EVALUATION OF AIR POLLUTION USING UNMANNED AERIAL VEHICLE BASED ON GENETIC ALGORITHM - Scientific conference
Main page
UA RU EN

Congratulation from Internet Conference!

Hello

Рік заснування видання - 2011

EVALUATION OF AIR POLLUTION USING UNMANNED AERIAL VEHICLE BASED ON GENETIC ALGORITHM

16.11.2022 00:08

[1. Information systems and technologies]

Author: Anastasiia Yuriivna Baranovska, student, Vinnytsia National Technical University, Vinnytsia; Maksym Andriyovych Leshok, student, Vinnytsia National Technical University, Vinnytsia


Air pollution is the main environmental cause of premature deaths in Europe. It causes a number of diseases, including cardiovascular diseases, lungs and others. Pollution of the surrounding atmospheric air is the cause of premature death of 600,000 people in the UN regions annually. Determining the exact contribution of individual factors to the development of the disease is often a very difficult task, which is complicated by a large number of effects they have, many of which can also occur among the population and without the influence of these factors.

Applying the approach of machine learning, this article proposes the use of UAV’s (drones) equipped with standard sensors to perform the tasks of monitoring air pollution. The purpose of the study is to choose the most optimal components of hardware and algorithm to solve the problem of monitoring atmospheric air pollution. Correction of the disadvantages of existing methods and increasing the quality of solving the above problem through the use of unmanned aerial vehicles based on a genetic algorithm, which offers accurate assessment of air quality in a short time.

To prove the versatility of the method, it is necessary to take into account the features of the territories on which the quality of air will be monitored:

• Densely populated city, with a lot of transport, industrial infrastructure, a large number of buildings.

• Rural territory with low population density, agricultural areas and complexes of the cultural industry.

• Territory with specific landscape, inaccessible areas, unstable soil (such as mountainous terrain).

• Water areas.

• Dangerous to human life and health places: certain places on industrial sites, closed areas (such as the Chernobyl Exclusion Zone), territories that have suffered from natural cataclysms.

Given all of the above factors, it will be a versatile and convenient method of using unmanned aerial vehicles, as their work is independent of a certain landscape, thanks to moving in the air, and also provides safe access to areas that can be dangerous for life and health of human.

Single computers are used for the design of such drone, as the system is very compact. For example, Jetson Nano or Raspberry Pi4 – a rather popular equipment that has enough resources to solve the task, small in size and adequate in price. Raspberry Pi is considered a more classic option and used in schools and universities. This article will look at the design of the drone on the basis of Raspberry Pi4, as its capabilities are sufficient to solve a given problem and it has a significant advantage in price.





Figure 1 – UAV design scheme

UAV is guided by Pixhawk autopilot, which controls its physical functioning. Raspberry Pi is mounted on a UAV chassis and connected to Pixhawk through a serial port. The sensors are connected to Raspberry Pi with Grove Raspberry Hat (GrovePi), which makes it easy to connect different types of Cots sensors. In particular, Figure 1 shows its components.

To solve this problem, the UAV must use the algorithm of machine learning. The purpose is to comprehend the whole given territory to ensure the quality of the tests on the most optimal route to complete the problem in a minimum time. The optimal route will be a graph through the vertices of which will pass UAVs and take the quality of air. A genetic algorithm was chosen to solve this problem. A genetic algorithm is an algorithm that is commonly used to solve a traveling salesman problem in planning research UAV trajectory. It is a random global search algorithm proposed in accordance with the theory of "survival of the most tangled" in the theory of evolution, which uses data in the form of lines of chromosomal data. According to the process of biological evolution in nature, the optimal solutions for genetic variation are selected; That is, iterative optimization. This algorithm requires more time to execute, but has more accuracy. Previously, the genetic algorithm was not used to UAV due to hardware restriction, but with the advent of computers such as RPI4 or Jetson Nano, the above can be performed to solve the task.

