An adaptive variable torque tobacco leaf harvesting device and method

Abstract

This invention discloses an adaptive variable torque tobacco leaf harvesting device and method. The device includes a rotating harvesting mechanism, a position adjustment mechanism, and a working parameter acquisition device. The rotating harvesting mechanism includes a rotating blade and a rotating drive mechanism. The position adjustment mechanism includes an adjustment mounting frame, a vertical adjustment mechanism, and a width adjustment mechanism. Two adjustment mounting frames are provided and arranged opposite to each other, with the two rotating harvesting mechanisms respectively mounted on the two adjustment mounting frames. Two sets of vertical adjustment mechanisms are provided, each connected to one of the two adjustment mounting frames. Two sets of width adjustment mechanisms are provided and connected to the two adjustment mounting frames respectively. The working parameter acquisition device includes a torque sensor, a height sensor, and a speed sensor. This invention can adjust the height, angle, and lateral distance of the rotating blade, and can also adaptively adjust the torque of the rotating blade, improving harvesting efficiency, reducing labor requirements, and reducing tobacco leaf loss.

Description

Technical Field

[0001] This invention relates to tobacco leaf harvesting devices and methods, specifically to an adaptive variable torque tobacco leaf harvesting device and method. Background Technology

[0002] The mechanization of tillage and ridging in tobacco production is relatively high, and mechanization of tobacco leaf production is an important vehicle for the modernization of tobacco leaf production. Among the various stages of tobacco production, ridging is the most widely mechanized operation. In contrast, the harvesting stage of tobacco production has a lower degree of mechanization. In northern tobacco-producing areas such as Heilongjiang, Shandong, and Henan, where planting areas are concentrated and the terrain is flat, mechanized operations are easier to carry out. In recent years, mechanized tobacco harvesting has developed rapidly, and mechanical equipment has been widely used, greatly reducing the number of workers and improving production efficiency. In southern tobacco-producing areas located in hilly and mountainous areas, where planting areas are more scattered and the terrain is rugged, mechanized harvesting is more difficult. However, all tobacco-producing areas are actively promoting mechanized operations and developing mechanized equipment suitable for local terrains, which is of great significance for promoting the modernization of tobacco agriculture. However, most tobacco leaf harvesting devices currently using rotary blades cannot adjust the height, angle, and lateral distance of the rotary blades, nor can they adaptively adjust the torque of the rotary blades, resulting in limited harvesting range and accuracy. Summary of the Invention

[0003] The purpose of this invention is to overcome the above-mentioned problems and provide an adaptive variable torque tobacco leaf picking device. This tobacco leaf picking device can adjust the height, angle, and lateral distance of the rotating blade, and can also adaptively adjust the torque of the rotating blade, thereby improving picking efficiency, reducing the number of workers, and reducing tobacco leaf loss.

[0004] Another objective of this invention is to provide an adaptive variable torque tobacco leaf harvesting method.

[0005] The objective of this invention is achieved through the following technical solution:

[0006] An adaptive variable torque tobacco leaf picking device includes a picking frame and a rotating picking mechanism and a position adjustment mechanism mounted on the picking frame.

[0007] The rotating harvesting mechanism is provided in two parts, each of which includes a rotating blade and a rotating drive mechanism for driving the rotating blade to rotate.

[0008] The position adjustment mechanism includes an adjustment mounting frame, a vertical adjustment mechanism, and a width adjustment mechanism. Two adjustment mounting frames are provided and arranged opposite each other, with two rotating picking mechanisms respectively mounted on the two adjustment mounting frames. Two sets of vertical adjustment mechanisms are provided, with the drive ends of each set connected to the two adjustment mounting frames. Each set of vertical adjustment mechanisms includes two vertical adjustment mechanisms. The arrangement direction of the two vertical adjustment mechanisms in the same set is parallel to the forward direction of the picking device during picking. Two sets of width adjustment mechanisms are provided, with the drive ends of each set connected to the two adjustment mounting frames. The arrangement direction of the two sets of width adjustment mechanisms is parallel to the arrangement direction of the two adjustment mounting frames.

