CN115563890A

CN115563890A - Environment monitoring sensor deployment method and experiment platform based on digital twins

Info

Publication number: CN115563890A
Application number: CN202211566201.9A
Authority: CN
Inventors: 徐亮; 陈龙; 其他发明人请求不公开姓名
Original assignee: Hubei Xiecheng Transportation Environmental Protection Co ltd
Current assignee: Hubei Xiecheng Transportation Environmental Protection Co ltd
Priority date: 2022-12-07
Filing date: 2022-12-07
Publication date: 2023-01-03
Also published as: CN115859838B; CN115859838A

Abstract

The invention relates to a digital twin-based environment monitoring sensor deployment method and an experimental platform, which are oriented to the deployment requirements of environment monitoring sensors under complex scenes of roads, bridges, rivers, lakes and the like, and decide to generate an environment monitoring sensor optimized deployment scheme with optimal cost, performance and service life, and mainly comprise the steps of constructing a performance evaluation model of the environment monitoring sensor optimized deployment scheme from multiple angles such as a monitoring value range, a signal transmission distance, resolution and the like; taking cost, performance and service life as optimization targets, and solving an optimal deployment scheme set of the environmental monitoring sensor in a complex scene by utilizing a multi-objective optimization genetic algorithm; and combining a Unity3D development engine and Matlab software to realize data interaction between the optimal deployment scheme of the environmental monitoring sensor and the digital twin model scene. The method can solve the technical problem of optimal deployment of the environment monitoring sensor, has the advantages of low calculation cost and high calculation efficiency, and has important practical significance.

Description

Environment monitoring sensor deployment method and experiment platform based on digital twins

Technical Field

The invention relates to the technical field of ecological environment monitoring, in particular to a digital twin-based environment monitoring sensor deployment method and an experimental platform.

Background

In the field of ecological environment monitoring, environmental data such as water, air, soil and the like are monitored, data support can be provided for confirming pollution sources and formulating corresponding solutions, and the method has important significance for environmental protection. However, when the number of deployed nodes of the environmental monitoring sensors increases, the number of the environmental monitoring sensors selectable by each node is hundreds of thousands, so that the number of alternatives for deployment of the environmental monitoring sensors in different environmental scenes increases exponentially, and how to select an optimal solution for deployment of the environmental monitoring sensors is an urgent problem to be solved, so as to provide reference for environmental monitoring enterprises.

In the prior art, a traditional traversal method is adopted to solve the optimal solution of the deployment of the environmental monitoring sensor. The traversal method refers to sequentially making an access to each node in the structure tree along a certain search route, wherein the operation performed by accessing the node depends on a specific application problem, and the specific access operation may be to check the value of the node, update the value of the node, and the like. Different traversal methods have different access node orders. The traversal method is also applicable to the case of multi-element sets, such as arrays.

However, the method for solving the optimal solution by adopting the traditional traversal method has the defects of great difficulty, high calculation cost and low calculation efficiency.

Disclosure of Invention

The invention aims to provide a deployment method and an experimental platform of an environment monitoring sensor based on digital twins, which are used for three-dimensional modeling based on a digital twins technology, can visually present an optimization process and an optimization result, can solve the technical problem of optimized deployment of the environment monitoring sensor, and have the advantages of low calculation cost and high calculation efficiency, thereby having important practical significance.

In order to achieve the purpose, the invention designs an environment monitoring sensor deployment method based on digital twins, which is characterized by comprising the following steps:

step 1: according to the environment monitoring requirement, arranging deployment nodes of an environment monitoring sensor in an environment monitoring scene to form a block-shaped ecological area comprising a road, a bridge and a river;

step 2: determining the type and the number of environment monitoring sensors used in the block-shaped ecological area;

and step 3: determining evaluation indexes of each environmental monitoring sensor in the massive ecological area, wherein the evaluation indexes comprise performance, cost and service life;

and 4, step 4: generating selectable individuals of each environment monitoring sensor according to the evaluation indexes, performing real number coding on the selectable individuals, and constructing gene codes deployed by the environment monitoring sensors by using a real number coding method;

and 5: according to different types and quantities of environment monitoring sensors, an initial population of the multi-objective optimization genetic algorithm is constructed through a random number generation method;

and 6: calculating the non-dominated sorting and the crowding degree of the initial population to obtain a population matrix containing the sorting level and the crowding degree, and evaluating the advantages and the disadvantages of the environment monitoring sensor optimized deployment scheme with a plurality of evaluation indexes;

and 7: combining competitive bidding competition selection and elite strategies, constructing a crossover operator of the genetic algorithm by adopting a random number threshold method, constructing a mutation operator of the genetic algorithm through random number updating operation, and solving an optimal set of environment monitoring sensor optimized deployment in a complex scene by a population circulation iterative updating strategy;

and 8: and constructing an environment monitoring sensor optimized deployment experimental platform based on a Unity virtual development engine and a Matlab genetic algorithm programming environment, importing an optimal set of environment monitoring sensor optimized deployment in the complex scene into a simulated environment monitoring scene of a block ecological area comprising roads, bridges and rivers, and dynamically presenting an environment monitoring sensor optimized deployment effect.

