One kind is based on wireless sensor network hen house microclimate environment intelligence control system
Technical field
The present invention relates to intelligent control device technical field is and in particular to one kind is stingy based on wireless sensor network hen house
Wait intelligent environment control system.
Background technology
With the fast development of feeding chicken in largely scale in recent years, intensive, the industrialization degree of poultry husbandry improve constantly, hen house
Microclimate environment is more and more big to the health of chicken and the impact of production performance, be chicken group create a suitable environmental condition be to prolong
Long laying hen in egg laying peak period, improves the key link of poultry economic benefit.Resistance against diseases with chicken and the general level of the health are in decline
Trend, this just more requires the environment controlling technique of feeding chicken in largely scale will constantly keep up with the needs that modern poultry produces, and adjusts in good time
The target of whole poultry environmental Kuznets Curves and strategy, the new environment of chicken house control technology of continuous development and application and equipment, be
The healthy and safe production of modern poultry husbandry provides safeguard and technical support.
The target of feeding chicken in largely scale hen house environment control is summed up production performance, the warp of operator mainly improving chicken
Ji benefit, health of guarantee chicken group etc..At present, the target that feeding chicken in largely scale gives up environmental Kuznets Curves is from raising chicken group's productivity mostly
The angle of energy, microclimate environment condition to determine suitable Environment Design parameter value with the correlation of chicken group's production performance.
Environment temperature is laid eggs to the surviving of chicken, sexal maturity, fertilization, egg size, quality of egg and efficiency of feed utilization have obvious
Impact.Chicken group lives in bad environment for a long time, and environment of chicken house is severe, and chicken resistance reduces, and chicken group's incidence of disease is high, the death rate
Increase, production performance declines relevant.Adjustment environment of chicken house, had both improved chicken all living creatures's dis environment, had improve the opposing to disease again
Power, had both saved unnecessary consumption cost, more improve production performance.Environment of chicken house controls and plays very in poultry production
Important effect, environmental Kuznets Curves are the optimal selections solving the problems, such as epidemic disease at present and improving production performance, and humiture is impact chicken
Group's health and the primary environment factor of performance.Temperature is the primary warm factor of impact chicken health and production performance.Work as ring
When border temperature reduces, the heat dissipation capacity of body increases, and chicken is in order to maintain thermostasis it is necessary to improving metabolic rate, increasing quantity of heat production.When
When temperature is too low, body occurs shivers, and so that metabolic heat production is greatly increased to maintain thermostasis.If under environment temperature continues
Fall, body can cause temperature decline because loss of heat is excessive, can cause death when serious.If temperature raises, body radiating is obstructed,
Physics adjusts and can not maintain thermostasis, internal accumulation of heat, thus causing metabolic rate to improve.When temperature continues to raise, machine is realized
The existing hot rising panted with body temperature.The optimum temperature of laying hen is 13 1 23 DEG C, and extremes of temperature is all unfavorable for laying eggs,
The too low efficiency of feed utilization of temperature is poor.The normal growth of suitable humidity and chicken develops closely related, the impact to chicken for the humidity
Combine with temperature and concur.In high temperature, high humidity easily causes going mouldy of feed and bedding and padding, and chick feather is dirty
Dirty, lead to bird to break out mycosis and global-worm illness.Relative humidity is higher, illustrates that air is the more humid, relative humidity is lower, then illustrates
Air is drier.Humidity is too low, and the feather of chicken is disorderly, dry skin, and the color and luster such as feather, beak, pawl is dim, and easily causes chicken body
Dehydration and cause various respiratory tract disease.The height of humidity to chick with become chicken to have considerable influence.As chick relative humidity
Too low, chick internal water is passed through breathing and is distributed in a large number, has a strong impact on remaining yolk in chick body and absorbs, feather growth
Bad, easy hyperposia and diarrhea, also easily cause breathing problem, relative humidity is too high, not only tide temperature in house, and chick
Regular heat produced and moisture evaporation are obstructed, and chick feels sultry discomfort, and this moist environment is beneficial to pathogenic fungi, carefully again
Bacterium and the growth and breeding of parasite, are easily caused feed, bedding and padding are mould loses, make chick break out infectious disease, respiratory tract, parasitic disease etc..