[0009] It also includes an operation parameter acquisition unit, which includes a torque sensor for detecting the rotational resistance torque of the rotating blade, a height sensor for detecting the height position of the rotating blade, and a speed sensor for detecting the operation speed of the harvesting device.

[0010] The working principle of the above-mentioned adaptive variable torque tobacco leaf picking device is as follows:

[0011] During operation, the main stalk of the tobacco plant extends into the gap between two rotating blades. The rotating blades are driven to rotate by a rotation drive mechanism, thereby knocking the tobacco leaves off the main stalk of the tobacco plant.

[0012] Based on the diameter of the tobacco plant's main stalk, the distance between the two adjustment mounting frames is adjusted by two sets of width adjustment mechanisms, allowing the tobacco plant's main stalk to smoothly extend into the two adjustment mounting frames so that the rotating blades can be used for harvesting.

[0013] Based on the height of the tobacco leaves to be harvested, the height of the two adjustment mounting frames is adjusted by two sets of vertical adjustment mechanisms so that the height of the rotating blades of the rotating harvesting mechanism corresponds to the height of the tobacco leaves, so that different parts of the tobacco leaves (upper leaves, middle leaves, and lower leaves) can be harvested flexibly.

[0014] Furthermore, when it is necessary to adjust the tilt of the rotating blade, different heights can be adjusted through two vertical adjustment mechanisms in the same set, so that the same adjustment mounting frame can be tilted at the desired angle to cope with different harvesting situations.

[0015] In a preferred embodiment of the present invention, the rotating blade has a structure symmetrical about the rotation center, and the cutting edge of the rotating blade has a wavy structure.

[0016] Furthermore, when the two rotating blades rotate to the same plane, the opposing cutting edges of the two rotating blades are misaligned and engaged.

[0017] In a preferred embodiment of the invention, the rotating blade is made of a resin material.

[0018] In a preferred embodiment of the present invention, the rotary drive mechanism includes a rotary drive motor, which is fixed on an adjustment mounting bracket, and the output shaft of the rotary drive motor is connected to the rotary blade via a coupling and a rotary mounting plate.

[0019] In a preferred embodiment of the present invention, the vertical adjustment mechanism includes a vertical drive component and a vertical transmission assembly, the vertical transmission assembly including a vertical transmission screw and a vertical transmission screw nut, the vertical transmission screw being connected to the adjustment mounting bracket.

[0020] Furthermore, the vertical drive component is a manual turntable;

[0021] The vertical transmission assembly also includes a bevel gear set; the manual turntable is connected to the vertical transmission lead screw via the bevel gear set.

[0022] Furthermore, a synchronous drive shaft is provided between the opposing bevel gear sets of the two sets of vertical adjustment mechanisms; the manual turntable is fixedly connected to one end of the synchronous drive shaft. In this way, the same end of the two adjustment mounting brackets can be driven to move vertically synchronously, making operation convenient.

[0023] Furthermore, each set of width adjustment mechanisms includes two width adjustment mechanisms, with the two width adjustment mechanisms in the same set respectively set on the two vertical adjustment mechanisms in the same set.

[0024] Furthermore, the width adjustment mechanism is fixedly connected to the vertical transmission screw nut, and the width adjustment mechanism is composed of a manual fine-tuning mechanism, which is fixedly connected to the adjustment mounting bracket.

[0025] In a preferred embodiment of the present invention, a walking mechanism is further included, which is disposed below the harvesting machine frame.

[0026] In a preferred embodiment of the present invention, a working parameter acquisition unit is further included. This unit comprises a torque sensor for detecting the rotational resistance torque of the rotating blade, a height sensor for detecting the height position of the rotating blade, and a speed sensor for detecting the operating speed of the harvesting device. Thus, during operation, by acquiring the resistance torque, the height of the rotating blade, and the operating speed of the harvesting device, and converting these into electrical signals, the controller is transmitted to the controller. The controller then uses a built-in algorithm and torque prediction model to adaptively control the rotating drive motor, ensuring that the output torque of the rotating drive motor is at its optimal state. The rotational speed of the rotating blade can be flexibly adjusted according to actual needs, which helps reduce energy consumption.

