prevent data of different scales from adversely affecting model training; the data is then organized in chronological order to construct a time series dataset so that the LSTM model can capture time dependencies in the future.

[0052] Data Acquisition: The sensor module continuously collects multi-dimensional data such as temperature, air pressure, fuel level, and speed; among them, the temperature sensor (T) records the temperature when the engine is running; the air pressure sensor (P) records the air pressure at the engine intake, within a certain range; the fuel level sensor (F) records the real-time fuel level in the tractor's fuel tank; and the speed sensor (S) records the real-time speed of the tractor.

[0053] Data standardization: In multi-dimensional data, the dimensions of data such as temperature (°C), air pressure (kPa), fuel volume (L), and speed (km / h) are different. Directly inputting them into the model will lead to poor training results. Therefore, it is necessary to standardize the data from each sensor and unify them to the same scale.

[0054] The min-max normalization method is used to scale the data to the range [0, 1]. The normalization formula is as follows:

[0055]

[0056] Where x represents the original data point, i.e., the original value to be normalized; x min x represents the minimum value of the original data, indicating the theoretical minimum or the lowest value within the actual measurement range of the dataset; max x is the maximum value of the original data, representing the theoretical maximum value or the highest value in the actual measurement range of the dataset; x′ is the normalized data value, ranging from [0, 1], used to unify data of different dimensions to the same scale, which is convenient for subsequent algorithm processing;

[0057] Standardized data will be used to train machine learning models to avoid the influence of different units on the models;

[0058] Time series construction: Since the working conditions of agricultural tractors have strong time correlations, the LSTM model needs time series data to capture the time dependence of various tractor parameters. In order to construct a time series dataset, a time window is set, that is, each input time window is a continuous time step of data to predict the next state of the tractor.

[0059] Numerical examples are as follows:

[0060] Data Acquisition: During a certain time period, the sensor module collected the following multi-dimensional data: temperature sensor (T) 80℃; air pressure sensor (P) 100kPa; fuel level sensor (F) 120L; speed sensor (S) 35km / h;

[0061] Data standardization: The theoretical minimum and maximum values ​​for the sensor are: temperature range of 0℃ to 120℃; air pressure range of 80kPa to 120kPa; fuel volume range of 0L to 200L; speed range of 0km / h to 50km / h.

[0062] Using min-max normalization, the normalized results are: Temperature T′=(80-0) / (120-0)=0.6667; Air pressure P′=(100-80) / (120-80)=0.5; Oil quantity F′=(120-0) / (200

[0063] -0)=0.6; Velocity S′=(35-0) / (50-0)=0.7;

[0064] The standardized data is: X = [0.6667, 0.5, 0.6, 0.7]

[0065] Time series construction: The time window is set to data with 3 time steps. The input is:

[0066] X1 = [0.6667, 0.5, 0.6, 0.7]

[0067] X2 = [0.7, 0.55, 0.62, 0.72]

[0068] X3 = [0.65, 0.52, 0.61, 0.71]

[0069] Use this data to build a time series dataset and predict the data for the next time step.

[0070] Step 2 Feature Extraction

[0071] An autoencoder mainly consists of two parts: an encoder and a decoder. The autoencoder compresses high-dimensional sensor data into a low-dimensional feature representation by the encoder, extracting the core features from the data. Then, the decoder reconstructs these low-dimensional features back into the original data, ensuring that the extracted features retain as much information as possible from the original data.

[0072] 2.1 Encoder Section

[0073] The function of an encoder is to compress high-dimensional input data into low-dimensional feature representations;

[0074] The encoder outputs a low-dimensional feature vector Z, whose formula is:

[0075] Z = f(W) e ·X+b e )

[0076] Where X represents the sensor input data vector, containing multi-dimensional data of T (temperature), P (air pressure), F (fuel quantity), and S (speed); W e The weight matrix represents the encoder, used to map the input data to a low-dimensional space; b e represents the encoder bias term; f represents the activation function, using a non-linear activation function (ReLU) to increase the model's expressive power; Z represents the low-dimensional feature representation, which represents the features of the input data after compression; these features will be used in subsequent LSTM networks for time series modeling;

[0077] 2.2 Decoder Section

[0078] The decoder's role is to reconstruct the low-dimensional feature representation Z into original data that is as close as possible to the input X, ensuring that the extracted features retain key information from the sensor data; the decoder's output formula is:

[0079]

[0080] in W represents the reconstructed output data, which is expected to be as close as possible to the original input X; d The weight matrix of the decoder is used to map low-dimensional features back to high-dimensional space; b d represents the bias term of the decoder; g represents the activation function, which is the same activation function used in the encoder; Z represents the low-dimensional features obtained from the encoder;

[0081] 2.3 Loss Function

[0082] The goal of an autoencoder is to minimize the difference between the input data X and the reconstructed data. The difference between them is used to ensure the effectiveness of feature extraction; the loss function here is the mean squared error (MSE), and its formula is:

[0083]

[0084] Where L represents the loss function, indicating the difference between the input and the reconstructed output; n represents the dimension of the input data; X i This represents the i-th data point in the original input; This represents the i-th data point after reconstruction;

[0085] By minimizing the loss function L, the autoencoder can learn key features in the sensor data and retain as much original information as possible.

[0086] 2.4 Numerical Examples

[0087] Encoder section: The input multidimensional sensor data is X = [T, P, F, S], and the data is compressed using a simple 2D encoder.

[0088] The encoder output is:

[0089] Z = f(W) e ·X+b e )

[0090] W e =[[0.2, 0.3, 0.5, 0.1], [0.4, 0.2, 0.3, 0.6]]

[0091] b e =[0.1,0.2]

[0092] Use ReLU as the activation function.

[0093] The calculation result is:

[0094] Z=f([0.2·0.6667+0.3·0.5+0.5·0.6+0.1·0.7]

[0095] +0.1[0.4·0.6667+0.2·0.5+0.3·0.6+0.6·0.7]+0.2)

[0096] Calculate each element:

[0097] Z1 = f(0.6133 + 0.1) = 0.7133

[0098] Z2 = f(0.91334 + 0.2) = 1.11334

[0099] Obtain encoder output:

[0100] Z = [0.7133, 1.11334]

[0101] Decoder section: The decoder uses the following weight matrix:

[0102] W a =[[0.3,0.4],[0.2,0.1],[0.5,0.2],[0.3,0.6]]

[0103] b d =[0.1,0.2,0.3,0.4]

[0104] Decoder output:

[0105]

[0106] calculate:

[0107]

[0108] The reconstructed data is as follows:

[0109]

[0110] Loss function (MSE): Given the original data X = [0.6667, 0.5, 0.6, 0.7], the MSE loss function is:

[0111]

[0112] calculate:

[0113]

[0114] Step 3: Time Series Modeling

[0115] The convolutional features are modeled in time series using an LSTM network. LSTM is suitable for processing time series data and capturing the state changes of agricultural tractors under different working conditions. The LSTM network learns the time dependence of sensor data step by step and can understand the changes of various working conditions of the tractor over time. The features extracted from the autoencoder are used as the input of the LSTM, and the working conditions are modeled through multiple layers of LSTM to predict the changes in working conditions in the future.

[0116] 3.1 LSTM Cell Design

[0117] LSTM networks are a special version of recurrent neural networks (RNNs) used to capture temporal dependencies, making them particularly suitable for processing time-series data. They learn long-term and short-term dependencies in sensor data through memory units, thereby understanding changes in tractor operating conditions. The core structure of an LSTM unit represents...

[0118] The ForgetGate is used to determine the historical information that needs to be forgotten; the formula is as follows:

[0119] f t =σ(W f ·[h t-1 x t ]+b f )

[0120] Where f t The output of the forget gate, with a value between 0 and 1, indicates how much past information is retained; W f The weight matrix of the forget gate; h t-1 Indicates the hidden state at the previous time step; xt b represents the feature vector input at the current time step; f σ represents the bias term of the forget gate; σ represents the activation function Sigmoid, which compresses the result to between 0 and 1.

