CN113625560A - Loss rate control method and device for corn harvester, storage medium and equipment - Google Patents

Abstract

A method, device, storage medium and equipment for controlling the loss rate of a corn harvester. The method for controlling the loss rate of a corn harvester includes the following steps: constructing a dual-input and single-output fuzzy neural network controller, wherein the fuzzy neural network controller includes the following steps: It uses genetic algorithm-particle swarm algorithm for offline learning, and determines the weights and the center and width of the membership function that need to be learned in the fuzzy neural network controller by learning the previous operating data of the system; using The BP algorithm conducts online learning, establishes the connection weight of the controller, detects the rotation speed of the corn harvester grain recovery device in real time, and adjusts the adjustable parameters in the controller in real time in combination with the online learning, so that the controller can adapt to the machinery of the corn harvester grain recovery device. Performance changes and track corn yield setpoints. The invention also provides a corresponding method, device, storage medium and equipment for controlling the loss rate of a corn harvester.

Description

Method, device, storage medium and equipment for controlling loss rate of corn harvester

technical field

The invention relates to a corn harvester grain loss control technology, in particular to a corn harvester loss rate control method, device, storage medium and terminal computer equipment based on a fuzzy neural network algorithm.

Background technique

The corn harvester system is a complex control object with nonlinear, large time delay, strong coupling and time-varying characteristics, and is disturbed by many uncertain factors, such as changes in natural environment, mechanical wear, motor speed or human factors, etc. , these factors will cause the loss of corn harvest. The motor speed can be optimized by algorithm, thereby reducing the loss of corn harvest.

Intelligent control has the ability of self-learning and self-adaptation, has good control effect on linear and nonlinear systems, and can well solve the control of complex systems such as corn harvesters. Among them, neural network and fuzzy control are two important branches of intelligent control. Among them, the neural network is an operation model that imitates the structure and function of the biological neural network. It is composed of a large number of neurons and is a nonlinear dynamic system. Neural network has nonlinear approximation ability, learning ability, adaptive ability and fault tolerance ability. However, neural networks are not suitable for expressing rule-based knowledge. Fuzzy control simulates human thinking and processes knowledge with fuzzy logic and reasoning. It is based on language-based control rules, and is very suitable for control objects whose dynamic characteristics are difficult to grasp or change significantly. However, due to the increase of ambiguity, part of the information will be lost, and it is difficult to learn and establish perfect control rules, and it lacks self-adaptive ability. The fuzzy neural network is widely used because of its simple principle, strong applicability and strong robustness. However, the performance of fuzzy neural network-based control is poor in controlling complex processes with nonlinear, time-varying, coupling, and uncertain parameters and mechanisms.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is aimed at the above-mentioned problems in the prior art, and provides a method, device, storage medium and terminal equipment for controlling the loss rate of a corn harvester based on a fuzzy neural network algorithm. Control to reduce corn harvest losses.

In order to achieve the above purpose, the present invention provides a method for controlling the loss rate of a corn harvester, which comprises the following steps:

S100, constructing a dual-input and single-output fuzzy neural network controller, where the fuzzy neural network controller includes an antecedent network and a consequent network;

S200, adopting the genetic algorithm-particle swarm algorithm for offline learning, and preliminarily determining the weights and the center and width of the membership function to be learned in the fuzzy neural network controller by learning the previous operating data of the system; and

S300, using the BP algorithm to perform online learning, establishing the connection weights of the fuzzy neural network controller, detecting the rotation speed of the grain recovery device of the corn harvester in real time, and adjusting the adjustable parameters in the fuzzy neural network controller in real time in combination with the online learning , so that the fuzzy neural network controller adapts to the mechanical performance changes of the corn harvester grain recovery device, and tracks the corn harvest rate set value.

The above-mentioned method for controlling the loss rate of a corn harvester, wherein the antecedent network is a four-layer network structure, comprising:

In the input layer, the active power deviation and the active power deviation change rate of the input corn harvester are taken as y ₁ and y ₂ respectively, and the number of nodes in the input layer is N ₁ =2, and the expression is as follows:

为性能参数变量跟踪误差变化率，c为机械性能变化值，y为机械性能实际检测值。where e is the tracking error,

is the change rate of the tracking error of the performance parameter variable, c is the change value of the mechanical property, and y is the actual detection value of the mechanical property.

Fuzzy layer, the input variables y ₁ , y ₂ are divided into 7 fuzzy subsets {NB, NM, NS, O, PS, PM, PB}, as the nodes of the fuzzy layer, each node represents a The values of linguistic variables; their membership functions are all Gaussian functions, and the membership functions of each linguistic variable are:

为第一层的输入在相应语言变量值的模糊论域的隶属度函数，y_i为输入玉米收获机的有功功率偏差，c_ij和σ_ij(i＝1，2，…，n，j＝1，2，…m_i)分别为隶属度函数的中心和宽度，n为输入变量个数，m_i为输入变量x_i的模糊分割数，n＝2，m₁＝m₂＝7；所述模糊化层的节点数N₂＝m₁+m₂＝14；in,

is the membership function of the input of the first layer in the fuzzy universe of the corresponding linguistic variable value, y _i is the active power deviation of the input corn harvester, c _ij and σ _ij (i=1, 2,...,n,j= _{1,2 , … _ _} The number of nodes in the fuzzification layer N ₂ =m ₁ +m ₂ =14;