[0027] An adaptive variable torque tobacco leaf harvesting method includes a pre-operation adjustment step and a harvesting operation step;

[0028] The pre-operation adjustment steps include:

[0029] When it is necessary to adjust the distance between the two adjustment mounting brackets, the distance between the two adjustment mounting brackets is adjusted by two sets of width adjustment mechanisms according to the diameter of the tobacco plant main stem, so that the tobacco plant main stem can be smoothly inserted into the two adjustment mounting brackets;

[0030] When the height of the rotating blade is required, the height of the two adjusting mounting brackets is adjusted by two sets of vertical adjusting mechanisms according to the height of the tobacco leaves to be picked, so that the height of the rotating blade of the rotating picking mechanism corresponds to the height of the tobacco leaves, so as to pick tobacco leaves from different parts.

[0031] When it is necessary to adjust the tilt of the rotating blade, the height of both ends of the adjustment mounting frame is adjusted by the two vertical adjustment mechanisms in the same group, so that the same adjustment mounting frame can be tilted at the desired angle to cope with different harvesting situations.

[0032] The harvesting process includes:

[0033] Place the harvesting device in the tobacco field; move the harvesting device along the direction of tobacco planting;

[0034] During the movement, the main stem of the tobacco plant extends into the gap between the two rotating blades. The rotating blades are driven to rotate by the rotation drive mechanism, which knocks the tobacco leaves off the main stem of the tobacco plant.

[0035] During the harvesting operation, data on the rotational resistance torque of the rotating blades is collected by a torque sensor; data on the height position of the rotating blades is collected by a height sensor; and data on the operating speed of the harvesting device is collected by a speed sensor.

[0036] The collected data is converted into electrical signals and transmitted to the controller. The controller then uses a built-in algorithm and torque prediction model to adaptively control the rotary drive motor, ensuring that the output torque of the rotary drive motor is at its optimal level. The speed of the rotary tool can be flexibly adjusted according to actual needs, which helps to reduce energy consumption.

[0037] Furthermore, based on the artificial neural network model, an output torque prediction model is established, and the torque prediction model is optimized using a genetic algorithm to finally obtain an ideal motor output torque prediction model, thereby achieving adaptive variable torque.

[0038] Furthermore, the operation of optimizing the torque prediction model using a genetic algorithm is as follows:

[0039] Parameter encoding;

[0040] The fitness function is determined based on the objective function. The expression for the fitness function is as follows:

[0041]

[0042] In the formula, n is the number of neurons in the output layer of the neural network; Let be the expected output of the i-th neuron; This represents the predicted output of the i-th neuron in the neural network.

[0043] Genetic output.

[0044] Furthermore, from R 2 The established torque prediction model is evaluated based on three aspects: RMSE, MAE, and R. The smaller the RMSE and MAE, the better the torque prediction performance. 2 The larger the value of R, the better the performance of the output torque prediction model. 2 The expressions for RMSE and MAE are as follows:

[0045]

[0046]

[0047]

[0048] In the formula, y i This represents the true value of the i-th neuron in the neural network. This represents the predicted output of the i-th neuron in the neural network. This represents the mean of the neural network samples.

[0049] Furthermore, a multi-layer feedforward neural network is used to train and predict the torque prediction model. The specific operation is as follows:

[0050] Set the number of neurons in the input layer, hidden layer, and output layer; initialize the weights, thresholds, and learning rates of connected neurons; determine the range of the number of neurons in the hidden layer using the following formula:

[0051]

[0052] In the formula, N h n is the number of neurons in the hidden layer; n0 is the number of neurons in the input layer; n0 is the number of neurons in the output layer; α is a constant between 1 and 10.

[0053] Calculate the hidden layer output:

[0054] In the formula, H j is the output of the j-th node in the hidden layer; n is the number of neurons in the input layer; f() is the activation function of the hidden layer;

[0055] Calculate the output of the output layer:

[0056] In the formula, H j is the output of the j-th node in the hidden layer; k is the number of neurons in the hidden layer; It is the output layer activation function;

[0057] Calculate the network prediction error: e k =Y k -O k ;

[0058] Based on the network prediction error e k Update neuron connection weights w ij With w jk :

[0059]

[0060] Based on the network prediction error e k Update neuron connection threshold a j With b k :

[0061]

[0062] b k =b k +e k ;

[0063] In the formula, η is the learning rate.