[0121] The input gate is used to determine the current information that needs to be updated; the formula is as follows:

[0122] i t =σ(W i ·[h t-1 x t ]+b i )

[0123] Where i t The output of the input gate controls the proportion of the current input that is retained; W i b represents the weight matrix of the input gate; i Indicates the bias term of the input gate; x t h represents the feature vector input at the current time step. t-1 This indicates the hidden state of the previous time step;

[0124] The candidate cell state (CandidateCellState) is used to generate the current candidate cell content; the formula is as follows:

[0125]

[0126] in W represents the candidate memory state at the current time step. C The weight matrix representing the candidate memory states; b C The bias term represents the candidate memory state; tanh represents the hyperbolic tangent function, which compresses the result to between -1 and 1; h t-1 Indicates the hidden state at the previous time step; x t This represents the feature vector input at the current time step;

[0127] CellStateUpdate is used to update the memory state through the forget gate and the input gate; the formula is as follows:

[0128]

[0129] Where C represents t Indicates the memory state at the current time step; C t-1 Indicates the memory state of the previous time step; i t This indicates the output of the input gate, controlling the proportion of the current input that is retained; Indicates the candidate memory state at the current time step; f t Indicates the output of the forget gate;

[0130] The output gate (OutputGate) is used to determine the hidden state of the output; the formula is as follows:

[0131] o t =σ(W o ·[h t-1 x t ]+b o )

[0132] The formula for calculating the hidden state is as follows:

[0133] h t =o t ·tanh(C t )

[0134] Among them o t The output of the output gate is represented by σ; the sigmoid activation function is represented by W. o The weight matrix [h] represents the output gate. t-1 x t ] represents the hidden state h of the previous time step. t-1 With the input x at the current time step t The concatenated vector; b o Represents the bias term of the output gate; C t Represents the memory state at the current time step; tanh represents the hyperbolic tangent function; C t This indicates the memory state at the current time step;

[0135] 3.2 Input and Output Sequences

[0136] Input sequence: In this invention, the feature map extracted from the autoencoder is used as the input of the LSTM network. The LSTM network models the time series data to understand the changes in the working conditions of the tractor.

[0137] Output sequence: The output of the LSTM network will be used to predict the operating condition of the tractor over a future period; the state output by the LSTM network is h. t LSTM can capture the impact of past data on the future, helping to predict the next state of the tractor;

[0138] Time series modeling is performed using a multi-layer LSTM, and the final prediction results are output:

[0139] y t =LSTM(x1, x2, ..., x t )

[0140] Where y tThis is the output of the current time step (which can represent the state of the fuel-consuming tractor: a prediction of normal operation, standby, or low-load operation), indicating the current operating condition of the tractor; x1, x2, ..., x t It is a time series of sensor input features;

[0141] 3.3 Algorithm Improvement

[0142] Traditional LSTM only considers information from the past to the present, but in some cases, future state information can also help with current predictions. By introducing bidirectional LSTM, the model can consider both past and future time step information, thus more accurately modeling the working conditions of tractors.

[0143] Bidirectional LSTM consists of two LSTM networks, one processing time series from front to back and the other from back to front; their outputs are the hidden states from both directions concatenated together to form the final feature representation.

[0144] Hidden state of a forward LSTM for:

[0145]

[0146] Hidden state of backward LSTM for:

[0147]

[0148] The final output of the bidirectional LSTM is:

[0149]

[0150] in: The hidden state of the forward LSTM captures past information at time step t; The hidden state of the backward LSTM captures future information at time step t; h t The output of the bidirectional LSTM contains comprehensive features of the past and future (which can characterize the state of the fuel-consuming tractor: predictions of normal operation, standby, or low-load operation).

[0151] 3.4 Numerical Examples

[0152] LSTM unit design: The features Z = [0.7133, 1.11334] extracted from the autoencoder are input into the LSTM.

[0153] Forget Gate Calculation:

[0154] f t =σ(W f ·[h t-1 ,Z]+bf )

[0155] in:

[0156] W f =[0.3,0.2]

[0157] b f =0.1

[0158] The hidden state h of the previous time step t-1 =0.5

[0159] calculate:

[0160] f t =σ(0.3·0.5+0.2·0.7133+0.1)=σ(0.35+0.14266+0.1)

[0161] =σ(0.59266) = 0.6448

[0162] Input gate calculation: Input gate weight W i = [0.2, 0.4], bias b i =0.2, then:

[0163] i t =σ(W i ·[h t-1 ,Z]+b i )=σ([0.2·0.5+0.4·0.7133]+0.2)

[0164] i t =σ(0.1+0.28532+0.2)=σ(0.58532)=0.6422

[0165] Output gate calculation: Calculate the output gate o t The weight matrix and bias are:

[0166] o t =σ(W o ·[h t-1 ,Z]+b o )

[0167] Hidden state update: The hidden state at the current time step is calculated last.