The fuzzy rule calculation layer is used to match the antecedents of the fuzzy rules, and calculate the fitness of each fuzzy rule:

为隶属度值，

为下一节点隶属度值，j₁＝j₂＝1，2，…，7；m＝m₁×m₂＝7×7＝49，所述模糊规则计算层的节点数N₃＝49；Among them, α _m is the calculation result of the fuzzy inference operation,

is the membership value,

is the membership degree value of the next node, j ₁ =j ₂ =1,2,...,7; m=m ₁ ×m ₂ =7×7=49, the number of nodes in the fuzzy rule calculation layer N ₃ =49;

The fourth layer of normalization layer is used to realize the normalization operation;

为加权系数，α_i为加权系数累加，所述归一化层的节点数N₄＝N₃＝49。in,

is the weighting coefficient, α _i is the weighting coefficient accumulation, and the number of nodes in the normalization layer is N ₄ =N ₃ =49.

The above-mentioned method for controlling the loss rate of a corn harvester, wherein the latter network is a three-layer network structure, including:

The first layer is used to pass the input variable to the next layer, the first layer has 3 nodes in total, and the input value of the first node is x ₀ =1, which is used to provide the constant term in the consequent of the fuzzy rule; The second and third nodes respectively input x ₁ , x ₂ ;

The second layer is used to calculate the consequent of each rule. There are 49 nodes in total, and each node represents a rule:

为连接权值，k＝0，1，2，m＝1，2，3，…，49，x₁为对应节点的规则，x₂为第二节点对应规则；where y _m is the consequent network of the mth rule,

is the connection weight, k=0, 1, 2, m=1, 2, 3,..., 49, x ₁ is the rule corresponding to the node, and x ₂ is the corresponding rule of the second node;

The third layer, used to calculate the controller output y:

Among them, α _m is the weighting coefficient, that is, the normalized fitness of each fuzzy rule, y _m is the weighted sum of the consequent components of each rule, and the output of the antecedent network is the connection weight of the consequent network.

The above-mentioned method for controlling the loss rate of a corn harvester, wherein the offline learning in step S200 includes the following steps:

S201, initialize a population parameter, where the population parameter is the initial position of each particle;

S202. Calculate the particle fitness F: F=abs(y-c), where y is the predicted output, and c is the expected output;

S203, find the individual extreme value and the group extreme value, and find out the individual minimum fitness value and the global minimum fitness value of each particle;

S204, the following formulas are used to calculate the speed update and position update of the particle:

Among them, ω is the inertia weight; d=1, 2, ..., D; i=1, 2, ..., n; k is the current number of iterations; V _id is the speed of the particle, P _id is the expected output, and X _id is the actual Output; c ₁ and c ₂ are acceleration factors, which are non-negative constants; r ₁ and r ₂ are random numbers distributed in [0, 1];

S205, according to the formula in step S202, calculate the particle fitness after speed and position update;

S206, update the individual extreme value and the group extreme value according to the formula in step S204;

S207, the following formula is used to calculate the intersection of the current individual and the individual extreme value, if the fitness value decreases, accept:

x _ij = x _ij (1-b)+P _ij b, in the formula, x _ij is a randomly selected individual extreme value, P _ij is a randomly selected group extreme value, and b is a random number between [0, 1];

S208, the following formula is used to calculate the intersection of the current individual and the group extreme value, and if the fitness value decreases, accept:

x _ij = x _ij (1-b)+P _gj , where x _ij is a randomly selected individual extreme value, P _gj is a randomly selected group extreme value, and b is a random number between [0, 1];

S209, the following formula is used to calculate the variation of the current individual itself, and if the fitness value decreases, accept:

In the formula, x _ij is a randomly selected individual extreme value, x _max is the upper bound of x _ij ; x _min is the lower bound of x _ij ; f(g)=r ₂ (1-g/G _max ) ² ; r ₂ is A random number; g is the current iteration number; G _max is the maximum number of evolutions; r is a random number of [0, 1];

S210. End if the maximum evolutionary algebra is satisfied, otherwise return to step S204.

The above-mentioned control method for the loss rate of a corn harvester, wherein the adjustable parameters that need to be learned include the connection weights and the central value c _ij and the width σ _ij of the membership function

The above-mentioned control method for the loss rate of a corn harvester, wherein, the learning algorithm of the connection weight is:

The learning algorithm of the central value c _ij of the membership function is:

c _ij (τ+1)=c _ij (τ)+Δc _ij (τ+1)+υ(c _ij (τ)-c _ij (τ-1));

The learning algorithm of the width σ _ij is:

σ _ij (τ+1)=σ _ij (τ)+Δσ _ij (τ+1)+υ(σ _ij (τ)-σ _ij (τ-1));

E为误差代价函数，η为学习速率。In the formula, i=1, 2, j=1, 2, 3,..., 7, T represents the moment, T+1 represents the next moment, T-1 represents the previous moment, u is the momentum factor,

E is the error cost function, and η is the learning rate.

The above-mentioned control method for the loss rate of corn harvester, wherein, the error cost function E is:

In the formula, c is the expected output and y is the actual output.