[0064] Furthermore, to determine whether the algorithm iteration has terminated, one of the following two conditions must be met: the network output reaches the set accuracy requirement and the total number of training iterations reaches the maximum number of iterations;

[0065] If the condition is met, training is terminated;

[0066] If the condition is not met, return to the hidden layer output for calculation.

[0067] Compared with the prior art, the present invention has the following advantages:

[0068] 1. By setting a width adjustment mechanism, the distance between the two adjustment mounting frames can be adjusted according to the diameter of the tobacco plant main stalk, so that the tobacco plant main stalk can be smoothly inserted into the two adjustment mounting frames for harvesting by rotating the blade.

[0069] 2. By setting up a vertical adjustment mechanism, the height of the two adjustment mounting frames can be adjusted according to the height of the tobacco leaves to be picked, so that the height of the rotating blade of the rotating picking mechanism corresponds to the height of the tobacco leaves, so that different parts of the tobacco leaves (upper leaves, middle leaves, and lower leaves) can be picked flexibly.

[0070] 3. When it is necessary to adjust the tilt of the rotating blade, different heights can be adjusted through two vertical adjustment mechanisms in the same group, so that the same adjustment mounting frame can be tilted at the desired angle to cope with different harvesting conditions, improve harvesting efficiency, reduce the number of workers and reduce tobacco leaf loss. Attached Figure Description

[0071] Figure 1 This is a three-dimensional structural diagram of the adaptive variable torque tobacco leaf harvesting device of the present invention.

[0072] Figure 2 This is a front view of the adaptive variable torque tobacco leaf harvesting device of the present invention.

[0073] Figure 3 This is a diagram of the artificial neural network structure of the present invention. Detailed Implementation

[0074] To enable those skilled in the art to fully understand the technical solutions of the present invention, the present invention will be further described below in conjunction with embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0075] See Figure 1-2 The adaptive variable torque tobacco leaf picking device of this embodiment includes a picking frame 1, a walking mechanism, a rotating picking mechanism, and a position adjustment mechanism; wherein, the walking mechanism is located at the bottom of the picking frame 1 and can directly adopt the existing structure.

[0076] See Figure 1-2 The rotating harvesting mechanism is provided in two parts, each of which includes a rotating blade 2 and a rotating drive mechanism for driving the rotating blade 2 to rotate; wherein, the rotating blade 2 has a structure symmetrical about the rotation center, and the blade of the rotating blade 2 has a wavy structure; the rotating blade 2 is made of resin material.

[0077] Furthermore, when the two rotating blades 2 rotate to the same plane, the opposing cutting edges of the two rotating blades 2 are misaligned and engaged.

[0078] See Figure 1-2 The rotary drive mechanism includes a rotary drive motor 3, which is fixed on the adjustment mounting bracket 4. The output shaft of the rotary drive motor 3 is connected to the rotary blade 2 through a coupling and a rotary mounting plate.

[0079] See Figure 1-2The position adjustment mechanism includes an adjustment mounting frame 4, a vertical adjustment mechanism, and a width adjustment mechanism. Two adjustment mounting frames 4 are provided and arranged opposite each other, with two rotating picking mechanisms respectively mounted on the two adjustment mounting frames 4. Two sets of vertical adjustment mechanisms are provided, with the drive ends of the two sets of vertical adjustment mechanisms respectively connected to the two adjustment mounting frames 4. The vertical surface of each adjustment mounting frame 4 has an elongated arc-shaped hole, and a rectangular area is cut off from the horizontal surface and has bearing fixing holes. Each set of vertical adjustment mechanisms includes two vertical adjustment mechanisms. The arrangement direction of the two vertical adjustment mechanisms in the same set is parallel to the forward direction of the picking device during picking. The four vertical adjustment mechanisms are arranged in a rectangular pattern on the picking machine frame 1.