Preferably, in the step 2, the method for determining the type of the environmental monitoring sensor used in the block ecological region includes selecting an air temperature sensor, a road pressure sensor and a rainfall sensor on a road, selecting a road pressure sensor and a rainfall sensor on a bridge, and selecting a river flow rate sensor in a river.

Preferably, in step 3, the performance evaluation index in the evaluation indexes specifically includes a monitoring value range, a signal transmission distance, and a resolution of the sensor.

Preferably, different evaluation indexes are normalized by a quantitative normalization method based on the isomerism of the evaluation indexes.

As a preferred scheme, in step 5, for different types of environmental monitoring sensors, one of 1000 selectable individuals of each environmental monitoring sensor is randomly selected, an initial population of the multi-objective optimization genetic algorithm is constructed, and parameters of each individual in the initial population are randomly generated; the random parameter generating method includes randomly generating an appointed number of integers in a certain range and filling the integers into a fixed-length array; the resolution parameter in the performance indicator in the sensor evaluation indicator is randomly generated by selecting among several already determined numbers.

As a preferred scheme, the evaluation index is used as an objective function in the initial population, the real number code of the selectable individual of each environment monitoring sensor is used as a decision variable, when the population is initialized, the decision variable and the objective function are connected in series to form a matrix for convenience of calculation, the decision variable value and the objective function value corresponding to each decision variable are respectively stored at the corresponding positions of a chromosome, the real number code is adopted in the gene coding process, the node number of one sensor deployment scheme forms a gene sequence of a feasible solution, and different sensor deployment schemes correspond to different feasible solutions, namely different decision variables.

As a preferred scheme, after an optimal set of optimal deployment of the environmental monitoring sensors in the complex scene is obtained in step 7, a scheme with the maximum selective energy parameter in the corresponding optimal solution set is used as an optimization result to be visualized for the air temperature sensor and the road pressure sensor; and for the rainfall sensor and the river flow velocity sensor, selecting a scheme with the largest cost parameter in the corresponding optimal solution set as an optimization result to be visualized.

As a preferred scheme, in step 8, the TXT text file is used as a middleware to realize data communication between the Unity virtual development engine and the Matlab genetic algorithm programming environment, an optimal deployment scheme is selected in the Matlab genetic algorithm programming environment, and corresponding data is imported into an external designated TXT text file; writing a script in a text component corresponding to the Unity virtual development engine to read the TXT text file, and displaying data; when the parameters of the optimal deployment scheme are changed, the optimization result is changed accordingly and is dynamically updated to the visualization platform for display.

The invention also designs an environmental monitoring sensor deployment experimental platform based on the digital twin, which is characterized in that the experimental platform is used for realizing the environmental monitoring sensor deployment method based on the digital twin.

As a preferred scheme, the environment monitoring sensor deployment experiment platform based on the digital twin is built based on a Unity virtual development engine and Matlab genetic algorithm programming environment, and a TXT text file is used as a middleware to achieve data communication.

Compared with the prior art, the invention aims at the problem of combined explosion of the deployment schemes of the environmental monitoring sensors in complex scenes, adopts the coding, local search and global search strategies of a multi-objective optimization genetic algorithm based on the type selection requirement of the environmental monitoring sensors, analyzes and generates the optimal deployment scheme of the environmental monitoring sensors in a circular iteration mode, sufficiently utilizes theories and technologies such as digital twinning, group intelligent optimization and the like, provides the optimal deployment method of the environmental monitoring sensors driven by a digital twinning model, generates the optimal deployment scheme set of the environmental monitoring sensors, carries out three-dimensional modeling based on the digital twinning technology, can visually present the optimization process and the optimization results, and deeply understands the optimal deployment scheme of the sensors from multiple dimensions. The method combines the multi-objective optimization genetic algorithm with the 3D virtual model to simulate the optimal deployment of the environmental monitoring sensor in actual production, and has important practical significance for optimizing deployment capability of the environmental monitoring enterprise sensor and reducing deployment cost of a monitoring network.