The optimum humidity of laying hen is 60% one 70%, if humidity is too low in house, may result in that laying hen feather is disorderly, skin is done
Dry, the color and luster such as feather and prolixity pawl is dim, and easily causes the dehydration of chicken body and the breathing problem causing chicken group.If wet in house
Spend height, the moisture of dissipation in the air during chicken breathing will be made to be difficult to distribute, cause chicken body filthy, germ amount reproduction, lead
Cause various diseases, cause the decline of egg production, the threshold moisture of laying hen raises with temperature and declines.Therefore, in daily management
In, the insulation relation wet with control will be handled well.Keep that in house air is fresh, well-ventilated is highly important.Especially in high density
The hen house raised, this problem is even more important.Because drawing badly, have substantial amounts of pernicious gas such as ammonia, titanium dioxide at any time
Carbon and hydrogen sulfide etc. discharge, and affect normally laying eggs and causing multiple diseases of chicken.In the winter time, the air velocity around chicken body
It is advisable with 0.1-0.2 meter per second, not more than 0.35 meter per second;Summer should improve the air velocity in hen house as far as possible, increases logical
Air quantity.Most of plants, using longitudinal exhaust ventilation, install blower fan on the wall at hen house end and are outwards evacuated now,
Air inlet is arranged on other end wall, the air intake into house.Require house interior air-flow uniformly, no dead angle.Harmful wind is larger to the harm of chicken,
Chicken body local can be made to catch a cold, cause part frozen, arthritis, muscle inflammation, neuritis, flu, or even the disease such as pneumonia, paralysis.
When ventilation, hot blast etc. can be sent inwards in the wall setting wall with flues of air inlet or with warm-air drier.Chicken belongs to needs the long-day
Animal, illumination not only affects the growth of chicken, and chicken production performance is had an impact.Illumination is larger on the egg laying performance impact of chicken,
Rational illumination energy induction of ovulation, promotes the normal growth of chicken to develop, increases egg production.Rational illumination energy induction of ovulation, increases
Plus egg production.Illumination is larger on the egg laying performance impact of chicken, and rational illumination can promote the normal growth of chicken to develop, and stimulates row
Ovum, increases egg production.Finishing period intensity of illumination is 5 luxs, and laying hen and breeder intensity of illumination are positively retained at 10 luxs, cage
Raising layer illumination should improve.The Controlling principle of light application time is chick nurturing period shortening illumination in production, and during laying eggs
Increase illumination.Therefore, hen house humiture, air-flow and illuminance produce material impact to the production of chicken.
Gold washes the problems such as be different from external environment once for totally enclosed Laying House design feature and the indoor environment factor,
Design closed Laying House environmental control system, be capable of temperature in hen house, humidity, illuminance, carbon dioxide, ammonia and
Hydrogen sulfide is according to setting value auto-control;Liu Huanping, from the actual demand of closed Laying House, devises closed laying hen
House long distance control system, and elaborate closed Laying House environmental control system hardware circuit and programming design, including profit
Carry out method and the program circuit of remote measurement and control information transfer with gsm module TC35i with reference to STC12C5A60AD single-chip microcomputer.Lee
The hen house humiture real-time monitoring system based on wireless sensor network for the beautiful China design, this system can be simultaneously to difference in hen house
Temperature, humidity carry out real-time monitoring, and send the data to remote monitoring center, be capable of hen house humiture long-range,
Accurately measure.The exploitation of these systems effectively promotes the fast development of poultry husbandry, but these control systems are all not bound with
The features such as hen house microclimate environment multivariable, non-linear, close coupling and big inertia, fundamentally solve hen house microclimate environment ginseng
The problems such as number is unstable, control accuracy is low and system response time is slow.
Content of the invention
It is an object of the invention to provide a kind of hen house microclimate environment intelligence control system, the present invention is by existing hen house
The feature of cultivation microclimate environment parameter multivariable, non-linear, close coupling and big inertia, devises one kind and is based on wireless senser
Network hen house microclimate environment intelligence control system.
