İNSANSIZ HAVA ARAÇLARINDA KÜMELENME VE SÜRÜ KONTROLÜ
Yükleniyor...
Dosyalar
Tarih
2022-03
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Bu çalışmada, insansız hava araçlarında sürü kontrolü ve kümelenmesi incelenerek, sürü davranışının matematiksel modeli için Parçacık Sürü Optimizasyonu (PSO), formasyon kontrol işlemi için PD tabanlı kontrol algoritması ve sürü İHA’lar arası çarpışma engelleme işleminde potansiyel fonksiyon kullanılmıştır. Diğer sürü kontrol algoritmaları, Kuş Sürüsü (BSA), Yapay Arı Kolonisi (ABA), Yarasa Algoritması (BA) ve Ateş Böceği (FA) algoritmaları modellenerek simülasyon ortamında minimum arama performansı karşılaştırılmıştır. Merkezi kontrol işlemleri için özgün yazılım çerçevesi geliştirilmiştir. Sürü üyelerinin merkezi kontrol olmadan lider tarafından kontrolü için algoritma geliştirilmiştir. Sürü liderinin belirlenmesi işleminde tek katmanlı yapay sinir ağı modeli oluşturularak sürü merkezine ve görev gereksinimlerine uygun lider atama işlemi gerçekleştirilmiştir. Her sürü bireyinin diğer sürü bireylerinin bilgilerine ulaşması için, oluşturulan Wi-Fi ağına, yer istasyonundan telemetri (konum, hız vb.) verilerini göndererek merkezi veya lider tarafından kontrolü sağlanmış ve konum eniyileme işlemi gerçekleştirilmiştir. Geometrik şekillerin (çizgi, üçgen, kare, beşgen, hilal vb.) formasyon noktaları oluşturulmuştur. Oluşturulan geometrik formasyon noktalarına hangi bireyin gideceğini belirlemek için eş zamanlı atama algoritması kullanılmıştır. Modellerin MATLAB ile analizleri yapılarak, GAZEBO simülasyon ortamında döner kanat insansız hava araçlarının modellenmesi ve uçuş testleri ile görevlerin doğruluğu sağlanmıştır. Çalışma da askeri ve sivil alanlarda sürü halinde görev icra edilerek, askeri alanlarda üstünlük, sivil alanlarda avantaj ve kolaylık elde edilmesine olanak sağlanacaktır.
In this paper, swarm control and clustering for unmanned aerial vehicles are investigated. Particle swarm optimisation (PSO) is used for the mathematical model of swarm behaviour, the PD -based control algorithm is used for the formation control process, and the potential function is used for the collision avoidance process between swarm UAVs. Other swarm control algorithms, Bird Swarm (BSA), Artificial Bee Colony (ABA), Bat Algorithm (BA) and Firefly (FA) were modelled and the minimum search performance was compared in the simulation environment. A unique software framework for central control operations was developed. An algorithm was developed for leader control of herd members without central control. In determining the herd leader, a single-layer artificial neural network model was created and the process of assigning the leader was carried out according to the herd centre and task requirements. In order for each individual of the herd to reach the information of the other herd members, the created Wi-Fi network was controlled by the control centre or the leader by sending telemetry data (position, speed, etc.) from the ground station and performing the location optimisation process. Formation points with geometric shapes (line, triangle, square, pentagon, crescent, etc.) were created. The simultaneous assignment algorithm was used to determine which person would go to the created geometric formation points. Analysis of the models using MATLAB, modelling of unmanned rotorcraft and flight tests in the simulation environment GAZEBO ensured the accuracy of the missions. In the study, it will be possible to achieve superiority in military domains and advantage and convenience in civil domains by performing tasks in a swarm in military and civil domains."
In this paper, swarm control and clustering for unmanned aerial vehicles are investigated. Particle swarm optimisation (PSO) is used for the mathematical model of swarm behaviour, the PD -based control algorithm is used for the formation control process, and the potential function is used for the collision avoidance process between swarm UAVs. Other swarm control algorithms, Bird Swarm (BSA), Artificial Bee Colony (ABA), Bat Algorithm (BA) and Firefly (FA) were modelled and the minimum search performance was compared in the simulation environment. A unique software framework for central control operations was developed. An algorithm was developed for leader control of herd members without central control. In determining the herd leader, a single-layer artificial neural network model was created and the process of assigning the leader was carried out according to the herd centre and task requirements. In order for each individual of the herd to reach the information of the other herd members, the created Wi-Fi network was controlled by the control centre or the leader by sending telemetry data (position, speed, etc.) from the ground station and performing the location optimisation process. Formation points with geometric shapes (line, triangle, square, pentagon, crescent, etc.) were created. The simultaneous assignment algorithm was used to determine which person would go to the created geometric formation points. Analysis of the models using MATLAB, modelling of unmanned rotorcraft and flight tests in the simulation environment GAZEBO ensured the accuracy of the missions. In the study, it will be possible to achieve superiority in military domains and advantage and convenience in civil domains by performing tasks in a swarm in military and civil domains."
Açıklama
Anahtar Kelimeler
İnsansız hava araçları, sürü kontrol, kümelenme, formasyon kontrol., Unmanned aerial vehicle, swarm control, clustering, formation control.