Drawings

FIG. 1 is a flow chart of the present invention;

FIG. 2 is a schematic diagram of an environmental monitoring sensor infrastructure arrangement according to the present invention;

FIG. 3 is an optimal deployment scenario of the environmental monitoring sensor according to the present invention;

FIG. 4 is a flow chart of the gene encoding process of the present invention;

FIG. 5 is a flow chart of a competitive bidding competition selection method of the present invention;

FIG. 6a is a graph of the running time versus the number of iterations for a population number of 20;

FIG. 6b is a graph of the relationship between the operation time and the number of iterations for a population number of 30;

FIG. 6c is a graph of run time versus iteration number for a population number of 40;

FIG. 7a is a graph of the relationship between the generation distance and the number of iterations for a population number of 20;

FIG. 7b is a graph of the relationship between the generation distance and the number of iterations for a population number of 30;

FIG. 7c is a graph of the relationship between the generation distance and the number of iterations for a population number of 40;

FIG. 8a is an optimization of air temperature sensor deployment over 50 iterations with a 200 population;

FIG. 8b is the optimization of the air temperature sensor deployment over 100 iterations with a cluster number of 200;

FIG. 8c is the optimization of air temperature sensor deployment over 150 iterations with a 200 population;

fig. 9 is a flowchart of visualization experiment platform deployment according to the present invention.

Detailed Description

In order to make the technical problems solved, the technical solutions adopted and the technical effects achieved by the present invention clearer, the technical solutions of the present invention are further described below by way of specific embodiments with reference to the accompanying drawings. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some but not all of the elements associated with the present invention are shown in the drawings.

In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. Wherein the terms "first position" and "second position" are two different positions.

In the description of the present invention, it should be noted that unless otherwise explicitly stated or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection or a removable connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.

The digital twin concept is derived from digital full-life-cycle management in the field of industrial manufacturing, is mainly applied to links such as product design, performance simulation, assembly process, measurement and inspection and the like, and can improve the research and development efficiency and reduce the research and development cost.

The invention discloses an environment monitoring sensor optimal deployment method based on a multi-objective optimization genetic algorithm, and a digital twin model modeling method based on Unity3D is used for visually presenting an optimization scheme, wherein the basic process is shown in figure 1.

The invention relates to a digital twin-based environmental monitoring sensor deployment method, which comprises the following steps:

step 1: dividing environment monitoring scenes into roads, bridges and rivers according to environment monitoring requirements; setting a deployment node of an environmental monitoring sensor in each environmental monitoring scene; forming a block ecological region comprising 'road-bridge-river';

step 2: determining the type and the number of environment monitoring sensors used in the block ecological area;

after the functions and the deployment sites of various sensors are comprehensively considered, 100 four sensors are determined to be deployed at the sites such as roads, bridges, rivers and the like according to the actual requirements of enterprises. The method for determining the type of the environment monitoring sensor used in the block ecological region comprises the steps of selecting an air temperature sensor, a road pressure sensor and a rainfall sensor on a road, selecting a road pressure sensor and a rainfall sensor on a bridge and selecting a river flow velocity sensor in a river.

The number of the environment monitoring sensors used in each environment monitoring scene is 60 on the road, wherein 30 air temperature sensors, 10 road pressure sensors and 20 rainfall sensors are arranged on the road; the number of the sensors on the bridge is 20, wherein 10 road pressure sensors and 10 rainfall sensors are arranged on the bridge; and 20 river sensors are used in the river. As shown in table 1 and figure 2 for details,

TABLE 1 number of environmental monitoring sensors used in environmental monitoring scenarios

in order to determine an optional population for the initial population input to the optimization algorithm, evaluation indexes of the sensors need to be defined, and then the optional population can be generated according to defined criteria.

The evaluation indexes of the sensor can be divided into two main categories of main performance indexes and other indexes. The main performance indexes are used for judging the quality of the sensor in the aspect of service performance, and comprise a monitoring value range, a signal transmission distance, resolution, linearity, sensitivity, drift, a threshold value and the like; other indexes are specifically analyzed according to specific use scenes, and the cost, the aging degree, the service life and the like are common.

Based on the basic deployment scheme of the sensor, on the premise of meeting research requirements according to the self condition of an enterprise and the application scene of the environment monitoring sensor, five evaluation indexes are selected to evaluate the deployed sensor. Which are respectively as follows: monitoring value range, signal transmission distance, resolution, service life and cost. Meanwhile, in order to simplify the optimized deployment process of the sensor, dimension unification and linear weighted summation are carried out on three main performance indexes of a monitoring value range, a signal transmission distance and resolution, and the three main performance indexes are comprehensively represented as evaluation indexes 'performance'. And based on the isomerism of the evaluation indexes, carrying out normalization processing on different evaluation indexes by a quantitative normalization method.

Set the monitoring value range asX1A signal transmission distance ofX2Resolution ofX3. Then performance is improvedY1As follows:

and then, the evaluation indexes of the sensor optimization deployment problem are as follows: performance, age, cost. The result of the multiobjective optimization is therefore to maximize the performance and lifetime of the overall solution, while at the same time minimizing costs. In order to ensure the consistency of the optimization goal, the cost needs to be subjected to parameter conversion.

Set the cost asX4Then cost parameterY2Expressed as:Y2=MAX+MIN-X4

in the above formulaMAXAndMINthe maximum value and the minimum value of the corresponding parameter are pointed.

Thus, the optimization objectives of the multi-objective optimization problem of the present invention are performance, age and cost.