[0168] h t =o t ·tanh(C t )

[0169] Step 4: Fuel Consumption Data Optimization

[0170] 4.1 Through the above working condition identification model, the tractor fuel consumption data under different working conditions can be automatically identified to reduce the impact of fuel consumption from ineffective operations; combined with the actual fuel consumption data under different working conditions, the fuel consumption management and operating efficiency of the tractor can be further optimized.

[0171] This implementation process combines two algorithms, autoencoder and LSTM, which has advantages in processing multi-dimensional sensor data and time series data, and can efficiently realize the automated identification of the working conditions of agricultural tractors.

[0172] Specifically, it is through the above formula

[0173] y t =LSTM(x1, x2, ..., x t )

[0174] A Long Short-Term Memory (LSTM) network model is used to predict the operating conditions at future times, and the final fuel consumption is calculated based on the prediction results; the LSTM output at each time step y t Each of these corresponds to a fuel consumption rate, and these fuel consumption rates are recorded.

[0175] The formula for calculating fuel consumption is: Fuel consumption = Fuel consumption rate × Time interval.

[0176] The total fuel consumption for the entire time series is obtained by summing up the fuel consumption at all time steps:

[0177]

[0178] Where T is the total number of time steps, and fuel consumption rate is... t It is the fuel consumption rate at the t-th time step;

[0179] 4.2 Numerical Cases

[0180] There are two main operating conditions for tractors:

[0181] Normal operation means the tractor is tilling the field at a high speed and with a heavy engine load; standby or low-load operation means the tractor is in standby or low-speed mode with relatively low fuel consumption.

[0182] In these two states, the fuel consumption data will vary depending on the operating conditions; the fuel consumption is 10L / hour when normal operation is detected; and the fuel consumption is 3L / hour when standby is detected.

[0183] The following LSTM prediction outputs are given (each time step is 1 hour):

[0184] Time step 1: y1 is predicted to be "normal operation", with a fuel consumption rate of 10L / hour;

[0185] Time step 2: y2 is predicted to be "standby", with a fuel consumption rate of 3L / hour;

[0186] Time step 3: Y3 is predicted to be "normal operation", with a fuel consumption rate of 10L / hour;

[0187] Time step 4: Y4 is predicted to be "standby", with a fuel consumption rate of 3L / hour;

[0188] Calculate fuel consumption at each time step

[0189] Time step 1: Fuel consumption = 10L / hour × 1 hour = 10L

[0190] Time step 2: Fuel consumption = 3L / hour × 1 hour = 3L

[0191] Time step 3: Fuel consumption = 10L / hour × 1 hour = 10L

[0192] Time step 4: Fuel consumption = 3L / hour × 1 hour = 3L

[0193] Therefore, the total fuel consumption is:

[0194] Total fuel consumption = 10 + 3 + 10 + 3 = 26L.

[0195] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An agricultural machinery operating condition acquisition device, comprising a data acquisition module, characterized in that: The data acquisition module is connected to a sensor module, a controller module, a GPS positioning system module, and a camera module. The sensor module is sequentially connected to a temperature sensor unit, a bar pressure sensor unit, a fuel level sensor unit, and a speed sensor unit. The controller module is sequentially connected to a stroke controller unit and an engine control unit. The data acquisition module is connected to a cloud platform via a network signal; The sensor module is a core component of the tractor data acquisition module, capable of collecting real-time data on the tractor's operating environment and mechanical equipment; the temperature sensor monitors the temperature of the agricultural machinery engine system in real time. A pressure sensor monitors the air pressure of the agricultural machinery engine system in real time. The fuel level sensor monitors the fuel level changes of the agricultural machinery's fuel tank system in real time; the speed sensor monitors the movement speed of the agricultural machinery in real time; the data collected by the sensor acquisition module can be used for comprehensive monitoring and data analysis of the tractor's working status. The controller module is another core component of the tractor data acquisition module, and it has the functions of controlling, converting and processing data. The stroke controller unit controls the stroke data of the agricultural machinery, and the engine controller unit controls the working data of the agricultural machinery engine. The controller module transmits the data of the stroke control unit and the engine controller unit, so that the stroke data and engine data of the agricultural machinery can be processed and analyzed. The cloud platform is a technology that combines tractor data collection and analysis with cloud computing; Based on the characteristics of agricultural operations and combined with the actual operation process, the data from the temperature sensor unit, air pressure sensor unit, fuel quantity sensor unit, speed sensor unit, stroke controller unit and engine control unit assist each other. According to the working condition recognition algorithm, the working state and stopping state of the tractor under different operations are identified, and all data of the tractor after starting are obtained. Invalid data of non-normal working state are eliminated and relevant algorithm analysis is performed to obtain the actual fuel consumption data during the operation. Cloud platform computing enables refined and centralized management of tractor data, and can accurately reflect the different fuel consumption data of different operating methods in the actual working process of agricultural tractors.