The present invention is achieved through the following technical solutions:
This system is by the hen house of hen house microclimate environment parameter acquisition and control platform, PID neutral net based on DRNN
Miniclimate breeding environment cascade control system and the setting based on economic optimum reasoning are controlled hen house microclimate environment parameter to manage
Think the expert system composition of value.
1) data acquisition of hen house microclimate environment and control platform:Including the detection node of environment of chicken house parameter, control section
Point and on-site supervision end composition, can be built into tracking and command network by wireless sensor network interconnection by them in hen house cultivation site
Network, realizes the parameters such as the temperature to hen house breeding environment, humidity, air-flow and illuminance and is detected and control.(see Fig. 1)
2) cultivate the feature of microclimate environment nonlinearity in parameters, close coupling and big inertia according to hen house, monitor at the scene
The hen house miniclimate breeding environment cascade control system designing the PID neutral net based on DRNN in end and control node improves
The stability of microclimate environment control system, precision and the response speed by control parameter.(see Fig. 2)
The hen house miniclimate breeding environment cascade control system of the PID neutral net based on DRNN for the design is realized to scale
Hen house cultivation microclimate environment parameter carries out uneoupled control;According to the different growth phases of chicken, design is based on economic optimum reasoning
Setting controlled the control to hen house microclimate environment parameter for the Implementation of Expert System of hen house microclimate environment parameter ideal value
Desired value Scientific Establishment.This control system by being controlled humiture, air-flow and illuminance parameter PID regulator, by control parameter
DRNN neural network identifier, neural network decoupling controller and the setting based on economic optimum reasoning controlled hen house stingy
Wait the expert system composition of ambient parameter ideal value, wherein monitoring client designs PID regulator, DRNN neural network identification at the scene
Device, neural network decoupling controller and expert system, realize controlling 4 PI pairs of hen house microclimate environment parameter in control node
Adjuster.The humiture of hen house microclimate environment, air-flow and illuminance are master variable, the voltage of heater, wet curtain motor
The voltage of rotating speed, the rotating speed of blower fan and light compensating lamp is secondary control variable.Hen house microclimate environment 4 is adjusted by 4 PID of control parameter
The PREDICTIVE CONTROL to them for the DRNN neural network identification model realization of section device and 4 parameters, by distinguishing to by control parameter
The error of knowledge and given ideal value to adjust the output of PID in time, solves the problems, such as the non-linear of control target and large time delay, god
Realize uneoupled control to hen house microclimate environment parameter through network decoupling controller, solve 4 by the coupling between control parameter
Conjunction problem, improves control accuracy and the response speed of system;Exporting of neural network decoupling controller is adjusted to 4 PI of inner ring are secondary
Device is controlling on the influential row mechanism of hen house microclimate environment parameter.
The main regulation loop of 4 PID regulator based on DRNN and neural network decoupling controller is stingy for hen house
Wait the multivariable of ambient parameter, the feature of nonlinear characteristic, close coupling and hysteresis quality, by have adaptive ability based on DRNN
4 PID and neural network decoupling controller be applied to the control of hen house microclimate environment parameter humiture, air-flow and illuminance
System, the method utilizes the identification model output of DRNN neutral net real-time online parameter and the error of preferable set-point to optimize PID
Master selector parameter, leads to control due to non-linear serious, close coupling and hysteresis quality for solving hen house microclimate environment parameter
The difficult problem of system has stronger pin when property, and this system suitability of the results show is strong, good stability, response speed and control
Precision processed meets the requirement of control system.When the humiture of controlled hen house microclimate environment, air-flow and illuminance parameter are inclined
From default value when, main regulation loop PID, neural network decoupling controller and DRNN identification model enter row operation, its
Output is entered row operation, is led to as the secondary set-point adjusting PI loop of 4 executing agencies, the secondary controller PI then corresponding to mechanism
The state crossing the corresponding executing agency of change makes being changed by control parameter of hen house microclimate environment, until little close to hen house
Till the set-point of climatic condition feature.The effect of 4 PID master selectors based on DRNN and neural network decoupling controller is
The set-point of 4 PI secondary controllers of the executing agency of adjustment in time, the serials control being made up of master selector and secondary controller
Effect is to ensure that the voltage of heater, the stablizing of the voltage of the rotating speed of wet curtain motor, the rotating speed of blower fan and light compensating lamp, and improves system
Response speed and control accuracy.