Data on which operations in the genetic algorithm are performed: chromosome - a sequence of peaks that cover the route; population, is the set of chromosomes (number of routes); person - a set of chromosomes that satisfy the solution. Fitness functions are used in genetic programming and genetic algorithms to direct simulations to optimal design solution.

Next, the following stages of the genetic algorithm can be distinguished:

• Creating an initial population.

• Calculation of the fitness function for populations (evaluation).

• Repetition to the criteria of stopping the algorithm. Such a criteria may be: finding a global or pre -optimal solution. In the case of the task described in this article, such a criterion will be the finding of a pre -optimal solution.

• Choice of individuals from the current population (selection)

• Crossing or/and mutation. Stages of crossing: generation of the gap point, formation of the first offspring - where the genes of the first father take to the point of the gap and the genes of the second father after the break point. If you remain unfilled with genes, unremarkable genes are added after the rupture point from the first parent. Similarly with the second offspring, only this time, the second father's genes to the break point are first used. Mutation - a random number is generated from 0 to 100. If the number is less than a given percentage of mutations, a mutation occurs.

• 2 random genes are selected and placed.

• Calculation of the fitness function for all persons. The worst solutions that have arisen in this step are rejected. The size of the population remains the same.

• Formation of a new generation.

Flowchart of this algorithm is shown on Figure 2. 






                      Figure 2 – Genetic Algorithm flowchart

In order to achieve a sample of full coverage of the region, when planning the path, an orderly sequence of all sub -regions is first generated that is to be investigated, which is the decomposition of the territory under study into smaller parts. This sequence represents the procedure for access to target sub -regions. On the basis of this sequence, the trajectory of the coverage is generated, and the UAV flies according to the trajectory and consistently visits each sub -region of interest, to perform operations of analysis of air quality.

References:

1. Як Україна вимірює забруднення повітря? [Internet resource] : [Website]. – Access link: https://ua-energy.org/uk/posts/yak-ukraina-vymiriuie-zabrudnennia-povitria.

2. Alvear, Oscar & Calafate, Carlos & Zema, Nicola & Natalizio, Enrico & Hernandez-Orallo, Enrique & Cano, Juan-Carlos & Manzoni, Pietro. (2018). PdUC-D: A Discretized UAV Guidance System for Air Pollution Monitoring Tasks. 10.1007/978-3-319-76111-4_38. 

3. Jetson Nano vs Raspberry Pi 4: The Differences [Internet resource] : [Website]. – Access link: https://all3dp.com/2/raspberry-pi-vs-jetson-nano-differences/.

4. Using UAV-Based Systems to Monitor Air Pollution in Areas with Poor Accessibility [Internet resource] : [Website]. – Access link: https://www.researchgate.net/publication/318971747.

5. Swarm Intelligence Algorithm [Internet resourse] : [Website]. – Access link: https://www.sciencedirect.com/topics/computer-science/swarm-intelligence-algorithm.

6. Evolution of a salesman: A complete genetic algorithm tutorial [Internet resource] : [Website]. – Access link: https://towardsdatascience.com/evolution-of-a-salesman-a-complete-genetic-algorithm-tutorial-for-python-6fe5d2b3ca35.

7. TSP problem using GA algorithm [Internet resourse] : [Website]. – Access link: https://sourceforge.net/projects/tsp-problem-ga-aco-comparisson/ .

_______________________



Academic supervisor: Yaroslav Anatoliyovych Kulyk, candidate of technical sciences, associate professor, Vinnytsia National Technical University, Vinnytsia




Creative Commons Attribution Ця робота ліцензується відповідно до Creative Commons Attribution 4.0 International License
допомога Знайшли помилку? Виділіть помилковий текст мишкою і натисніть Ctrl + Enter

Another articles in this section

Сonferences

Conference 2024

Conference 2023

Conference 2022

Conference 2021



Міжнародна інтернет-конференція з економіки, інформаційних систем і технологій, психології та педагогіки

Наукова спільнота - інтернет конференції

:: LEX-LINE :: Юридична лінія

Інформаційне суспільство: технологічні, економічні та технічні аспекти становлення