2. The agricultural machinery operating condition acquisition device according to claim 1, characterized in that: The GPS positioning system module provides accurate latitude and longitude data for agricultural machinery in real time, enabling the positioning of agricultural machinery in the operating area.

3. The agricultural machinery operating condition acquisition device according to claim 1, characterized in that: The camera module allows for remote monitoring of on-site operations.

4. A method for collecting agricultural machinery operating conditions, applied to the agricultural machinery operating condition collection device according to any one of claims 1-3, characterized in that, The method is as follows: Step 1: Connect the temperature sensor unit and the air pressure sensor unit to the agricultural tractor engine system through the sensor module to access the relevant engine operating status and collect data. Connect the fuel level sensor unit of the sensor module to the agricultural tractor fuel tank system to access the relevant real-time fuel consumption and collect data. Connect the speed sensor unit of the sensor module to the agricultural tractor speed control system to access the real-time operating speed of the tractor speed control system, collect data, and upload it to the data acquisition module. Step 2: Connect the stroke control unit to the agricultural tractor stroke control system through the controller module, access the relevant working status of the tractor's working stroke, collect data, and connect the engine controller unit to the agricultural tractor engine system through the controller module, access the relevant working status of the tractor engine control, collect data, and upload the data to the data acquisition module. Step 3: The GPS positioning system module provides accurate latitude and longitude data of the agricultural machinery in real time and uploads it to the data acquisition module; Step 4: Based on the characteristics of agricultural operations and the actual operation process, the data from the temperature sensor unit, air pressure sensor unit, fuel quantity sensor unit, speed sensor unit, stroke controller unit and engine control unit are used to assist each other. According to the working condition recognition algorithm, the working state and stopping state of the tractor under different operations are identified, and all data of the tractor after starting are obtained. Invalid data of non-normal working state are eliminated and relevant algorithm analysis is performed to obtain the actual fuel consumption data during the operation. Step 5: The cloud platform enables refined and centralized management of tractor data, which can accurately reflect the different fuel consumption data of different operating methods in the actual working process of agricultural tractors; Step 6: Conduct real-time video monitoring via the camera module.