[0080] See Figure 1-2 The vertical adjustment mechanism includes a vertical drive component and a vertical transmission assembly. The vertical drive component is a manual turntable 5. The vertical transmission assembly includes a vertical transmission screw 6 and a vertical transmission screw nut 7. The vertical transmission screw 6 and the vertical transmission screw nut 7 are connected to the adjustment mounting bracket 4. The vertical transmission assembly also includes a bevel gear set. The manual turntable 5 is connected to the vertical transmission screw 6 through the bevel gear set.

[0081] See Figure 1-2 A synchronous drive shaft 8 is provided between the opposing bevel gear sets of the two sets of vertical adjustment mechanisms; the manual turntable 5 is fixedly connected to one end of the synchronous drive shaft 8. In this way, the same end of the two adjustment mounting brackets 4 can be driven to move vertically synchronously, making operation convenient.

[0082] See Figure 1-2 The width adjustment mechanism is provided in two groups, each group including two width adjustment mechanisms; the arrangement direction of the two groups of width adjustment mechanisms is parallel to the arrangement direction of the two adjustment mounting brackets 4; the two width adjustment mechanisms in the same group are respectively set on the two vertical adjustment mechanisms in the same group.

[0083] Furthermore, the width adjustment mechanism is fixedly connected to the vertical transmission screw nut 7. The width adjustment mechanism is composed of a manual fine-tuning mechanism 9, which is fixedly connected to the adjustment mounting bracket 4.

[0084] Specifically, this embodiment also includes an operation parameter acquisition device (not shown in the figure), which includes a torque sensor for detecting the rotational resistance torque of the rotating blade 2, a height sensor for detecting the height position of the rotating blade 2, and a speed sensor for detecting the operation speed of the harvesting device.

[0085] See Figure 1-2 The adaptive variable torque tobacco leaf picking method of this embodiment includes a pre-operation adjustment step and a picking operation step.

[0086] The pre-operation adjustment steps include:

[0087] When it is necessary to adjust the distance between the two adjustment mounting brackets 4, the distance between the two adjustment mounting brackets 4 is adjusted by two sets of width adjustment mechanisms according to the diameter of the tobacco plant main stem, so that the tobacco plant main stem can be smoothly inserted into the two adjustment mounting brackets 4.

[0088] When the height of the rotating blade 2 is required, the height of the two adjusting mounting brackets 4 is adjusted by two sets of vertical adjusting mechanisms according to the height of the tobacco leaves to be picked, so that the height of the rotating blade 2 of the rotating picking mechanism corresponds to the height of the tobacco leaves, so as to pick tobacco leaves from different parts.

[0089] When it is necessary to adjust the tilt of the rotating blade 2, the height of both ends of the adjustment mounting frame 4 is adjusted by the two vertical adjustment mechanisms in the same group, so that the same adjustment mounting frame 4 can be tilted at the desired angle to cope with different harvesting situations.

[0090] The harvesting process includes:

[0091] The harvesting device is placed in the tobacco field; the harvesting device is moved along the direction of tobacco planting.

[0092] During the movement, the main stem of the tobacco plant extends into the gap between the two rotating blades 2, and the rotating blades 2 are driven to rotate by the rotation drive mechanism, which knocks the tobacco leaves off the main stem of the tobacco plant.

[0093] During the harvesting operation, the torque sensor collects data on the rotational resistance torque of the rotating blade 2; the height sensor collects data on the height position of the rotating blade 2; and the speed sensor collects data on the operating speed of the harvesting device. The collected data are converted into electrical signals and transmitted to the controller. The controller uses a built-in algorithm and torque prediction model to adaptively control the rotating drive motor 3, so that the output torque of the rotating drive motor 3 is in the optimal state. The rotational speed of the rotating blade 2 can be flexibly adjusted according to actual needs, which helps to reduce energy consumption.

[0094] Furthermore, the torque sensor is a photoelectric torque sensor, which can be obtained through the following formula:

[0095]

[0096] In the formula, T is the torque (N·m); G is the shear modulus (GPa); I p It is the polar moment of inertia. d is the diameter of the elastic shaft (mm); L is the relative rotation angle between the two ends of the elastic shaft (rad); L is the length of the elastic shaft between the two gratings (mm).