Based on the method, an evaluation model of the environment monitoring sensor optimized deployment scheme is constructed.

in the step 4, 1000 selectable individuals of each environment monitoring sensor are set, and real number encoding is performed on the 1000 selectable individuals from 0 to 1000.

in each environment monitoring scene, aiming at different types of environment monitoring sensors, randomly selecting one from 1000 selectable individuals of each environment monitoring sensor, constructing an initial population of a multi-objective optimization genetic algorithm, and randomly generating parameters of each individual in the initial population; the random parameter generating method includes randomly generating an appointed number of integers in a certain range and filling the integers into a fixed-length array; the resolution parameter in the performance indicator in the sensor evaluation indicator is randomly generated by selecting among several already determined numbers.

The evaluation index is used as a target function in the initial population, the real number code of the selectable individual of each environment monitoring sensor is used as a decision variable, in order to facilitate calculation, the decision variable and the target function are connected in series to form a matrix, the decision variable value and the target function value corresponding to each decision variable are respectively stored at the corresponding positions of chromosomes, the real number code is adopted in the gene coding process, the node number of one sensor deployment scheme forms a gene sequence of feasible solutions, and different sensor deployment schemes correspond to different feasible solutions, namely correspond to different decision variables.

Thereafter, in terms of a multiobjective optimization genetic algorithm based on Pareto (Pareto) optimal solution and crowding distance:

an evaluation model based on an environment monitoring sensor optimized deployment scheme needs to generate an optional population, and performs gene coding and initialization on the initial population, and the basic process is as follows:

first, an initial alternate population needs to be randomly generated, containing 1000 alternate individuals for each sensor, the 1000 individuals being customized in the database.

And (3) writing a program by using MATLAB, randomly generating parameters of each individual, and writing the parameters into corresponding cells of the Excel file. The random parameter generation comprises two program segments, wherein one program segment is used for randomly generating an appointed number of integers in a certain range and filling the integers into a fixed-length array; alternatively, each datum can only be selected among several already determined numbers, and the resolution in the performance indicator in the sensor evaluation indicator needs to be generated in this way.

And randomly generating simulation data of the four sensors, and filling the generated data into four different sheets of the Excel table respectively. In the present invention, there are 100 environmental monitoring sensors in total, so the initial selectable population has 100 ten thousand individuals containing four sensors in total.

Secondly, the gene coding and initialization process of the population;

in both cases, the gene coding and the gene decoding represent a solution candidate of a problem by a chromosome, and mapping of a decoding space to a coding space is realized. There are many encoding methods, such as binary encoding, real vector encoding, integer permutation encoding, and universal data structure encoding.

Wherein, the decision variable refers to the serial number (a certain number between 1 and 1000) of the sensor selected by the node; the objective function refers to three evaluation indexes of the environment monitoring sensor. In the algorithm design, the gene coding process adopts real number coding. Specifically, the number of nodes in one sensor deployment scenario may form a feasible gene sequence, and different sensor deployment scenarios correspond to different feasible solutions, that is, different decision variables, as shown in fig. 4 in detail.

The process of population initialization is to randomly generate a matrix consisting of population individuals. All decision variables of each individual are randomly valued within the constraint.

In addition, for convenience of calculation, the decision variables and the objective function are connected in series to form a matrix during population initialization. If there are V decision variables and M objective functions, randomly generated decision variable values are stored at positions from 1 to V of the chromosome, and corresponding objective function values are stored at positions from V +1 to V + M of the chromosome.

Step 6: calculating the non-dominated sorting and the crowding degree of the initial population to obtain a population matrix containing the sorting level and the crowding degree, and evaluating the advantages and the disadvantages of the environment monitoring sensor optimized deployment scheme with a plurality of evaluation indexes;

after the initialization of the population is completed, the non-dominated sorting and congestion degree calculation needs to be carried out on the initial population:

according to the Pareto optimal solution theory, when the solution A is superior to the solution B, the solution A dominates the solution B; when neither A nor B are mutually supporting, they constitute a pair of non-dominated solutions. The fast non-dominant ranking process is a process of ranking decision vectors according to a dominant relationship. When A dominates B, the non-dominated ranking value of A is higher than that of B.

Meanwhile, in order to obtain the crowdedness of a specific solution in the population, the average distance between points on both sides of the point needs to be calculated according to each objective function. This value is used as an estimate of the perimeter of the cuboid with the nearest neighbors as vertices (called the crowding factor).

Thereafter, each individual calculates a corresponding non-dominant ranking value and congestion distance. Meanwhile, the non-dominant ranking value and the crowding distance are also added to the chromosome matrix for the convenience of subsequent processing. Thus, a population matrix including the ranking level and the crowdedness can be obtained, and the matrix is already ranked according to the ranking level.

And 7: combining competitive bidding competition selection and elite strategies, adopting a random number threshold method to construct a crossover operator of the genetic algorithm, constructing a mutation operator of the genetic algorithm through random number updating operation, and solving an optimal set of environment monitoring sensor optimized deployment in a complex scene according to a population circulation iterative updating strategy.