, on the basis of serials control, hen house microclimate environment parameter is using 4 PID regulator based on DRNN for the present invention
With neural network decoupling controller come hen house microclimate environment parameter is controlled, this control method guarantees hen house miniclimate
Ambient parameter is stablized near set-point, and this control algolithm can give full play to PID, uneoupled control, serials control and pre- observing and controlling
The advantage of system, this control method either antijamming capability still in terms of robustness and traditional cas PID control and routine
ANN Control compare and be all greatly improved.The hen house miniclimate cultivation ring of the PID neutral net based on DRNN
Border cascade control system, improves the Control platform of hen house microclimate environment parameter, improves system response time, stabilizes chicken
House microclimate environment temperature, humidity, air-flow and illuminance are it is suppressed that the impact to hen house microclimate environment parameter for the many factors.
(see part under Fig. 2)
3) controlled the expert system of hen house microclimate environment parameter ideal value based on the setting of economic optimum reasoning:Existing
The Implementation of Expert System based on economic optimum reasoning for the field monitoring client design is to hen house microclimate environment humiture, air-flow and illuminance
The control targe of parameter carries out Scientific Establishment.This system is based on microclimate environment humiture, air-flow becomes with illuminance state modulator
Basis, the growth model of chicken, feed-weight ratio and growth phase, Economic Benefit Model constitute setting hen house microclimate environment ideal parameters
Expert system, this system can be according to hen house microclimate environment humiture, air-flow and illuminance parameter regulation and control cost, growth mould
Type, feed-weight ratio and growth phase science determine the hen house microclimate environment temperature being in different growth phases broiler chicken or laying hen growth
The ideal parameters value of humidity, air-flow and illuminance, improves poultry benefit.Effectively overcome and only set not with poultry operating personnel's experience
The setting value of hen house microclimate environment parameter in the same time, improves utilization rate and the fanning economics of feed.(see part on Fig. 2)
The utility model compared with prior art, has following obvious advantage:
1st, the hen house miniclimate breeding environment cascade control system of the PID neutral net based on DRNN for the present invention employing, 4
The PI secondary controller of executing agency is the master selector output according to hen house microclimate environment humiture, air-flow and illuminance parameter
To adjust the servomechanism of hen house microclimate environment executing agency state, 4 secondary controller loops are as far as possible controlled mistake
In journey, the impact to hen house microclimate environment humiture, air-flow and illuminance parameter changes violent, frequent, big mainly the disturbing of amplitude
Moving includes in secondary controller loop, the electricity on the voltage being included in impact heater, wet curtain for these secondary controller loops
The Secondary Disturbance of the voltage of machine rotating speed, rotation speed of fan and light compensating lamp has very strong rejection ability and an adaptive ability, secondary disturbs
Dynamic adjusting to main controlled volume hen house microclimate environment humiture, air-flow and illuminance parameter by major and minor regulator loop
Impact very little, so the output valve amplitude of variation of hen house microclimate environment parameter is little, stable, system being capable of fast speed response control
The requirement to hen house microclimate environment parameter for the system.
2nd, the present invention passes through hen house microclimate environment humiture, air-flow and illuminance parameter DRNN neural network identifier mould
Type is estimated to hen house microclimate environment parameter, by the error of this discreet value and hen house microclimate environment parameter set-point not
The output of disconnected 4 PID regulator revising microclimate environment humiture, air-flow and illuminance parameter, by neural network decoupling
Controller exports makes the actual value of hen house microclimate environment parameter close to predetermined setting value at any time to 4 PI secondary controllers,
This control system does not allow controlled variable hen house microclimate environment humiture, air-flow and illuminance parameter to there is deviation and accurately protect
Hold regulated variable and meet production requirement, when hen house microclimate environment parameter occurs disturbance, just can quickly hen house microclimate environment be joined
Count required for being adjusted to numerically.What this control system was implemented to error is prediction closed-loop control, can adapt to many disturbances
Factor affects, and has good robustness.