[0097] The speed sensor is a photoelectric speed sensor, and the speed calculation formula is as follows:

[0098]

[0099] In the formula, R is the diameter of the wheelset (m), N is the number of pulses, Z is the number of through holes / gratings on the rotating disk, and t is the measurement time (s).

[0100] The controller in this embodiment can be a processor that can perform data processing based on programming, such as an STM32F1* series MCU processor, and is powered by a 24V DC power supply.

[0101] Furthermore, the motor output torque required for harvesting the upper, middle, and lower leaves of the tobacco plant is different. Therefore, the height models of the upper, middle, and lower leaves of the tobacco plant need to be preset in the controller. When the height sensor collects the height data, it is compared with the height model in the controller to determine the position of the rotating blade, so that the motor outputs the appropriate torque.

[0102] Furthermore, the torque fed back by the rotating blade, the working height, and the working speed of the harvesting device are all dynamically changing, requiring an accurate output torque prediction model to achieve adaptive torque variation. Based on an artificial neural network (ANN) model, a motor output torque prediction model is established, and a genetic algorithm is used to optimize the torque prediction model, ultimately obtaining an ideal motor output torque prediction model to achieve adaptive torque variation.

[0103] Furthermore, to fully consider the practicality of the prediction model and ensure that the prediction model has sufficient training samples, a motor output torque prediction model was constructed, taking the torque fed back by the rotating blade, the working height, and the working speed of the picking device as influencing factors, and the output torque of the motor as the evaluation index. 100 prediction model sample data obtained from the test were randomly selected, and the dataset was randomly divided into training dataset and test dataset in an 8:2 ratio to ensure the uniformity of the sample data.

[0104] Furthermore, when building a machine learning model, the setting of some parameters is random, such as the weights and thresholds of the artificial neural network. These parameter settings determine the training speed and prediction accuracy of the model. To improve the training speed and accuracy of motor output torque, it is necessary to optimize the randomly set parameters to obtain ideal model parameters. A genetic algorithm is used to optimize the prediction model, improving its computational efficiency and prediction accuracy. The specific process is as follows:

[0105] Parameter encoding.

[0106] The fitness function is determined based on the objective function, and its specific expression is as follows:

[0107]

[0108] In the formula, n is the number of neurons in the output layer of the neural network; Let be the expected output of the i-th neuron; This represents the predicted output of the i-th neuron in the neural network.

[0109] Genetic output.

[0110] Furthermore, from R respectively 2 The established output torque model is evaluated based on three aspects: coefficient of determination (RMSE), root mean square error (RMSE), and mean absolute error (MAE). The smaller the RMSE and MAE, the better the output torque model. 2 The larger the value of R, the better the performance of the output torque prediction model. 2 The formulas for expressing RMSE and MAE are as follows:

[0111]

[0112]

[0113]

[0114] In the formula, y i This represents the true value of the i-th neuron in the neural network. This represents the predicted output of the i-th neuron in the neural network. This represents the mean of the neural network samples.

[0115] Furthermore, a multi-layer feedforward neural network is used for training and predicting the output torque. The basic elements of an artificial neural network (ANN) consist of neurons, which have input and output functions for transmitting and processing information. See also Figure 3 The neurons are distributed across three layers of the neural network: the input layer, the hidden layer, and the output layer. The torque fed back by the rotating blade, the working height, and the working speed of the harvesting device serve as the input variables of the neural network, while the motor's output torque is the output variable. The input layer acts as the network's entry point, transmitting input data throughout the network. The output layer acts as the network's exit point, outputting the network model's predictions. The hidden layer acts as an intermediate layer, establishing connections with the input and output layers by adjusting the weights, thresholds, and activation functions of connected neurons.

[0116] Furthermore, the training process for the artificial neural network model is as follows:

[0117] Set the number of neurons in the input, hidden, and output layers; initialize the weights, thresholds, and learning rates of connected neurons. Determine the range of the number of neurons in the hidden layer using the following formula:

[0118]

[0119] In the formula, N h n is the number of neurons in the hidden layer; n is the number of neurons in the input layer; n0 is the number of neurons in the output layer; and a is a constant between 1 and 10.