In order to better achieve the above purpose, the present invention also provides a loss rate control device for a corn harvester, which includes a fuzzy neural network controller, the fuzzy neural network controller includes an antecedent network and a consequent network, and adopts a fuzzy neural network controller. The above method for controlling the loss rate of the corn harvester can reduce the loss rate of the corn harvest by optimally controlling the rotational speed of the corn harvester grain recovery device.

In order to better achieve the above object, the present invention also provides a storage medium, wherein the storage medium stores a computer program, and the computer program is configured to execute the above-mentioned method for controlling the loss rate of a corn harvester when running.

In order to better achieve the above purpose, the present invention also provides an electronic device, which includes:

processor; and

a memory for storing executable instructions for the processor;

Wherein, the processor is configured to execute the above-mentioned corn harvester loss rate control method by executing the executable instructions.

The technical effect of the present invention is:

Based on the fuzzy neural network algorithm, the invention integrates the learning and computing functions of the neural network into the fuzzy system, and embeds the human-like IF-Then rules of the fuzzy system into the neural network, so as to maintain the strong knowledge expression ability of the fuzzy control system and improve its It has self-adaptive ability and self-learning ability. Through the optimization of the neural network algorithm for controlling the motor speed of the grain recovery device of the corn planter, the loss rate of corn harvesting is reduced.

The present invention is described in detail below with reference to the accompanying drawings and specific embodiments, but is not intended to limit the present invention.

Description of drawings

1 is a schematic diagram of a method for controlling the loss rate of a corn harvester according to an embodiment of the present invention;

2 is a schematic structural diagram of a fuzzy neural network according to an embodiment of the present invention;

FIG. 3 is a flowchart of a GA-PSO offline learning algorithm according to an embodiment of the present invention.

Detailed ways

Below in conjunction with accompanying drawing, structure principle and working principle of the present invention are described in detail:

Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments, however, can be embodied in various forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided in order to give a thorough understanding of the embodiments of the present invention. However, those skilled in the art will appreciate that the technical solutions of the present invention may be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be employed. In other instances, well-known solutions have not been shown or described in detail to avoid obscuring aspects of the present invention.

Furthermore, the drawings are merely schematic illustrations of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus their repeated descriptions will be omitted. Some of the block diagrams shown in the figures are functional entities that do not necessarily necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor devices and/or microcontroller devices.

分别为输入的跟踪误差和跟踪误差变化率，u(t)表示输出控制量，此处du/dt表示后移算子的功能，亦即求取u(t-1)，上一时刻的控制量。FNN表示模糊神经网络控制器。K表示模糊神经网络控制器的比例系数，该参数根据运行结果不断调整。Referring to FIG. 1, FIG. 1 is a schematic diagram of a method for controlling the loss rate of a corn harvester according to an embodiment of the present invention. As shown in Figure 1, the corn harvester loss rate control method of the present invention is based on a fuzzy neural network, and a fuzzy neural network controller with dual input and single output is constructed to detect the speed tracking output of the grain recovery device in real time and the setting of the standard speed. The fixed value, combined with the online learning mechanism, adjust the adjustable parameters in the controller in real time, so that it can adapt to the change of the grain recovery device and track the set value of the standard speed. The fuzzy neural network controller is a two-dimensional structure with two inputs and one output, taking e and

are the input tracking error and tracking error rate of change, respectively, u(t) represents the output control amount, where du/dt represents the function of the backshift operator, that is, to obtain u(t-1), the control at the previous moment quantity. FNN stands for Fuzzy Neural Network Controller. K represents the proportional coefficient of the fuzzy neural network controller, and this parameter is continuously adjusted according to the running results.

The learning stage of fuzzy neural network controller is divided into offline learning and online learning stage. The offline learning stage is to preliminarily determine the center and width of the weights and membership functions that need to be learned in the fuzzy neural network by learning the previous system operating data. These determined values are not very precise, and then the controller whose parameter structure is preliminarily determined is precisely adjusted by the BP algorithm in the online learning stage, so that the control performance is better. Offline learning adopts improved particle swarm algorithm, namely genetic algorithm-particle swarm algorithm, and online learning is BP algorithm. The BP algorithm is overly dependent on the initial value of the network, and a poor initial value may lead to poor performance or no convergence at all. In addition, the global search ability of BP algorithm is poor, and it is easy to fall into local minima. The combination of PSO (Particle Swarm Optimization) and BP algorithm can not only ensure the global convergence of learning, but also overcome the dependence of the gradient method on the initial value and local convergence problems, and also overcome the randomness caused by the simple particle swarm algorithm. question of probability and probability.

Specifically, the following steps may be included:

Step S100, constructing a dual-input single-output fuzzy neural network controller, where the fuzzy neural network controller includes an antecedent network and a consequent network;

Step S200, adopting the genetic algorithm-particle swarm algorithm for offline learning, and preliminarily determining the weights and the center and width of the membership function to be learned in the fuzzy neural network controller by learning the previous operating data of the system, wherein the system is first For example, the operating data can set the speed regulation model of the motor, the input quantity of the drive system, etc., and set the initial speed of the motor; and

Step S300, using the BP algorithm for online learning, establishing the connection weights of the fuzzy neural network controller, detecting the rotating speed of the corn harvester grain recovery device in real time, and adjusting the adjustable parameters in the fuzzy neural network controller in real time in combination with the online learning. parameters, so that the fuzzy neural network controller adapts to the mechanical performance changes of the corn harvester grain recovery device, and tracks the corn harvest rate set value.