3rd, hen house microclimate environment humiture of the present invention, air-flow and illuminance parameter are by PID regulator, neutral net solution
Coupling controller and the collective effect of neural network identifier model, the feedback forecasting cascade control system of PI secondary controller composition
Under, the stability of hen house microclimate environment humiture, air-flow and illuminance parameter is greatly improved, and this control system is to chicken
The control of the house hysteresis quality of microclimate environment parameter and close coupling object has very strong robustness, control system to set-point with
Track characteristic is good.The Control platform of this control system, response speed and stability all substantially carry compared with original conventional control
Height, the stability of the control accuracy height, strong antijamming capability and parameter of hen house microclimate environment parameter is good, preferably should have
With and promotional value.
4th, the present invention, according to decoupling principle and neutral net thought, proposes a kind of self-adaptive PID warp of two-stage series connection structure
Network decoupling controller. prime is the PID neural network decoupling controller based on DRNN, and rear class PI secondary controller is proposed
Decoupling controller has good robustness and adaptive decoupling ability, is to solve the problems, such as multivariable, non-linear and close coupling
A kind of simple and effective control algolithm, improves the multivariable of hen house microclimate environment humiture, air-flow and illuminance parameter, strong coupling
Close the dynamic characteristic with delayed sexual system and decoupling ability.
5th, the present invention adopts expert system to the hen house microclimate environment humiture of breeding process, air-flow and illuminance parameter
Ideal value carry out scientific offering, improve setting breeding process in chicken different growth phases to microclimate environment humiture,
Air-flow and the science of illuminance parameter requirements amount, improve the returns and efficiency of cultivation it is achieved that scientific culture and efficiently supporting
Grow.
6th, PID control, neutral net, PREDICTIVE CONTROL, uneoupled control, serials control and expert system are mutually tied by the present invention
Close, the hen house miniclimate breeding environment cascade control system of the PID neutral net based on DRNN for the design.This control system overcomes
Simple PID control is big to big inertia to postpone the shortcoming that object regulation quality is poor, anti-interference is weak, hen house microclimate environment parameter god
Effectively compensate for Classical forecast through network identifier model and control the limitation based on linear model.This control system is used for
The control of hen house microclimate environment humiture, air-flow and illuminance parameter has stronger performance of dynamic tracking and antijamming capability
And good dynamic and static state performance index.
Brief description
Fig. 1 is hen house microclimate environment detecting and control of parameter system construction drawing of the present invention;
Fig. 2 is the hen house miniclimate breeding environment cascade control system structure based on the PID neutral net of DRNN for the present invention
Figure;
Fig. 3 is DRNN neural network structure of the present invention and its approaches structure;
Fig. 4 is detector unit data acquisition software workflow diagram of the present invention;
Fig. 5 is on-site supervision end of the present invention software architecture diagram;
Fig. 6 is control system plane arrangement structure figure of the present invention.
Specific embodiment
(1) detection node design
Detector unit is by humiture, illuminance and pneumatic sensor, filter circuit, signal conditioning circuit, communication nRF2401
Module and MSP430 single-chip microcomputer are constituted.Hen house microclimate environment humiture is gathered using humiture digital sensor SHT11 and send
Enter the A/D conversion input circuit of single-chip microcomputer, radio communication adopts nRF2401 module to pass through self-organizing form by observing and controlling hen house structure
Build the information exchange of radio fest and control real-time performance and on-site supervision end, system adopts nRF2401 module to connect the RS232 of computer
Interface realizes on-site supervision end and the communication of hen house microclimate environment measuring and controlling node.Detection node structure (as shown in Figure 1), detection
Node realizes the detection to hen house microclimate environment parameter humiture, illuminance and air-flow, software communication flow process (Fig. 4).