[0120] Calculate the hidden layer output:

[0121] In the formula, H j is the output of the j-th node in the hidden layer; n is the number of neurons in the input layer; f() is the activation function of the hidden layer.

[0122] Calculate the output of the output layer:

[0123] In the formula, H j is the output of the j-th node in the hidden layer; k is the number of neurons in the hidden layer; It is the activation function of the output layer.

[0124] Calculate the network prediction error: e k =Y k -O k .

[0125] Based on the network prediction error e k Update neuron connection weights w ij With w jk :

[0126]

[0127] Based on the network prediction error e k Update neuron connection threshold a j With b k :

[0128]

[0129] b k =b k +e k ;

[0130] In the formula, η is the learning rate.

[0131] To determine whether the algorithm iteration should terminate, one of the following two conditions must be met: ① The network output reaches the set accuracy requirement; ② The total number of training iterations reaches the maximum number of iterations. If both conditions are met, training terminates; otherwise, the algorithm returns to calculating the hidden layer output.

[0132] The above are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above content. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. An adaptive variable torque tobacco leaf harvesting device, characterized in that, This includes the harvesting machine frame and the rotating harvesting mechanism and position adjustment mechanism mounted on the harvesting machine frame; The rotating harvesting mechanism is provided in two parts, each of which includes a rotating blade and a rotating drive mechanism for driving the rotating blade to rotate. The position adjustment mechanism includes an adjustment mounting frame, a vertical adjustment mechanism, and a width adjustment mechanism. Two adjustment mounting frames are provided and arranged opposite each other, with two rotating picking mechanisms respectively mounted on the two adjustment mounting frames. Two sets of vertical adjustment mechanisms are provided, with the drive ends of each set connected to the two adjustment mounting frames. Each set of vertical adjustment mechanisms includes two vertical adjustment mechanisms. The arrangement direction of the two vertical adjustment mechanisms in the same set is parallel to the forward direction of the picking device during picking. Two sets of width adjustment mechanisms are provided, with the drive ends of each set connected to the two adjustment mounting frames. The arrangement direction of the two sets of width adjustment mechanisms is parallel to the arrangement direction of the two adjustment mounting frames. It also includes an operation parameter acquisition device, which includes a torque sensor for detecting the rotational resistance torque of the rotating blade, a height sensor for detecting the height position of the rotating blade, and a speed sensor for detecting the operation speed of the harvesting device. The adaptive variable torque tobacco picking method of the adaptive variable torque tobacco picking device includes a pre-operation adjustment step and a picking operation step. The pre-operation adjustment steps include: When it is necessary to adjust the distance between the two adjustment mounting brackets, the distance between the two adjustment mounting brackets is adjusted by two sets of width adjustment mechanisms according to the diameter of the tobacco plant main stem, so that the tobacco plant main stem can be smoothly inserted into the two adjustment mounting brackets; When the height of the rotating blade is required, the height of the two adjusting mounting brackets is adjusted by two sets of vertical adjusting mechanisms according to the height of the tobacco leaves to be picked, so that the height of the rotating blade of the rotating picking mechanism corresponds to the height of the tobacco leaves, so as to pick tobacco leaves from different parts. When it is necessary to adjust the tilt of the rotating blade, the height of both ends of the adjustment mounting frame is adjusted by the two vertical adjustment mechanisms in the same group, so that the same adjustment mounting frame can be tilted at the desired angle to cope with different harvesting situations. The harvesting process includes: Place the harvesting device in the tobacco field; move the harvesting device along the direction of tobacco planting; During the movement, the main stem of the tobacco plant extends into the gap between the two rotating blades. The rotating blades are driven to rotate by the rotation drive mechanism, which knocks the tobacco leaves off the main stem of the tobacco plant. During the harvesting operation, data on the rotational resistance torque of the rotating blades is collected by a torque sensor; data on the height position of the rotating blades is collected by a height sensor; and data on the operating speed of the harvesting device is collected by a speed sensor. The collected data is converted into electrical signals and transmitted to the controller. The controller then uses a built-in algorithm and torque prediction model to adaptively control the rotary drive motor, ensuring that the output torque of the rotary drive motor is at its optimal state. The speed of the rotary tool can be flexibly adjusted according to actual needs to reduce energy consumption. Based on an artificial neural network model, an output torque prediction model is established, and a genetic algorithm is used to optimize the torque prediction model, thereby improving the computational efficiency and prediction accuracy of the prediction model. The operation of optimizing the torque prediction model using a genetic algorithm is as follows: Parameter encoding; The fitness function is determined based on the objective function. The expression for the fitness function is as follows: In the formula, n is the number of neurons in the output layer of the neural network; Let be the expected output of the i-th neuron; This represents the predicted output of the i-th neuron in the neural network. Genetic output; From R 2 The established torque prediction model is evaluated based on three aspects: RMSE, MAE, and R. The smaller the RMSE and MAE, the better the torque prediction performance. 2 The larger the value of R, the better the performance of the output torque prediction model. 2 The expressions for RMSE and MAE are as follows: In the formula, y i This represents the true value of the i-th neuron in the neural network. This represents the predicted output of the i-th neuron in the neural network. The mean of the neural network samples; A multi-layer feedforward neural network is used to train and predict the torque prediction model. The specific operation is as follows: Set the number of neurons in the input layer, hidden layer, and output layer; initialize the weights, thresholds, and learning rates of connected neurons; determine the range of the number of neurons in the hidden layer using the following formula: In the formula, N h n is the number of neurons in the hidden layer; n is the number of neurons in the input layer; n0 is the number of neurons in the output layer; a is a constant between 1 and 10. Calculate the hidden layer output: In the formula, H j is the output of the j-th node in the hidden layer; n is the number of neurons in the input layer; f() is the activation function of the hidden layer; Calculate the output of the output layer: In the formula, H j is the output of the j-th node in the hidden layer; k is the number of neurons in the hidden layer; It is the output layer activation function; Calculate the network prediction error: e k =Y k -O k ; Based on the network prediction error e k Update neuron connection weights w ij With w jk : Based on the network prediction error e k Update neuron connection threshold a j With b k : b k =b k +e k ; In the formula, h is the learning rate.