Referring to FIG. 2, FIG. 2 is a schematic structural diagram of a fuzzy neural network according to an embodiment of the present invention. The fuzzy neural network in this embodiment is composed of an antecedent network and a consequent network. The antecedent network is used to match the antecedents of the fuzzy rule, and the consequent network is used to generate the consequent of the fuzzy rules. The antecedent network is a four-layer network structure, including:

The first layer: the input layer, the active power deviation and the active power deviation change rate of the input corn harvester are taken as y ₁ and y ₂ respectively, and the number of nodes in the input layer is N ₁ =2, and the expression is as follows:

为性能参数变量跟踪误差变化率，c为机械性能变化值，y为机械性能实际检测值；where e is the tracking error,

is the change rate of the tracking error of the performance parameter variable, c is the change value of the mechanical property, and y is the actual detection value of the mechanical property;

Fuzzy layer, the input variables y ₁ , y ₂ are divided into 7 fuzzy subsets {NB, NM, NS, O, PS, PM, PB}, as the nodes of the fuzzy layer, each node represents a The values of linguistic variables; their membership functions are all Gaussian functions, and the membership functions of each linguistic variable are:

为第一层的输入在相应语言变量值的模糊论域的隶属度函数，y_i为输入玉米收获机的有功功率偏差，c_ij和σ_ij(i＝1，2，…，n，j＝1，2，…m_i)分别为隶属度函数的中心和宽度，n为输入变量个数，m_i为输入变量x_i的模糊分割数，n＝2，m₁＝m₂＝7；所述模糊化层的节点数N₂＝m₁+m₂＝14；in,

is the membership function of the input of the first layer in the fuzzy universe of the corresponding linguistic variable value, y _i is the active power deviation of the input corn harvester, c _ij and σ _ij (i=1, 2,...,n,j= _{1,2 , … _ _} The number of nodes in the fuzzification layer N ₂ =m ₁ +m ₂ =14;

The third layer: the fuzzy rule calculation layer, which is used to match the antecedents of the fuzzy rules and calculate the fitness of each fuzzy rule. Since there are two inputs y ₁ and y ₂ , the fuzzy inference operation is to combine the two A continuous multiplication operation is performed on the fuzzified input quantities, and the fuzzy operator used is the continuous multiplication operator:

为隶属度值，

为下一节点隶属度值，j₁＝j₂＝1，2，…，7；，m＝m₁×m₂＝7×7＝49，所述模糊规则计算层的节点数N₃＝49；Among them, α _m is the calculation result of the fuzzy inference operation,

is the membership value,

is the membership value of the next node, j ₁ =j ₂ =1,2,...,7; m=m ₁ ×m ₂ =7×7=49, the number of nodes in the fuzzy rule calculation layer N ₃ =49 ;

The fourth layer: the normalization layer, which is used to realize the normalization operation;

为加权系数，α_i为加权系数累加，所述归一化层的节点数N₄＝N₃＝49，该层节点数与第三层节点数相同。in,

is the weighting coefficient, α _i is the weighting coefficient accumulation, the number of nodes in the normalization layer is N ₄ =N ₃ =49, and the number of nodes in this layer is the same as the number of nodes in the third layer.

The postware network in this embodiment is a three-layer network structure, including:

The first layer, the input layer, is used to pass the input variable to the next layer, that is, the second layer. The first layer has 3 nodes in total. The input value of the first node in the input layer is x ₀ =1, and its The function is used to provide the constant term in the consequent of the fuzzy rule; the second and third nodes respectively input x ₁ , x ₂ ;

The second layer is used to calculate the consequent of each rule. There are 49 nodes in total, and each node represents a rule:

为连接权值，k＝0，1，2，m＝1，2，3，…，49，x₁为对应节点的规则，x₂为第二节点对应规则；where y _m is the consequent network output of the mth rule,

is the connection weight, k=0, 1, 2, m=1, 2, 3,..., 49, x ₁ is the rule corresponding to the node, and x ₂ is the corresponding rule of the second node;

The third layer, used to calculate the controller output y:

Among them, α _m is the weighting coefficient, that is, the normalized fitness of each fuzzy rule, y _m is the weighted sum of each rule consequent, the fuzzy neural network output y is the weight of each fuzzy rule consequent, and the weighting coefficient is the normalization of each fuzzy rule. The normalized usage degree, that is, the output of the antecedent network is the connection weight of the consequent network.

The offline learning stage of the T-S fuzzy neural network learning algorithm: the offline learning stage is to provide an excellent initial value of the network for online learning, because the online learning algorithm - BP algorithm is very dependent on the selection of the initial value, and the initial value is better than the optimal training effect. it is good. Offline learning adopts an improved particle swarm algorithm, namely GA-PSO algorithm.

Particle swarm optimization is a swarm intelligence optimization algorithm, first proposed by Kennedy and Eberhart in 1995. It stems from the study of bird predation behavior. When birds prey, the way each bird finds food is to search the surrounding area of the bird that is currently closest to the food.