(2) control node design
Control node is by Phototube Coupling output driving circuit, MSP430 single-chip microcomputer, 4 PI secondary controllers and communication
NRF2401 module composition, by detection node communication nRF2401 module, control node communication nRF2401 module, on-site supervision end
NRF2401 module, realize the communication of on-site supervision end and detection node and control node, Control system architecture and control node
Structure (as shown in Figure 1).In this node according to control external equipment to need the PI secondary controller designing 4 loops to realize right
Heater, the control of blower fan, wet curtain and light filling equipment, the wherein input Connection Neural Network of PI decouples the output of master selector,
Thus constituting many cascade control systems (as shown in Fig. 2 the latter half) of major-minor adjuster.
(3) on-site supervision end
The timing of on-site supervision end reads the hen house microclimate environment humiture from detection node collection, air-flow and light from serial ports
Illumination parameter, it can be shown in real time to the microclimate environment data of test point, curve shows, data storage and history are looked into
Ask (as shown in Figure 5).Monitoring client design humiture, air-flow and illuminance are controlled+nerve by the PID1-4 of control parameter at the scene
Network decoupling controller+correspondence is constituted the homophony of hen house microclimate environment by the neural network identifier NN1-NN4 of control parameter
Section device (as shown in Fig. 2 the latter half), this correspondence is realized by the neural network identifier NN of control parameter and its PID controller
PREDICTIVE CONTROL to hen house microclimate environment parameter, can be adjusted effectively due to the non-thread of hen house microclimate environment environment temperature
Property, the control accuracy that causes such as large time delay and time variation is low, the problem of low-response and stability difference.In order to solve between them
Close coupling problem, after PID regulator, the decoupling control between hen house microclimate environment realized by series connection neural network decoupling controller
System, improves hen house microclimate environment control accuracy and response speed.Improve hen house culture benefit and efficiency, it is to avoid chicken is artificially set
House microclimate environment parameter leads to the actual state of culture benefit and inefficiency, according to chicken different growth phases and culture benefit
Needs to microclimate environment humiture, air-flow and illuminance parameter, the Implementation of Expert System based on economic optimum reasoning for the design
Scientific offering (as shown in Fig. 2 top half) to hen house microclimate environment humiture, air-flow and illuminance parameter.Wherein:
Ith, PID controller design
Using Position Form PID controller, if riT () is respectively the specified rate of t hen house microclimate environment parameter, yi(t)
It is respectively the output actual value of t hen house microclimate environment parameter, the control error of corresponding parameter is:
errori(k (t))=ri(t)-yi(t) (i=1,2,3) (1)
If the sampling period is T, then corresponds to controller and be output as:
ui(t)=kpi(t)xi1(t)+kIi(t)xi2(t)+kdi(t)xi3(t) (i=1,2,3) (2)
Three term coefficient k of PID controllerpi(t)、kIi(t) and kdiT () is entered using corresponding parameter identification neutral net DRNN
Row is adjusted.
If corresponding by the evaluation function controlling microclimate environment parameter be:
Then the index variation amount of PID controller is respectively:
In formulaRepresent the spirit being controlled hen house microclimate environment factor reality output to correspond to PID controller output
Sensitivity function, it can be carried out identifier by corresponding hen house by neutral net DRNN of control parameter and obtain.
IIth, DRNN neural network identifier
DRNN diagonally recurrent neural network model (diagonal recurrent neural network) is that one kind has feedback
Dynamic Recurrent Neural Network, this network can more directly vivo reflection be identified the change of hen house microclimate environment dynamic state of parameters
Performance, and there is the ability adapting to time-varying characteristics.Fig. 4 is used for the nerve net of hen house microclimate environment PID controller parameter of adjusting
Network DRNN1-4 adopts the 3 layer network structures of 3-7-1, and its hidden layer is to return layer.