After an optimal set of optimal deployment of the environmental monitoring sensors in a complex scene is obtained, for the air temperature sensor and the road pressure sensor, a scheme with the maximum selective energy parameter in a corresponding Pareto optimal solution set is used as an optimization result needing visualization; and for the rainfall sensor and the river flow velocity sensor, selecting a scheme with the largest cost parameter in the corresponding Pareto optimal solution set as an optimization result to be visualized.

And 8: and constructing an environment monitoring sensor optimized deployment experimental platform based on a Unity virtual development engine and a Matlab genetic algorithm programming environment, importing an optimal set of environment monitoring sensor optimized deployment in the complex scene into a simulated environment monitoring scene of a block ecological area comprising a road, a bridge and a river, and dynamically presenting an environment monitoring sensor optimized deployment effect.

The data communication between the Unity virtual development engine and the Matlab genetic algorithm programming environment is realized by using the TXT text file as the middleware: selecting an optimal deployment scheme in a Matlab genetic algorithm programming environment, and importing corresponding data into an external specified TXT text file; writing a script in a text component corresponding to the Unity virtual development engine to read the TXT text file, and displaying data; when the parameters of the optimal deployment scheme change, the optimization result changes along with the change of the parameters, and the optimization result is dynamically updated to the visualization platform for displaying.

After the initial population is sorted in a non-dominant mode and the crowding degree is calculated, the initial population is selected, crossed and mutated through a genetic algorithm (including selection, crossing and mutation), and the basic process is as follows:

wherein, the selection algorithm uses competitive competition selection; the cross algorithm selects analog binary cross; the mutation algorithm selects polynomial mutation. These processes will be explained separately below.

Competitive competition selection and elite strategy:

the core thought is as follows: two individuals are randomly selected each time, the individual with the high non-dominant ranking value is preferentially selected, and if the ranking levels are the same, the individual with the high crowdedness is preferentially selected. The basic flow chart is shown in fig. 5.

The competitive bidding competition selection in the genetic algorithm can be divided into the following steps:

(1) randomly selecting different competition individuals;

(2) recording the sequencing level and the crowdedness of each competition individual;

(3) selecting a competition individual with a smaller ranking grade value, and returning the index of the competition individual by using find; smaller ranking grade values represent higher ranking grades;

(4) if the ranking levels of the two participating individuals are equal, the crowdedness is continuously compared, and the individual with the larger crowdedness is selected.

Through competitive bidding competition selection, an excellent initial parent population can be obtained, and then the parent population is subjected to cross and variation operations to obtain a child population.

After the crossover and mutation algorithm operations are completed, a new population comprising a parent population and a child population can be obtained. Thereafter, a new elite population is generated therefrom for the next iteration using the elite strategy.

The method for iterating the population by using the elite strategy is the biggest difference between the NSGA-II algorithm and the NSGA algorithm, and comprises the following specific processes:

firstly, combining the filial generation population Q produced by t generation with the parent generation population P to form R _t ，R _t The population size was 2N. Then to R _t Performing fast non-dominant sorting to generate a series of non-dominant sets Z _i And calculates the degree of congestion. Because the individuals of the offspring population Q and the parent population P are contained in R _t In (3), a series of non-dominating sets Z after non-dominating sorting _i All of the individuals contained in (A) are R _t The best of them; firstly, a first-stage non-dominating set Z is set ₁ Put into a new parent population P _t+1 In case of a new parent population P _t+1 If the population size of (2) is less than N, continuing to P _t+1 Middle filling next level non-dominating set Z ₂ Until addition of Z _i Chrononew parent population P _t+1 When the population size of Z is larger than N, stopping _i The individuals in (1) are compared and selected for crowdedness, so that P _t+1 Then generating a new filial generation population Q through a genetic algorithm (selection, crossing and mutation) _t+1 。

Crossing and mutation:

the decimal of 0~1 is randomly chosen using the rand function. When the selected natural number is smaller than the crossover probability, implementing crossover operation; and when the selected natural number is smaller than the mutation probability, performing mutation operation. In addition, there are two issues that need to be addressed when implementing analog binary interleaving and polynomial diversity:

(1) when new filial generation individuals are generated through crossing and variation, rounding operation needs to be carried out on a calculation result so as to ensure that data in a population are positive integers:

(2) when random mutation operation is carried out on data, the constraint condition is added, so that the data beyond the data limit range is forced to be converted into an upper limit or a lower limit of the constraint.

Finally, in the aspect of the method for creating the digital twin model and integrating the system in the environment monitoring scene:

after the Pareto optimal solution set of the sensor deployment schemes is obtained, each deployment scheme contained in the Pareto optimal solution set is independent of each other. However, when the optimization result is visualized, only one complete deployment scheme needs to be imported. The optimal solution needs to be selected.