The standard particle swarm algorithm first randomly initializes a group of particles in the feasible solution space. Each particle represents a potential optimal solution of the extreme value optimization problem. The three indicators of position, speed and fitness value are used to represent the characteristics of the particle. It is calculated by the fitness function, and the quality of its value indicates the quality of the particle. The particle moves in the solution space, and the individual position is updated by tracking the individual extremum Pbest and the group extremum Gbest. The individual extremum Pbest represents the position with the best fitness value among the positions experienced by the individual, and the group extremum Gbest represents that all particles of the population have searched for it. The position where the fitness value is optimal.

Suppose that in a D-dimensional search space, a _{population Xi} = ( _{x 1} , _{x 2} . x _i2 ... x _iD ) ^T , which represents the position of the i-th particle in the D-dimensional search space, and also represents a potential solution to the problem. The velocity of the i-th particle is V _i =(v _i1 , v _i2 ...v _iD ) ^T , its individual extreme value P _i =(p ₁ , p ₂ ... p _n ) ^T , the population global extreme value P _g = (p _g1 , p _g2 . . . p _gn ) ^T .

The velocity and position update formulas of the standard particle swarm algorithm are as follows:

In the formula, ω is the inertia weight; d=1, 2,..., D; i=1, 2,...,n; k is the current iteration number; V _id is the speed of the particle, P _id is the expected output, and X _id is Actual output; c ₁ and c ₂ are acceleration factors, which are non-negative constants; r ₁ and r ₂ are random numbers distributed in [0, 1].

The fitness function adopted in the present invention is the absolute value of the error between the predicted output and the expected output as the individual fitness value F, and the calculation formula is:

F=abs(y-c); (9)

where y is the predicted output and c is the expected output.

The standard particle swarm algorithm completes extremum optimization by following the individual extremum and the group extremum. Although the operation is simple and it can converge quickly, as the number of iterations increases, while the population converges and concentrates, the particles become more and more similar. It is impossible to jump out near the local optimal solution. The hybrid particle swarm algorithm abandons the method of updating the particle position by tracking the extreme value in the standard particle swarm algorithm, but introduces the crossover and mutation operations in the genetic algorithm. mutation to search for the optimal solution.

The crossover operation adopts the real number crossover method, and randomly selects the ith individual extremum and the crossover position j of the ith individual extremum and the ith individual, and then performs the crossover. The method is as follows:

x _ij = x _ij (1-b)+P _ij b; (10)

x _ij =x _ij (1-b)+P _gj ; (11)

In the formula, x _ij is a randomly selected individual extreme value, P _ij is a randomly selected group extreme value, and b is a random number between [0, 1].

The particle self-mutation operation first randomly selects the i-th individual mutation position j, and then mutates, the method is as follows:

In the formula, x _max is the upper bound of x _ii ; x _min is the lower bound of x _ij ; f(g)=r ₂ (1-g/G _max ) ² ; r ₂ is a random number; g is the current iteration number; G _max is the maximum number of evolutions; r is a random number in [0, 1].

Referring to FIG. 3 , FIG. 3 is a flowchart of a GA-PSO offline learning algorithm according to an embodiment of the present invention. In this embodiment, the offline learning in step S200 includes the following steps:

Step S201, initializing a population parameter, where the population parameter is the initial position of each particle;

Step S202, calculating the particle fitness F: F=abs(y-c), where y is the predicted output, and c is the expected output;

Step S203, searching for the individual extreme value and the group extreme value, and finding out the individual minimum fitness value and the global minimum fitness value of each particle;

Step S204, using the following formula to calculate the speed update and position update of the particle:

Among them, ω is the inertia weight; d=1, 2,..., D; i=1, 2,...,n; k is the current iteration number; V _id is the velocity of the particle; c ₁ and c ₂ are the acceleration factors, which are Non-negative constant; r ₁ and r ₂ are random numbers distributed in [0, 1];

Step S205, calculate the particle fitness after speed and position update according to the formula in step S202;

Step S206, update the individual extreme value and the group extreme value according to the formula in step S204;

In step S207, the following formula is used to calculate the intersection of the current individual and the individual extreme value, and if the fitness value decreases, accept:

x _ij =x _ij (1-b)+P _ij b, where b is a random number between [0, 1];

In step S208, the following formula is used to calculate the intersection of the current individual and the group extreme value, and if the fitness value decreases, accept:

x _ij = x _ij (1-b)+P _gj , where x _ij is a randomly selected individual extreme value, P _gj is a randomly selected group extreme value, and b is a random number between [0, 1];

Step S209, the following formula is used to calculate the variation of the current individual itself, and if the fitness value decreases, accept:

In the formula, x _ij is a randomly selected individual extreme value, x _max is the upper bound of x _ii ; x _min is the lower bound of x _ij ; f(g)=r ₂ (1-g/G _max ) ² ; r ₂ is A random number; g is the current iteration number; G _max is the maximum number of evolutions; r is a random number of [0, 1];

In step S210, if the maximum evolutionary algebra is satisfied, the process ends; otherwise, the process returns to step S204.

Wherein, the adjustable parameters that need to be learned include the connection weight and the central value c _ij and the width σ _ij of the membership function.

