Multi-unmanned aerial vehicle path optimization method for industrial park pollution source monitoring
Technical Field
The invention belongs to the technical field of atmospheric pollution source monitoring, and relates to an industrial park pollution source monitoring path optimization method based on multiple unmanned aerial vehicles.
Background
According to incomplete statistics, the existing industrial parks of various levels in China exceed 7000, and make great contribution in promoting industrial aggregation and economic development. Meanwhile, environmental problems caused in the production process of the park enterprises are increasingly prominent, and the park enterprises become an accumulation area of environmental pollution. The government department is favorable for timely evaluating the atmospheric pollution condition in the garden range, monitoring the current emission situation of each pollution source and quickly determining the accurate position of the overproof emission pollution source by strengthening daily environmental monitoring and patrol of the pollution source in the industrial park, and has important significance for realizing targeted treatment of the atmospheric pollution in the park and formulating an effective emergency disposal scheme.
At present, the daily environment monitoring of pollution sources generally adopts modes of a fixed monitoring station, a wireless sensor network, a monitoring vehicle or a handheld monitor and the like. The location of the overproof pollution source is generally obtained by integrating the monitored location thereof with the pollutant concentration information to obtain a relatively rough concentration distribution map, and then the location of the pollution source is estimated. However, due to the limitations of ground conditions, buildings and urban roads, the above monitoring means is often inefficient. Meanwhile, the monitoring points are often unevenly distributed, and a monitoring station is not necessarily arranged or surrounded by the monitoring points right nearby the accident pollution source. Although the environmental monitoring quality can be improved by increasing the number of monitoring stations and the observation frequency, the problems of limited monitoring range, high cost and poor mobility are faced. The research result shows at present that unmanned aerial vehicle environmental monitoring platform has advantages such as flexible, nimble and monitoring range is wide, can compensate current fixed monitoring station and monitoring vehicle not enough, more is favorable to carrying out daily environment inspection to the atmospheric pollution source.
Industrial parks are often a collection of pollution sources, which are numerous and vary in the type of pollutants emitted by different sources. When single unmanned aerial vehicle carries out the time measuring of garden interior atmosphere pollution source, owing to receive unmanned aerial vehicle self duration limit, the monitoring task of partial pollution source position can only be accomplished in flight once, and the efficiency of execution of monitoring task is not high, does not satisfy the requirement that a plurality of pollution sources kept the synchronism as far as possible on the monitoring time yet. Therefore, it is necessary to adopt multiple unmanned aerial vehicles to complete the monitoring tasks of multiple pollution sources. How to research and construct an objective function and a constraint condition according to the specific problem of monitoring the pollution source of the industrial park and adopt a proper technical means to find an optimal monitoring path of the atmospheric pollution source in the park is the problem which is mainly solved by the invention.
The method is different from the traditional multi-station carrier (mTSP) problem, and aims at the specific problem background of pollution source monitoring of the industrial park, the total length of a monitoring path of each unmanned aerial vehicle, the flight path turning angle of the unmanned aerial vehicle, the avoiding radius of the unmanned aerial vehicle to an obstacle, the endurance time limit of a single unmanned aerial vehicle, the pollutant emission type and pollutant emission quantity limit of each pollution source, the residence time limit of each pollution source unmanned aerial vehicle and the optimization of the quantity of the unmanned aerial vehicles are considered in the construction of a planning model. Meanwhile, a multi-chromosome genetic algorithm based on a complex mutation tree is adopted in the solution of the planning model, and the algorithm can reasonably select mutation operators according to given optimization direction rules to optimize multiple unmanned aerial vehicle monitoring paths aiming at pollution sources of the industrial park.
Disclosure of Invention
The invention aims to provide an industrial park pollution source monitoring path optimization method based on multiple unmanned aerial vehicles. The object of the present invention is achieved by the following technique.