Among the three evaluation indexes, the service life is the index with the smallest relative influence, and the performance parameters and the cost parameters of the other two indexes have different emphasis requirements under different conditions and need to be specifically considered.

For an air temperature sensor and a road pressure sensor, the performance of the sensor is more strong and stable, so that the scheme with the maximum selective energy parameter in the corresponding Pareto solution set is used as an optimization result to be visualized; for the rainfall sensor and the river flow rate sensor, the cost control is more important than the excellence of the performance, so the most visual optimization result is needed for selecting the scheme with the largest cost parameter in the corresponding Pareto solution set, and the basic process is shown in fig. 3.

And combining the selected schemes of the four sensors to obtain a deployment result of the current environment monitoring sensor optimized deployment, and importing the corresponding data into the Unity corresponding model to realize the visualization of the optimization result.

In order to realize the construction of an environment monitoring sensor optimized deployment experiment platform, namely the real-time visualization of the deployment scheme optimization result, data connection needs to be carried out on Matlab and Unity.

The invention also relates to an environmental monitoring sensor deployment experiment platform based on the digital twins, and the experiment platform is used for realizing the environmental monitoring sensor deployment method based on the digital twins.

The environment monitoring sensor deployment experiment platform based on the digital twin is built based on a Unity virtual development engine and a Matlab genetic algorithm programming environment, and a TXT text file is used as a middleware to realize data communication.

The construction of the environment monitoring sensor deployment experiment platform based on the digital twins comprises the following steps:

sequentially constructing virtual models of roads, bridges and rivers through three steps of standard resource packet import, material coverage and terrain drawing, and constructing a digital twin model of an environment monitoring scene through a virtual assembly form;

marking deployment nodes of the environment monitoring sensors in each environment monitoring scene, performing simulation modeling on the deployment nodes of the environment monitoring sensors, and importing an optimization scheme corresponding to external data to realize the construction, test and analysis of a visual experiment platform;

considering the heterogeneity of data, according to the data interaction characteristics of a group intelligent optimization algorithm and a digital twin model development environment, utilizing a TXT text as a data interaction middleware to realize effective transmission of the labels of selectable individuals of each environment monitoring sensor and the deployment scheme of the environment monitoring sensors, and integrating a Unity3D virtual engine and a Matlab computing environment;

when the parameters of the optimization algorithm are changed, the optimization result is changed, and the change can be reflected in the visual experiment platform in time.

The present invention uses the TXT text file as an intermediate file to implement data communication, and the basic flow chart is shown in fig. 9.

The whole thought is as follows: selecting an optimal deployment scheme in Matlab, and importing corresponding data into an external specified TXT text file; and then writing a script in the Unity corresponding text component to read the TXT text file and displaying the data.

In order to realize the visualization of the optimal deployment scheme of the sensor, firstly, the modeling of the basic terrain needs to be carried out in Unity, namely, a small ecological region containing 'road-bridge-river' is visually presented.

In order to complete the basic modeling, a Standard Assets package (Standard Assets package) needs to be imported first, and after the Standard Assets package is imported, enough basic Assets are available to support the construction of the basic terrain.

First, a terrain is created and the relief of the terrain is initially mapped. And then, the terrains need to be covered by materials, the terrains are generally covered by green grassland materials, and the mountains are covered by brown stone materials. Then, a part of the landform is reduced to be used as a river channel, so that the river can be conveniently manufactured later. After the work is finished, the construction of the terrain is finished, and a foundation is laid for the following manufacture of roads, bridges and rivers and the introduction of an optimization result.

For roads and bridges, the method selects the Road Architect plug-in to construct the three-dimensional visual model of the roads and bridges.

After the Road Architect resource package is introduced, a Road system needs to be newly built under a Window drop-down column. The Road system is then selected, and "Add Road" is selected at the component panel, creating a Road component. And clicking and selecting the Road1, and then pressing the shift and left mouse button to quickly create a Road model on the terrain.

For bridges, only material correction needs to be carried out on the side faces of roads generated in a river channel area. The material used here is aluminum from Road Architect. After the material is changed and the subsequent river covering is carried out, the effect of the bridge can be simulated on the part of the road.

This section can be described in two sub-sections. The first part is the formation of the water surface effect, and the second part is the manufacture of the underwater effect.

First, it is the formation of the surface effect, which is relatively simple. The introduction of standard resource packages has been described previously. The environment resource package can be found in the Standard resource package (Standard Assets package) that has been imported. And (4) clicking an open environment resource package, searching for water, and then seeing the preset (Prefabs) of the water resource.

The resource package has several presets of water resources, including static water, dynamic water, day water, night water, etc. As used herein, preset is "dynamic water resources during the day" (waters basic daytime). And splicing the water surface effect, and finally covering the whole river channel, thereby finishing the manufacture of the river water surface effect.