2. The adaptive variable torque tobacco leaf harvesting device according to claim 1, characterized in that, The rotating blade has a structure symmetrical about the rotation center, and the cutting edge of the rotating blade has a wavy structure.

3. The adaptive variable torque tobacco leaf harvesting device according to claim 1, characterized in that, When the two rotating blades rotate to the same plane, the opposite cutting edges of the two rotating blades are misaligned and engaged.

4. The adaptive variable torque tobacco leaf harvesting device according to claim 1, characterized in that, The rotary drive mechanism includes a rotary drive motor, which is fixed on an adjustment mounting bracket. The output shaft of the rotary drive motor is connected to the rotary blade via a coupling and a rotary mounting plate.

5. The adaptive variable torque tobacco leaf harvesting device according to claim 1, characterized in that, The vertical adjustment mechanism includes a vertical drive component and a vertical transmission assembly. The vertical transmission assembly includes a vertical transmission screw and a vertical transmission screw nut. The vertical transmission screw and the vertical transmission screw nut are connected to the adjustment mounting bracket.

6. The adaptive variable torque tobacco leaf harvesting device according to claim 5, characterized in that, The vertical drive component is a manual turntable; The vertical transmission assembly also includes a bevel gear set; the manual turntable is connected to the vertical transmission lead screw via the bevel gear set.

7. The adaptive variable torque tobacco leaf harvesting device according to claim 6, characterized in that, A synchronous drive shaft is provided between the opposing bevel gear sets of the two sets of vertical adjustment mechanisms; the manual turntable is fixedly connected to one end of the synchronous drive shaft.

8. The adaptive variable torque tobacco leaf harvesting device according to claim 5, characterized in that, Each set of width adjustment mechanisms includes two width adjustment mechanisms, and the two width adjustment mechanisms in the same set are respectively set on the two vertical adjustment mechanisms in the same set; The width adjustment mechanism is fixedly connected to the vertical transmission screw nut. The width adjustment mechanism consists of a manual fine-tuning mechanism, which is fixedly connected to the adjustment mounting bracket.

Images