In the DRNN neutral net of design identification hen house microclimate environment humiture, air-flow and illuminance parameter, if I=
[I1(t),I2(t),…,In(t)], wherein I vectorial for network inputsiT () is hen house microclimate environment parameter identification network inputs
The input of i-th neuron t of layer, if parameter identification network inputs take I={ u (t-1), y (t), 1.0 }, j-th of recurrence layer
Neuron is output as xj(t), sjT () is j-th recurrent nerve unit input summation, f () is the function of S, then O (t) is DRNN
The output of network.Then the output layer of the hen house microclimate environment parameter of this network identification is output as:
The layer that returns of the hen house microclimate environment parameter of this network identification is output as:
xj(t)=f (Sj(t)) (5)
The hen house microclimate environment parameter of this network identification recurrence layer input be:
The approximate error of the hen house microclimate environment parameter of this network identification is:
em(t)=y (t)-ym(t) (7)
The identifying index of this hen house microclimate environment parameter of this network identification takes:
If this identification network obtains the desired output of this hen house microclimate environment humiture, air-flow and illuminance parameter,
This identification network learning algorithm terminates;Otherwise, this identification network goes to backpropagation, by the mistake of this hen house microclimate environment parameter
Difference signal presses connecting path backwards calculation, is adjusted weights and the threshold values of this identification network each layer neuron by gradient descent method, makes
The parameter error of this network identification reduces.Of the identification network of hen house microclimate environment humiture, air-flow and illuminance parameter
Practise algorithm and adopt gradient descent method, wherein, the gradient descent method study returning layer to output layer connection weight of this identification network is calculated
Method is:
The input layer of this identification network to return layer connection weight gradient descent method learning algorithm be:
This identification network recurrence layer connection weight gradient descent method learning algorithm be:
If the recurrence layer neuron of identification network takes double S function to be:
And approx takeThen parameter P in this identification networkj(t) and QijT the value of () is:
η in formulao、ηDAnd ηIIt is respectively the learning rate of input layer, recurrence layer and output layer, α is inertia coeffeicent.Then should
The Jacobian information of this hen house microclimate environment humiture, air-flow and illuminance parameter of network identification is:
IIIth, neural network decoupling
The neural network decoupling controller architecture of the hen house microclimate environment factor is as shown in Fig. 2 wherein ri(t) and yi(t) point
Do not represent the desired output and reality output being controlled environment of chicken house parameter, controlled environment of chicken house factor desired output and reality
The minimum evaluation function as this control system performance of the quadratic sum of the difference of output, that is,:
Evaluation function E (t) is the function of decoupling neutral net weight coefficient as can be seen from above, by on-line tuning with certainly
The connection weight of study decoupling neutral net can make E (t) minimum, realizes the purpose of environment of chicken house factor Multivariable Decoupling Control,
Improve the control performance of the environment of chicken house factor.According to gradient descent method, decouple nerve network controller link weight coefficients in t+1
The iterative formula in moment is:
The learning rate that in formula, η weighs for uneoupled control network connection, wherein Represent j-th input of hen house microclimate environment factor control systemAct on during control target with respect to defeated
Go out yiT the rate of change of (), it represents the influence degree of the Dynamic Coupling to i-th output for j-th input, and that is, this is controlled system
The dynamical couple characteristic of system.According to neutral net to the identification being controlled hen house, available following formula calculates the Dynamic Coupling of this system
Characteristic, that is,:, in formula, sgn () is sign function.
IVth, Expert System Design
This expert system mainly sets hen house microclimate environment humiture, air-flow and illuminance according to Economic Benefit Model
The ideal parameters value of parameter, main basis:The cost model of microclimate environment humiture, air-flow and illuminance state modulator, chicken
The parameter such as feed-weight ratio and chicken growth phase under growth model, different microclimate ambient parameter, by the economic benefit of output/consumption
Model obtains the economic optimum micro-climate parameter of present period chicken growth, sets hen house miniclimate by the reasoning of expert system
Ambient temperature and humidity, air-flow and illuminance parameter control system desired value, this value is as the setting value of control system.This management software
Have selected Microsoft Visual++6.0 as developing instrument, the reasoning process of expert system is shown in the top half of Fig. 2.
(4) design example of hen house microclimate environment parameter control system
According to the relative position of hen house, system arranges detection node and control node, on-site supervision end, and whole system is put down
Face layout drawing is shown in Fig. 6.
Technological means disclosed in the present invention program is not limited only to the technological means disclosed in above-mentioned embodiment, also includes
By the formed technical scheme of above technical characteristic any combination.It should be pointed out that for those skilled in the art
For, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications are also considered as
Protection scope of the present invention.