The main idea for making the underwater effect is that the scene main camera has a special view shade when entering underwater. Corresponding to this idea, there are several items that need to be completed:

(1) compiling a corresponding script file and importing the script file into a newly-built transparent square block;

(2) adjusting the position of the transparent square block to enable the transparent square block to just cover the whole river;

(3) and importing the configuration file required to be used into the scene main camera.

First, a script file underwater. Cs is written. The core idea is as follows:

(1) when the camera enters water, turning on the Fog and Blur components, and setting the color to light blue;

(2) when the camera is out of the water, the Fog and Blur components are turned off:

after the script UnderWater.cs is written, the script file is added into the established transparent square model, namely the istriggerCube. And then adding a own Blur script in the environment resource package into the main camera of the scene. Parameters are adjusted, and projects are operated, so that the underwater effect can be observed to be successfully manufactured.

So far, the water surface effect and the underwater effect are manufactured, and the model of the river is basically realized. In conclusion, an environment monitoring scene comprising 'road-bridge-river' is obtained.

And partial sensor deployment points are marked by using round points, so that the observation is convenient. And the test and analysis of the visual platform are realized only by performing simulation modeling on the sensor deployment node and importing a corresponding optimization scheme.

After the 3D modeling of the environment monitoring scene is realized, external data needs to be imported and displayed, and the building, testing and analysis of a visual platform are realized.

Before external data is imported, a process of performing simulation modeling on the deployment nodes of the environmental monitoring sensors is specifically to establish 100 white small blocks to simulate the deployment nodes of the 100 environmental monitoring sensors, and import a text plug-in on the white small blocks to record labels of the environmental monitoring sensors deployed on the nodes.

Importing external data into the label of the environment monitoring sensor and the evaluation index of the optimization scheme; for the import of the environmental monitoring sensor label, a txt text is mainly used as a middleware for importing;

the Text plug-in Text (TMP) of each block (sensor deployment node) is added with a corresponding script file, so that the Text content of the external file "sensor label. Txt" can be read into Unity in a row and displayed in the corresponding Text box.

It should be mentioned that, in order to ensure the convenience of searching text content, a Streaming Assets folder needs to be newly created under the Unity project Assets folder, and several text files storing the optimization results of the sensor deployment scheme are dragged into the Streaming Assets folder. The text file is named "sensor number + index of corresponding category sensor".

The numbers represent the sensor numbers deployed on the nodes. The deployment of different sensors is marked with fonts of different colors. The red font represents that an air temperature sensor is deployed, the yellow font represents that a road pressure sensor is deployed, the blue font represents that a rainfall sensor is deployed, and the green font represents that a river flow rate sensor is deployed. At this point, the sensor label is successfully imported and displayed normally.

In order to observe the optimization result of the sensor deployment scheme completely and intuitively, besides the import of the sensor label, the sensor evaluation indexes (performance, cost and service life) of the currently selected optimal scheme also need to be displayed intuitively.

The method used here is similar to the method used when introducing the sensor label, and therefore, the description thereof is omitted. In contrast, it is assumed that the evaluation index should be displayed on the uppermost UI interface, and thus a script file for automatically inserting text should be added to a newly created UI component, not a text component attached to an object.

When the evaluation index of the current scheme is displayed, F1 represents the sum of the service life of the current deployment sensor, F2 represents the sum of the performance parameters of the current deployment sensor, and F3 represents the sum of the cost parameters of the current deployment sensor.

When the parameters of the optimization algorithm are changed, the optimization result is changed, and the change can be reflected in the visualization platform in time. Through comparison, the display result is found to be consistent with the actual calculation result, which shows that the result change caused by the parameter change can be timely updated to the visualization platform for display.

Example (b):

in this embodiment, parameters such as the number of populations, the number of iterations, the cross rate, and the variation rate need to be set, and specific parameter values are shown in table 2.

TABLE 2 set parameters for optimization algorithm

Number of population	Corresponding number of iterations	Rate of variation	Cross rate
				20	25、50、75、100	0.1	0.9
30	25、50、75、100	0.1	0.9
				40	25、50、75、100	0.1	0.9
200	50、100、150	0.1	0.9

In the aspect of parameter selection, experiments are carried out under the conditions that the population size is 20, 30 and 40 and the iteration times are 25, 50, 75 and 100, meanwhile, the population number is kept at 200, the iteration times are changed, and different optimization results are observed; each set of experiments was repeated five times and boxed to observe the effect of parameter changes on algorithm run time and performance.

The presentation of the optimization results is divided into two parts. One part is a three-dimensional representation drawn by three evaluation index values of each scheme in the Pareto solution set, three coordinate axes of the three-dimensional representation respectively correspond to the service life (F1), the performance parameter (F2) and the cost parameter (F3) in the evaluation index, and the other part is a data set of all the schemes in the Pareto solution set.