5. The method for collecting agricultural machinery operating conditions according to claim 4, characterized in that, The working condition identification algorithm mentioned in step four has the following specific steps: By constructing a working condition recognition model based on autoencoders and long short-term memory networks (LSTM), the algorithm can learn to extract representative and important features from multi-dimensional data such as temperature, air pressure, fuel level, and speed collected by sensors. The autoencoder can extract the core features of the data by compressing and reconstructing the data, while the LSTM can capture long-term dependencies in the time series, thereby more accurately identifying different working conditions of agricultural tractors. Step 1: Data Preprocessing First, multi-dimensional data is collected through sensor modules; then, the data from each sensor is standardized to prevent data of different scales from adversely affecting model training; the data is then organized in chronological order to construct a time series dataset so that the LSTM model can capture time dependencies in the future. Data acquisition: The sensor module continuously collects multi-dimensional data such as temperature, air pressure, fuel level, and speed; The temperature sensor records the engine temperature T during operation; the air pressure sensor records the air pressure P at the engine intake; the fuel level sensor records the real-time fuel level F in the tractor's fuel tank; and the speed sensor records the real-time tractor speed S. Data standardization: In multi-dimensional data, the dimensions of temperature, air pressure, fuel volume, and speed are different. Directly inputting these data into the model will lead to poor training results. Therefore, it is necessary to standardize the data from each sensor and unify them to the same scale. The min-max normalization method is used to scale the data to the range [0, 1]. The normalization formula is as follows: Where x represents the original data point, i.e., the original value to be normalized; x min x represents the minimum value of the original data, indicating the theoretical minimum or the lowest value within the actual measurement range of the dataset; max x is the maximum value of the original data, representing the theoretical maximum value or the highest value in the actual measurement range of the dataset; x′ is the normalized data value, ranging from [0, 1], used to unify data of different dimensions to the same scale, which is convenient for subsequent algorithm processing; Standardized data will be used to train machine learning models to avoid the influence of different units on the models; Time series construction: Since the working conditions of agricultural tractors have strong time correlations, the LSTM model needs time series data to capture the time dependence of various tractor parameters. In order to construct a time series dataset, a time window is set, that is, each input time window is a continuous time step of data to predict the next state of the tractor. Step 2 Feature Extraction An autoencoder mainly consists of two parts: an encoder and a decoder. The autoencoder compresses high-dimensional sensor data into a low-dimensional feature representation through the encoder, extracting the core features from the data. Then, these low-dimensional features are reconstructed into the original data through a decoder, ensuring that the extracted features retain as much information as possible from the original data. 2.1 Encoder Section The function of an encoder is to compress high-dimensional input data into low-dimensional feature representations; The encoder outputs a low-dimensional feature vector Z, whose formula is: Z=f(W e ·X+b e ) Where X represents the sensor input data vector, containing multidimensional data such as temperature T, air pressure P, fuel quantity F, and speed S; W e The weight matrix represents the encoder, used to map the input data to a low-dimensional space; b e Indicates the encoder's bias term; f represents the activation function, using the non-linear activation function ReLU to increase the model's expressive power; Z represents the low-dimensional feature representation, which represents the features of the input data after compression. These features will be used in subsequent LSTM networks for time series modeling; 2.2 Decoder Section The role of the decoder is to reconstruct the low-dimensional feature representation Z into the original data that is as close as possible to the input X, so as to ensure that the extracted features can retain the key information in the sensor data. The decoder's output formula is: in W represents the reconstructed output data, which is expected to be as close as possible to the original input X; d The weight matrix of the decoder is used to map low-dimensional features back to high-dimensional space; b d represents the bias term of the decoder; g represents the activation function, which is the same activation function used in the encoder; Z represents the low-dimensional features obtained from the encoder; 2.3 Loss Function The goal of an autoencoder is to minimize the difference between the input data X and the reconstructed data. The difference between them is used to ensure the effectiveness of feature extraction; the loss function here is the mean squared error (MSE), and its formula is: Where L represents the loss function, indicating the difference between the input and the reconstructed output; n represents the dimension of the input data; X i This represents the i-th data point in the original input; This represents the i-th data point after reconstruction; By minimizing the loss function L, the autoencoder can learn key features in the sensor data and retain as much original information as possible. Step 3: Time Series Modeling Time series modeling of convolutional features is performed using an LSTM network; LSTM is suitable for processing time series data and capturing the state changes of agricultural tractors under different working conditions. The LSTM network learns the time dependence of sensor data step by step and can understand the changes of various working conditions of the tractor over time. The features extracted from the autoencoder are used as the input of the LSTM, and the working conditions are modeled through multiple layers of LSTM to predict the changes in working conditions in the future. 