A method for optimizing monitoring paths of pollution sources in an industrial park based on multiple unmanned aerial vehicles comprises the steps of constructing an objective function and a constraint condition aiming at a specific problem of monitoring the pollution sources in the industrial park, wherein the objective function considers that the total length of monitoring paths of each unmanned aerial vehicle is shortest, the turning angle of a flight path of each unmanned aerial vehicle is minimum (the smoothness of flight is ensured so as to reduce the flight energy consumption of the unmanned aerial vehicles), and the avoiding radius of each unmanned aerial vehicle to an obstacle is maximum (the method can ensure that the path adaptability value of the unmanned aerial vehicles to the obstacle is small so as to ensure that the finally obtained path and the obstacle keep a certain distance); the specific constraint conditions comprise the endurance time limit of a single unmanned aerial vehicle, the pollutant emission category and the pollutant emission quantity limit of each pollution source, the residence time limit of each pollution source unmanned aerial vehicle, the takeoff starting point and the terminal point limit of the unmanned aerial vehicle and the like. And then solving the planning model by adopting a multi-chromosome genetic algorithm based on the complex mutation tree to obtain an optimal industrial park pollution source monitoring path. The method comprises the following steps:
s1: acquiring environment information of a target industrial park, wherein the environment information comprises coordinate information of a starting base of an unmanned aerial vehicle and n pollution source positions, the endurance time of a single unmanned aerial vehicle, the average flight speed of the unmanned aerial vehicle, the flight unit energy consumption of the unmanned aerial vehicle, the hovering unit energy consumption of the unmanned aerial vehicle, pollutant discharge types and pollutant discharge amounts of pollution sources, the residence time of the unmanned aerial vehicles of the pollution sources, the number m of the unmanned aerial vehicles and the like;
s2: constructing a path optimization problem model for monitoring the atmospheric pollution source according to the environmental information of the industrial park, and setting a specific objective function and constraint conditions of the model;
s3: solving the planning model based on a multi-chromosome genetic algorithm of a complex mutation tree to obtain optimal industrial park pollution source monitoring paths of a plurality of unmanned aerial vehicles;
further, the objective function F in step S2 is:
min F=w1·L/Lmax+w2·Dmax/D+w3·θ/θmax (1)
wherein, w1,w2And w3In order to monitor the total length L of the path, the avoiding radius D of the unmanned aerial vehicle to the obstacle and the weight of three sub-targets of the turning angle theta of the flight path of the unmanned aerial vehicle, w1+w2+w3=1;Lmax、DmaxAnd thetamaThe purpose is to de-dimensionalize the maximum value which is set in advance. The reason why the avoidance radius D is reciprocal is that it is expected that the larger the avoidance radius of the drone is, the better.
Wherein L is
kTotal time (distance) flown for the k-th drone's route; m is the number of unmanned aerial vehicles; n is the number of pollution sources in the park; c
ijThe flight time of the unmanned aerial vehicle from pollution source i to pollution source j; (x)
i,y
i) Is the ith pollution source position; (x)
j,y
j) Is the j-th contamination source position;
monitoring the residence time of the unmanned aerial vehicle in the pollution source j; v is the average flight speed of the unmanned aerial vehicle; s is the longest endurance time (maximum flight distance) of a single unmanned aerial vehicle; λ is penalty factor, when L of k route
kAnd when the sum is greater than S, the path is not the optimal path, the path is made to be a poor solution through a penalty factor lambda, and the path is removed from the feasible solution.
Wherein D ismThe sum of the avoidance radii of the flight path of the mth unmanned aerial vehicle to the obstacles, g is the number of the obstacles, dkDistance of flight path from kth obstacle, rkThe maximum peripheral radius of the obstruction.
Wherein, theta
mFor the mth frame without peopleThe sum of the turning angles in the flight path of the aircraft, w being the number of turning angles,
the k-th turning angle on the flight path.
Further, the constraint conditions in step S2 are:
all pollution source monitoring points, unmanned aerial vehicle takeoff bases and terminal bases in the traversal industrial park are starting points:
each source of contamination remains only once:
the cruising ability of single unmanned aerial vehicle retrains:
other constraints are:
xijk∈{0,1};1≤k≤m;1≤m≤M (13)
wherein M is the preset number of unmanned aerial vehicles; m is the number of the optimized unmanned aerial vehicles;
further, step S3 includes the following steps:
s31: setting parameters of a multi-chromosome genetic algorithm based on a complex variation tree, wherein the parameters comprise population number Pop, iteration times S and unmanned aerial vehicle number (namely chromosome number) M;
s32: initializing a population, and coding the population by adopting a multi-chromosome coding mode according to the current chromosome number M, wherein one chromosome represents a pollution source monitoring point sequence of an unmanned aerial vehicle;
s33: calculating an algorithm fitness function according to the target function F and the constraint conditions thereof in the step S2;
s34: operators with different complexity degrees in the complex mutation tree are randomly selected to optimize various groups;
wherein the complex mutation tree is divided into a simple mutation operator, a general mutation operator and a complex mutation operator.
The simple mutation operators include a Swap operator, a Reverse operator, a Slide operator, an Insert operator, and a cross operator.
The general mutation operator is formed by combining two random simple operators, including a Swap and cross (Swap & Crossover) operator, a Reverse and cross (Reverse & Crossover) operator, a Slide and cross (Slide & Crossover) operator, an Insert and Slide (Insert & Slide) operator and an Insert and cross (Insert & Crossover) operator.
The complex mutation operator is formed by combining three random simple operators, including a Swap & Reverse & Crossover operator, a Slide & Insert & Crossover operator, a Swap & Slide & Crossover operator and a Slide & Reverse & Crossover operator.
S35: judging whether the current iteration times reach the maximum iteration times S or not, and if so, recording the current optimal solution; otherwise, return to step S33.