In the relationship between the influence of the parameters on the running time of the algorithm, when the number of the population is 20, 30, 40, the running time of the algorithm increases as the number of iterations increases, and the influence of the number of iterations on the running time of the algorithm can be summarized as shown in fig. 6a, fig. 6b, and fig. 6c, respectively.

And analyzing the operation performance of the algorithm under different iteration times by using a Generation Distance (GD) index in the influence relation of the parameters on the performance of the algorithm. Generally, a smaller generation distance value indicates a higher quality solution, as shown in fig. 7a, 7b, and 7 c.

When the population number is 200 and the iteration times are 50, 100 and 150 respectively, the optimization results of the deployment schemes of the air temperature sensors are shown in fig. 8a, 8b and 8c respectively, and blue points in the diagrams represent the air temperature sensors.

Finally, based on the Unity virtual development engine and the Matlab genetic algorithm programming environment, when the parameters of the optimization algorithm change, the optimal solution obtained by the genetic algorithm can be dynamically presented in the digital twin model, as shown in fig. 9.

The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present invention. It should be noted that various changes and modifications can be made by those skilled in the art without departing from the spirit of the invention, and these changes and modifications are all within the scope of the invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A deployment method of an environment monitoring sensor based on digital twinning is characterized by comprising the following steps: the method comprises the following steps:

step 2: determining the types and the number of environment monitoring sensors used in the block ecological region;

and 5: according to different types and quantities of environment monitoring sensors, constructing an initial population of a multi-objective optimization genetic algorithm by a random number generation method;

2. The digital twin-based environmental monitoring sensor deployment method of claim 1, wherein: in the step 2, the method for determining the type of the environment monitoring sensor used in the block ecological region comprises the steps of selecting an air temperature sensor, a road pressure sensor and a rainfall sensor on a road, selecting a road pressure sensor and a rainfall sensor on a bridge and selecting a river flow velocity sensor in a river.

3. The digital twin-based environmental monitoring sensor deployment method of claim 1, wherein: in step 3, the performance evaluation index in the evaluation indexes specifically comprises a monitoring value range, a signal transmission distance and a resolution ratio of the sensor.

4. The digital twin-based environmental monitoring sensor deployment method of claim 3, wherein: and based on the isomerism of the evaluation indexes, carrying out normalization processing on different evaluation indexes by a quantitative normalization method.

5. The digital twin-based environmental monitoring sensor deployment method of claim 1, wherein: step 5, randomly selecting one of 1000 selectable individuals of each environment monitoring sensor aiming at different types of environment monitoring sensors, constructing an initial population of the multi-objective optimization genetic algorithm, and randomly generating parameters of each individual in the initial population; the random parameter generating method includes randomly generating an appointed number of integers in a certain range and filling the integers into a fixed-length array; the resolution parameter in the performance indicator in the sensor evaluation indicator is randomly generated by selecting among several already determined numbers.

6. The digital twin-based environmental monitoring sensor deployment method of claim 5, wherein: the evaluation index is used as a target function in the initial population, the real number code of the selectable individual of each environment monitoring sensor is used as a decision variable, in order to facilitate calculation, the decision variable and the target function are connected in series to form a matrix, the decision variable value and the target function value corresponding to each decision variable are respectively stored at the corresponding positions of chromosomes, the real number code is adopted in the gene coding process, the node number of one sensor deployment scheme forms a gene sequence of feasible solutions, and different sensor deployment schemes correspond to different feasible solutions, namely correspond to different decision variables.

7. The digital twin-based environmental monitoring sensor deployment method of claim 1, wherein: step 7, after an optimal set of optimal deployment of the environment monitoring sensors in the complex scene is obtained, for the air temperature sensors and the road pressure sensors, a scheme with the maximum selective energy parameter in the corresponding optimal solution set is used as an optimization result needing visualization; and for the rainfall sensor and the river flow velocity sensor, selecting a scheme with the largest cost parameter in the corresponding optimal solution set as an optimization result to be visualized.

8. The digital twin-based environmental monitoring sensor deployment method of claim 1, wherein: in the step 8, the TXT text file is used as a middleware to realize data communication between the Unity virtual development engine and the Matlab genetic algorithm programming environment, an optimal deployment scheme is selected in the Matlab genetic algorithm programming environment, and corresponding data are imported into an external designated TXT text file; writing a script in a text component corresponding to the Unity virtual development engine to read the TXT text file, and displaying data; when the parameters of the optimal deployment scheme are changed, the optimization result is changed accordingly and is dynamically updated to the visualization platform for display.

9. The utility model provides an environmental monitoring sensor deploys experiment platform based on digital twin which characterized in that: the experimental platform is used for implementing the digital twin-based environmental monitoring sensor deployment method of any one of claims 1~8.

10. The digital twin-based environmental monitoring sensor deployment experiment platform of claim 9, wherein: the environment monitoring sensor deployment experiment platform based on the digital twin is built based on a Unity virtual development engine and a Matlab genetic algorithm programming environment, and a TXT text file is used as a middleware to realize data communication.