3.1 LSTM Cell Design LSTM networks are a special version of recurrent neural networks (RNNs) used to capture temporal dependencies, and are particularly well-suited for processing time series data. It learns long-term and short-term dependencies in sensor data by designing memory units, thereby understanding changes in the tractor's operating conditions; the core structure of the LSTM unit represents... The forget gate is used to determine the historical information that needs to be forgotten; the formula is: f t =σ(W f ·[h t-1 ,x t ]+b f ) Where f t The output of the forget gate, with a value between 0 and 1, indicates how much past information is retained; W f The weight matrix of the forget gate; h t-1 Indicates the hidden state at the previous time step; x t b represents the feature vector input at the current time step; f σ represents the bias term of the forget gate; σ represents the activation function Sigmoid, which compresses the result to between 0 and 1. The input gate is used to determine the current information that needs to be updated; the formula is as follows: i t =σ(W i ·[h t-1 ,x t ]+b i ) Where i t The output of the input gate controls the proportion of the current input that is retained; W i b represents the weight matrix of the input gate; i Indicates the bias term of the input gate; x t h represents the feature vector input at the current time step. t-1 This indicates the hidden state of the previous time step; Candidate memory states are used to generate the current candidate memory content; The formula is to represent in W represents the candidate memory state at the current time step. C The weight matrix representing the candidate memory states; b C The bias term represents the candidate memory state; tanh represents the hyperbolic tangent function, which compresses the result to between -1 and 1; h t-1 Indicates the hidden state at the previous time step; x t This represents the feature vector input at the current time step; Memory update is used to update the memory state through the forget gate and the input gate; the formula is expressed as follows: Where C represents t Indicates the memory state at the current time step; C t-1 Indicates the memory state of the previous time step; i t This indicates the output of the input gate, controlling the proportion of the current input that is retained; Indicates the candidate memory state at the current time step; f t Indicates the output of the forget gate; The output gate is used to determine the hidden state of the output; the formula is as follows: the t =σ(W o ·[h t-1 ,x t ]+b o ) The formula for calculating the hidden state is as follows: h t =o t ·tanh(C t ) Among them o t The output of the output gate is represented by σ; the sigmoid activation function is represented by W. o The weight matrix [h] represents the output gate. t-1 x t ] represents the hidden state h of the previous time step. t-1 With the input x at the current time step t The concatenated vector; b o The output gate's bias term is represented by tanh; tanh represents the hyperbolic tangent function; C t This indicates the memory state at the current time step; 3.2 Input and Output Sequences Input sequence: The feature map extracted from the autoencoder is used as the input to the LSTM network. The LSTM network models the time series data to understand the changes in the working conditions of the tractor. Output sequence: The output of the LSTM network will be used to predict the operating condition of the tractor over a future period; the state output by the LSTM network is h. t LSTM can capture the impact of past data on the future, helping to predict the next state of the tractor: prediction of normal operation, standby or low-load operation; Time series modeling is performed using a multi-layer LSTM, and the final prediction results are output: y t =LSTM(x1,x2,…,x t ) Where y t This is the output of the current time step, representing the current operating state of the tractor; x1, x2, ..., x t It is a time series of sensor input features; 3.3 Algorithm Improvement Traditional LSTM only considers information from the past to the present, but in some cases, future state information can also help with current predictions. By introducing bidirectional LSTM, the model can consider both past and future time step information, thus more accurately modeling the working conditions of tractors. Bidirectional LSTM consists of two LSTM networks, one processing time series from front to back and the other from back to front; their outputs are the hidden states from both directions concatenated together to form the final feature representation. Hidden state of forward LSTM for: Hidden state of backward LSTM for: The final output of the bidirectional LSTM is: in: The hidden state of the forward LSTM captures past information at time step t; The hidden state of the backward LSTM captures future information at time step t; h t The output of a bidirectional LSTM contains combined features of the past and future; LSTM represents the Long Short-Term Memory network constructed above. Step 4: Fuel Consumption Data Optimization The above-mentioned working condition identification model automatically identifies tractor fuel consumption data under different working conditions, reducing the impact of fuel consumption from ineffective operations; combined with actual fuel consumption data under different working conditions, the fuel consumption management and operating efficiency of tractors are further optimized. Specifically, it is through the above formula y t =LSTM(x1,x2,…,x t ) The Long Short-Term Memory (LSTM) network model is used to predict the operating conditions at future times, and the final fuel consumption is calculated based on the prediction results; the LSTM output at each time step y t Each of these corresponds to a fuel consumption rate, and these fuel consumption rates are recorded. The formula for calculating fuel consumption is: Fuel consumption = Fuel consumption rate × Time interval; The total fuel consumption for the entire time series is obtained by summing up the fuel consumption at all time steps: Where T is the total number of time steps, and fuel consumption rate is... t It is the fuel consumption rate at the t-th time step; This implementation combines two algorithms, autoencoder and LSTM, which has advantages in processing multi-dimensional sensor data and time series data, and can efficiently realize the automated identification of the working conditions of agricultural tractors.

Images