The learning algorithm of the connection weight is:

The learning algorithm of the central value c _ij of the membership function is:

c _ij (τ+1)=c _ij (τ)+Δc _ij (τ+1)+υ(c _ij (τ)-c _ij (τ-1));

The learning algorithm of the width σ _ij is:

σ _ij (τ+1)=σ _ij (τ)+Δσ _ij (τ+1)+υ(σ _ij (τ)-σ _ij (τ-1));

E为误差代价函数，η为学习速率。In the formula, i=1, 2, j=1, 2, 3,..., 7, T represents the moment, T+1 represents the next moment, T-1 represents the previous moment, u is the momentum factor,

E is the error cost function, and η is the learning rate.

式中，c为期望输出，y为实际输出。The error cost function E is:

In the formula, c is the expected output and y is the actual output.

Online learning stage of T-S fuzzy neural network learning algorithm: BP algorithm is used for online learning.

Since the number of fuzzy divisions of each input variable is predetermined, the parameters to be learned are mainly the connection weights of the posterior network and the center value c _ij and the width σ _ij of the Gaussian membership function.

Define the error cost function E as:

In the formula, c is the expected output and y is the actual output.

的学习算法：About connection weights

The learning algorithm of:

In the learning stage of the controller, the offline learning results are unsatisfactory, so the BP algorithm in the online learning stage is used to further improve the center and width of the weights and membership functions required in the fuzzy neural network. The membership function of the input component adopts a Gaussian function, and the parameters to be learned are the connection weight and the central value c _ij and the width σ _ij of the membership function.

Among them, τ represents the time, τ+1 represents the next time, and τ-1 represents the previous time.

已知。Then, the learning algorithm of the center value c _ij and the width σ _ij is discussed. At this point the connection weight

A known.

In the above formulas, η>0 is the learning rate.

中心值c_ij和宽度σ_ij的学习算法分别为：The standard BP algorithm has a slow convergence speed, and the objective function has a local minimum problem. Nowadays, there are many methods to improve the above problems. There are two commonly used methods: when the momentum term is introduced, the BP algorithm can find a better solution; when the adaptive learning rate is introduced, the BP algorithm can appropriately shorten the training time . Therefore, this paper combines the two. At this time, the connection weights are

The learning algorithms for the center value c _ij and the width σ _ij are:

c _ij (τ+1)=c _ij (τ)+Δc _ij (τ+1)+υ(c _ij (τ)-c _ij (τ-1)); (20)

σ _ij (τ+1)=σ _ij (τ)+Δσ _ij (τ+1)+υ(σ _ij (τ)-σ _ij (τ-1)); (21)

E为误差代价函数，η为学习速率。In the formula, i=1, 2, j=1, 2, 3,..., 7, T represents the moment, T+1 represents the next moment, T-1 represents the previous moment, u is the momentum factor,

E is the error cost function, and η is the learning rate.

Although the various steps of the methods of the present invention are depicted in the figures in a particular order, it is not required or implied that the steps must be performed in the particular order or that all illustrated steps must be performed to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step for execution, and/or one step may be decomposed into multiple steps for execution, and the like.

In addition, the present invention also provides a corn harvester loss rate control device, including a fuzzy neural network controller. The structure of the dual-input and single-output fuzzy neural network controller in this embodiment can be determined according to fuzzy rules and their physical meanings. The fuzzy neural network controller includes an antecedent network and a consequent network, and adopts the above-mentioned control method for the loss rate of the corn harvester to reduce the loss rate of corn harvesting by optimally controlling the rotational speed of the corn harvester grain recovery device. It should be noted that this means is used for the execution of actions, which may realize its functions by several modules or units. Indeed, according to embodiments of the present invention, the features and functions of two or more modules or units may be embodied in one module or unit. Conversely, the features and functions of one module or unit can be further divided into multiple modules or units to be embodied.

Correspondingly, based on the same inventive concept, the present invention also provides a storage medium storing a computer program, and the computer program is configured to execute the above-mentioned method for controlling the loss rate of a corn harvester when running. From the description of the above embodiments, those skilled in the art can easily understand that the exemplary embodiments described herein may be implemented by software, or may be implemented by software combined with necessary hardware. Therefore, the technical solution according to the embodiment of the present invention can be embodied in the form of a software product, and the software product can be stored in a non-volatile storage medium (for example, it can be a CD-ROM, a U disk, a mobile hard disk, etc.) or a network Above, several instructions are included to cause a computing device (which may be a personal computer, a server, a mobile terminal, or a network device, etc.) to execute the method according to an embodiment of the present invention.

In some possible implementations, aspects of the present invention can also be implemented in the form of a program product comprising program code for enabling the program product to run on a terminal device The terminal device performs the steps according to various exemplary embodiments of the present invention described in the "Example Method" section above in this specification. Optionally, in this embodiment, the above-mentioned storage medium may include, but is not limited to: a USB flash drive, a read-only memory (Read-Only Memory, referred to as ROM for short), a random access memory (Random Access Memory, referred to as RAM for short), mobile Various media that can store computer programs, such as hard disks, magnetic disks, or optical disks.

A program product for implementing the above method according to an embodiment of the present invention may adopt a portable compact disc read only memory (CD-ROM) and include program codes, and may run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto, and in this document, a readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus or device, or a combination of any of the above. More specific examples (non-exhaustive list) of readable storage media include: electrical connections with one or more wires, portable disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

A computer readable signal medium may include a propagated data signal in baseband or as part of a carrier wave with readable program code embodied thereon. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing. A readable signal medium can also be any readable medium, other than a readable storage medium, that can transmit, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.