S36: judging whether the number (chromosome number) of the current unmanned aerial vehicles reaches a lower limit, and if so, outputting a current optimal solution; otherwise, the number of unmanned planes is-1, and the step S32 is returned.
S37: and setting pollution source monitoring points of each unmanned aerial vehicle and hovering time of each monitoring point according to the current optimal pollution source monitoring path of the industrial park, and performing pollution source monitoring tasks of the industrial park.
Drawings
FIG. 1 is a flow chart of a multiple chromosome genetic algorithm employed in the present invention;
FIG. 2 is a schematic diagram of the Crossover operator employed in the present invention:
FIG. 3 is a schematic diagram of the Insert operator employed in the present invention;
FIG. 4 is a schematic diagram of the Reverse operator employed in the present invention;
FIG. 5 is a schematic diagram of Slide operators employed in the present invention;
FIG. 6 is a schematic diagram of the Swap operator employed in the present invention;
FIG. 7 is a schematic diagram of the complex operator employed in the present invention;
FIG. 8 is a diagram of monitoring point placement of pollution sources in an industrial park according to an embodiment of the present invention;
FIG. 9 shows the optimization results of the monitoring path of the pollution source in the industrial park according to the present invention;
FIG. 10 is a MATLAB simulation result of the monitoring path of the pollution source of the industrial park in the present invention;
detailed description of the preferred embodiment
The specific embodiment of the invention is as follows:
the technical problems to be solved in this example are: knowing the position coordinate information of each pollution source in an industrial park, the pollutant discharge category and the pollutant discharge quantity of each pollution source, taking the central point of a monitoring area as an unmanned aerial vehicle starting base, and navigating m unmanned aerial vehicles to n atmospheric monitoring points and finally returning to the starting base; every unmanned aerial vehicle can go to a plurality of monitoring points and carry out the sampling monitoring, require every monitoring point all to be sampled the monitoring and every monitoring point can only be sampled by an unmanned aerial vehicle, discharge pollutant classification and discharge pollutant quantity according to each pollution source and confirm its sampling dwell time, m unmanned aerial vehicles need get back to the base of starting after accomplishing the monitoring task again, require to plan out the optimal monitoring route for every unmanned aerial vehicle, and unmanned aerial vehicle quantity finally reaches the optimization.
In view of the above technical problems, an embodiment of the present invention provides a method for optimizing monitoring paths of pollution sources in an industrial park based on multiple unmanned aerial vehicles, as shown in fig. 1, the method includes the following steps:
s1: acquiring environment information of a target industrial park, wherein the floor area of the target industrial park is known to be 36 square kilometers, 20 industrial plants are randomly distributed in the park, and the distribution diagram is shown in FIG. 8; the coordinate information, pollutant emission types and unmanned aerial vehicle residence time of 20 pollution sources are shown in table 1; the endurance time of the single unmanned aerial vehicle is 40min, and the average flying speed is 5 m/s.
Table 1 details of the sources of pollution
S2: constructing a path optimization problem model for monitoring the atmospheric pollution source according to the environmental information of the industrial park, and setting a specific objective function and constraint conditions of the model;
setting specific objective function formulas (1) to (7) of the model; in this embodiment, the radius of dodging of unmanned aerial vehicle to the barrier is not considered for the time, and the stability of turning when unmanned aerial vehicle flies is stronger, therefore w1Take 0.95, w2Take 0, w30.05 is taken.
Setting specific constraint conditions of the model from formula (8) to formula (13);
s3: and solving the planning model based on a multi-chromosome genetic algorithm of the complex mutation tree. In this embodiment, the population number Pop is set to 80, the number of iterations is set to 500, the initial number of unmanned aerial vehicles is preset to 8, and finally, the optimal industrial park pollution source monitoring paths of a plurality of unmanned aerial vehicles are obtained, as shown in fig. 9.
As can be seen from fig. 9, the optimized number of the unmanned aerial vehicles is 6, the flight path of the unmanned aerial vehicle 1 is 1 → 18 → 2 → 1, the flight path of the unmanned aerial vehicle 2 is 1 → 15 → 20 → 1, the flight path of the unmanned aerial vehicle 3 is 1 → 12 → 19 → 1, the flight path of the unmanned aerial vehicle 4 is 1 → 16 → 17 → 11 → 10 → 1, the flight path of the unmanned aerial vehicle 5 is 1 → 6 → 5 → 14 → 1, and the flight path of the unmanned aerial vehicle 6 is 1 → 4 → 9 → 13 → 8 → 1; the total length of the flight path of 6 drones is 49.0073km, and the number of iterations of the calculation program is 435.
According to the calculation result, the method can be used for solving the problem of path optimization of multiple unmanned aerial vehicles for monitoring the pollution source in the industrial park.