Program code embodied on a readable medium may be transmitted using any suitable medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Program code for carrying out operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages, such as Java, C++, etc., as well as conventional procedural A programming language, such as the "C" language or similar programming language. The program code may execute entirely on the user computing device, partly on the user device, as a stand-alone software package, partly on the user computing device and partly on a remote computing device, or entirely on the remote computing device or server execute on. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (eg, using an Internet service provider business via an Internet connection).

Correspondingly, based on the same inventive concept, the present invention also provides an electronic device, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the Instructions are executable to perform the corn harvester loss rate control method described above.

As will be appreciated by one skilled in the art, various aspects of the present invention may be implemented as a system, method or program product. Therefore, various aspects of the present invention can be embodied in the following forms: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software aspects, which may be collectively referred to herein as implementations "circuit", "module" or "system". The electronic device of this embodiment is represented as a general-purpose computing device. The components of the electronic device may include, but are not limited to: the above-mentioned at least one processor, the above-mentioned at least one memory, and a bus connecting different system components (including the memory and the processor). The memory is used to store executable instructions of the processor; the processor is configured to perform the above-described corn harvester loss rate control method by executing the executable instructions. Wherein, the memory stores program codes that can be executed by the processor to cause the processor to perform the various exemplary embodiments of the present invention described in the "Exemplary Methods" section above in this specification. step.

The memory may include readable media in the form of volatile storage units, such as random access storage units (RAM) and/or cache storage units, and may further include read only storage units (ROM).

The memory may also include a program/utility having a set (at least one) of program modules including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of these examples One or some combination may include an implementation of a network environment.

The bus can be one or more representing several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of a variety of bus structures .

The electronic device may also communicate with one or more external devices (eg, keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device, and/or communicate with the electronic device. A device can communicate with any device (eg, router, modem, etc.) that communicates with one or more other computing devices. This communication can take place through an input/output (I/O) interface. Also, the electronic device may communicate with one or more networks (eg, a local area network (LAN), a wide area network (WAN), and/or a public network such as the Internet) through a network adapter. The network adapter communicates with other modules of the electronic device through the bus. Other hardware and/or software modules may be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

From the description of the above embodiments, those skilled in the art can easily understand that the exemplary embodiments described herein may be implemented by software, or may be implemented by software combined with necessary hardware. Therefore, the technical solutions according to the embodiments of the present invention can be embodied in the form of software products, and the software products can be stored in a non-volatile storage medium (which can be CD-ROM, U disk, mobile hard disk, etc.) or on the network , including several instructions to cause a computing device (which may be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to an embodiment of the present invention.

Based on the fuzzy neural network algorithm, the invention integrates the learning and computing functions of the neural network into the fuzzy system, and embeds the human-like IF-Then rules of the fuzzy system into the neural network, so as to maintain the strong knowledge expression ability of the fuzzy control system and improve its It has self-adaptive ability and self-learning ability. Through the optimization of the neural network algorithm for controlling the motor speed of the grain recovery device of the corn planter, the loss rate of corn harvesting is reduced.

Of course, the present invention can also have other various embodiments, without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and modifications according to the present invention, but these corresponding Changes and deformations should belong to the protection scope of the appended claims of the present invention.

Claims (10)

1. a corn harvester loss rate control method, is characterized in that, comprises the steps: S100, constructing a dual-input and single-output fuzzy neural network controller, where the fuzzy neural network controller includes an antecedent network and a consequent network; S200, adopting the genetic algorithm-particle swarm algorithm for offline learning, and preliminarily determining the weights and the center and width of the membership function to be learned in the fuzzy neural network controller by learning the previous operating data of the system; and S300, using the BP algorithm to perform online learning, establishing the connection weights of the fuzzy neural network controller, detecting the rotation speed of the grain recovery device of the corn harvester in real time, and adjusting the adjustable parameters in the fuzzy neural network controller in real time in combination with the online learning , so that the fuzzy neural network controller adapts to the mechanical performance changes of the corn harvester grain recovery device, and tracks the corn harvest rate set value. 2. The method for controlling the loss rate of a corn harvester as claimed in claim 1, wherein the antecedent network is a four-layer network structure, comprising: In the input layer, the active power deviation and the active power deviation change rate of the input corn harvester are taken as y ₁ and y ₂ respectively, and the number of nodes in the input layer is N ₁ =2, and the expression is as follows:

where e is the tracking error,

is the change rate of the tracking error of the performance parameter variable, c is the change value of the mechanical property, and y is the actual detection value of the mechanical property; Fuzzy layer, the input variables y ₁ , y ₂ are divided into 7 fuzzy subsets {NB, NM, NS, O, PS, PM, PB}, as the nodes of the fuzzy layer, each node represents a The values of linguistic variables; their membership functions are all Gaussian functions, and the membership functions of each linguistic variable are:

in,

is the membership function of the input of the first layer in the fuzzy domain of the corresponding linguistic variable value, y _i is the active power deviation of the input corn harvester, c _ij and σ _ij (i=1,2,...,n,j= _{1 , 2 , …} The number of nodes in the fuzzification layer N ₂ =m ₁ +m ₂ =14; The fuzzy rule calculation layer is used to match the antecedents of the fuzzy rules, and calculate the fitness of each fuzzy rule:

Among them, α _m is the calculation result of the fuzzy inference operation,

is the membership value,

is the membership value of the next node, j ₁ =j ₂ =1,2,...,7; m=m ₁ ×m ₂ =7×7=49, the number of nodes in the fuzzy rule calculation layer N ₃ =49 ; The normalization layer is used to realize the normalization operation;

in,

is the weighting coefficient, α _i is the weighting coefficient accumulation, and the number of nodes in the normalization layer is N ₄ =N ₃ =49. 3. The method for controlling the loss rate of a corn harvester as claimed in claim 2, wherein the latter network is a three-layer network structure, comprising: The first layer is used to pass the input variable to the next layer, the first layer has 3 nodes in total, and the input value of the first node is x ₀ =1, which is used to provide the constant term in the consequent of the fuzzy rule; The second and third nodes respectively input x ₁ , x ₂ ; The second layer is used to calculate the consequent of each rule. There are 49 nodes in total, and each node represents a rule:

where y _m is the consequent network output of the mth rule,

is the connection weight, k=0, 1, 2, m=1, 2, 3,..., 49, x ₁ is the rule corresponding to the node, and x ₂ is the corresponding rule of the second node; The third layer, used to calculate the controller output y:

Among them, α _m is the weighting coefficient, that is, the normalized fitness of each fuzzy rule, y _m is the weighted sum of the consequent components of each rule, and the output of the antecedent network is the connection weight of the consequent network. 4. corn harvester loss rate control method as claimed in claim 3, is characterized in that, the off-line learning in step S200 comprises the following steps: S201, initialize a population parameter, where the population parameter is the initial position of each particle; S202. Calculate the particle fitness F: F=abs(y-c), where y is the predicted output, and c is the expected output; S203, find the individual extreme value and the group extreme value, and find out the individual minimum fitness value and the global minimum fitness value of each particle; S204, the following formulas are used to calculate the speed update and position update of the particle:

Among them, ω is the inertia weight; d=1,2,...,D; i=1,2,...,n; k is the current iteration number; V _id is the speed of the particle, P _id is the expected output, and X _id is the actual Output; c ₁ and c ₂ are acceleration factors, which are non-negative constants; r ₁ and r ₂ are random numbers distributed in [0,1]; S205, according to the formula in step S202, calculate the particle fitness after speed and position update; S206, update the individual extreme value and the group extreme value according to the formula in step S204; S207, the following formula is used to calculate the intersection of the current individual and the individual extreme value, if the fitness value decreases, accept: x _ij = x _ij (1-b)+P _ij b, where x _ij is a randomly selected individual extreme value, P _ij is a randomly selected group extreme value, and b is a random number between [0,1]; S208, the following formula is used to calculate the intersection of the current individual and the group extreme value, and if the fitness value decreases, accept: x _ij = x _ij (1-b)+P _gj , where x _ij is a randomly selected individual extreme value, P _gj is a randomly selected group extreme value, and b is a random number between [0,1]; S209, the following formula is used to calculate the variation of the current individual itself, and if the fitness value decreases, accept:

In the formula, x _ij is a randomly selected individual extreme value, x _max is the upper bound of x _ij ; x _min is the lower bound of x _ij ; f(g)=r ₂ (1-g/G _max ) ² ; r ₂ is A random number; g is the current iteration number; G _max is the maximum number of evolutions; r is a random number of [0,1]; S210. End if the maximum evolutionary algebra is satisfied, otherwise return to step S204. 5 . The method for controlling the loss rate of a corn harvester according to claim 4 , wherein the adjustable parameters to be learned include connection weights and the central value c _ij and width σ _ij of membership functions. 6 . 6. corn harvester loss rate control method as claimed in claim 5, is characterized in that, the learning algorithm of described connection weight is:

The learning algorithm of the central value c _ij of the membership function is: c _ij (τ+1)=c _ij (τ)+Δc _ij (τ+1)+υ(c _ij (τ)-c _ij (τ-1)); The learning algorithm of the width σ _ij is: σ _ij (τ+1)=σ _ij (τ)+Δσ _ij (τ+1)+υ(σ _ij (τ)-σ _ij (τ-1)); In the formula, i=1,2, j=1,2,3,...,7, τ represents the moment, τ+1 represents the next moment, τ-1 represents the previous moment, υ is the momentum factor,

E is the error cost function, and η is the learning rate. 7. corn harvester loss rate control method as claimed in claim 6, is characterized in that, described error cost function E is:

In the formula, c is the expected output and y is the actual output. 8. A corn harvester loss rate control device, characterized in that it comprises a fuzzy neural network controller, the fuzzy neural network controller comprises an antecedent network and a consequent network, and adopts any one of the above claims 1-7. In the method for controlling the loss rate of the corn harvester described in item 1, the reduction of the loss rate of the corn harvest is realized by optimally controlling the rotational speed of the grain recovery device of the corn harvester. 9 . A storage medium, characterized in that the storage medium stores a computer program, and the computer program is configured to execute the method for controlling the loss rate of a corn harvester according to any one of claims 1 to 7 when running. . 10. An electronic device, comprising: processor; and a memory for storing executable instructions for the processor; Wherein, the processor is configured to perform the corn harvester loss rate control method of any one of the preceding claims 1-7 by executing the executable instructions.

Images