Kapalı ortamlarda kişi tespitinde makine öğrenmesi algoritmalarının karşılaştırmalı başarım analizi
Yükleniyor...
Tarih
2021
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Günümüzde, iç mekan konumlandırma ve kişi takip sistemlerinin uygulama alanları her geçen gün artış göstermektedir. Özellikle, hasta, personel, cihaz ve müşteri takip sistemleri ile akıllı binalar ve kalabalık tahminleme gibi birçok alanda, kişilerin konumlarının veya mekan içerisinde bulunma durumlarının doğru tespiti büyük önem taşımaktadır. İç mekan konumlandırma sistemlerinde genellikle, hedef mobil varlığın üzerine periyodik olarak radyo sinyali gönderen küçük bir cihaz yerleştirilir ve bu cihazdan elde edilen sinyaller ile varlığın konumu belirlenir. Fakat bazı ortamlarda, üzerinde herhangi bir sinyal göndericisi taşımayan varlıkların konumlarının tespit edilmesine ihtiyaç duyulmaktadır. Dolayısıyla, mobil cihaz kullanılmayan radyo tabanlı takip sistemlerinde, radyo sinyallerinde meydana gelen dalgalanmalar analiz edilerek, ortamdaki hareketlilik tahmin edilmeye çalışılır. Bu sistemlerde, ortamın çeşitli noktalarına periyodik olarak radyo sinyalleri gönderen ve diğer cihazların gönderdiği sinyalleri alabilen cihazlar yerleştirilir. Ortamda bulunan herhangi bir nesnenin hareket etmesi durumunda, sinyal gücündeki dalgalanmalar analiz edilerek, ortamdaki hareketlilik ve yoğunluk tahmin edilebilir. Ancak bazı durumlarda, radyo sinyallerinde hareketten kaynaklanmayan, geçici ama nispeten şiddetli dalgalanmalar yaşanabilmektedir. Bu tür dalgalanmalar, yanlış tespitlere sebep olarak, sistemin hassasiyetini ve doğruluğunu düşürmektedir. Makine öğrenmesi tekniklerinin, veriler arasındaki gizli örüntü ve karmaşık ilişkileri ortaya çıkarmadaki başarıları sayesinde, makine öğrenmesi tekniklerine dayalı kişi tespit sistemleri, geleneksel yöntemlere göre doğruluğu daha yüksek tahminleme becerisi sunmaktadır. Dolayısıyla, bu çalışmada, kapalı ortamda kişi tespiti için makine öğrenmesi algoritmalarından yararlanılmıştır. Deneysel çalışmalar kapsamında, 10 farklı geleneksel (Naive Bayes, Çok Katmanlı Algılayıcı (MLP), Destek Vektör Makineleri (SVM) ve K-en Yakın Komşuluk (K-NN)) ve karar ağacı tabanlı (C4.5, Random Forest, Random Tree, REPTree, Decision Stump ve HoeffdingTree) sınıflandırma algoritmaları, kapalı ortamdaki 3 farklı telsiz duyarga düğümünden elde edilen ve 23585 kayıttan oluşan veri seti üzerinde ayrı ayrı uygulanmış, doğruluk oranı ve model oluşturma süresi performanslarına göre karşılaştırılmıştır. Elde edilen sonuçlar incelendiğinde, çalışmada uygulanan tüm algoritmaların kapalı alandaki kişi tespitinde %80’in üzerinde başarı performansı sunduğu ve en başarılı algoritmanın %99.68 doğruluk oranı ile Random Forest olduğu gözlemlenmiştir. Ayrıca, geleneksel ve karar ağacı tabanlı sınıflandırma algoritmaları sağlamış oldukları ortalama doğruluk oranlarına göre kıyaslandığında ise karar ağacı tabanlı algoritmaların %95.78 ile daha yüksek tahminleme becerisi sunduğu görülmektedir.
Nowadays, the application areas of indoor positioning and person tracking systems are increasing day by day. Especially, in many fields such as patient, personnel, device and customer tracking systems and smart buildings and crowd forecasting, it is crucial to determine the location of humans or their presence in the environment. Generally, in the indoor tracking systems, a small device, that periodically sends a radio signal, is attached to the mobile target asset and the location of the asset is determined with the signals obtained from this device. However, in some environments, it is necessary to determine the location of assets that do not carry any signal transmitters on them. Hence, in the systems without mobile devices, the mobility in the environment is estimated by analyzing fluctuations in the radio signals. In these systems, there are some static devices at various points in the environment that send radio signals periodically and can receive signals sent by other devices. If an object moves in the environment, the mobility and density of mobile objects in the environment can be detected by analyzing the fluctuations in signal strength. However, in some cases, temporary but relatively violent fluctuations, which are not related to any movement, may occur in radio signals. These fluctuations reduce the sensitivity and accuracy of the system by causing false detections. Thanks to the success of machine learning techniques in revealing hidden patterns and complex relations among data, human detection systems based on machine learning techniques offer higher accuracy estimation ability than traditional methods. Therefore, in this study, machine learning algorithms are utilized for human detection in indoor environment. Within the experimental studies, 10 different traditional (Naive Bayes, Multilayer Perceptron (MLP), Support Vector Machines (SVM), and K-nearest Neighbors (K-NN)) and decision tree-based (C4.5, Random Forest, Random Tree, REPTree, Decision Stump, and HoeffdingTree) classification algorithms were applied separately on a data set consisting of 23585 records, obtained from 3 different wireless sensor nodes in the indoor environment, and were compared according to accuracy rate and model build time performances. When the obtained results are examined, it is observed that all algorithms applied in the study offer over 80% success performance in the detection of human in indoor environments and the most successful algorithm is Random Forest with an accuracy rate of 99.68%. In addition, when the traditional and decision tree-based classification algorithms compared according to the average accuracy rates they provided, it is seen that decision tree-based algorithms present higher estimation ability with 95.78%.
Nowadays, the application areas of indoor positioning and person tracking systems are increasing day by day. Especially, in many fields such as patient, personnel, device and customer tracking systems and smart buildings and crowd forecasting, it is crucial to determine the location of humans or their presence in the environment. Generally, in the indoor tracking systems, a small device, that periodically sends a radio signal, is attached to the mobile target asset and the location of the asset is determined with the signals obtained from this device. However, in some environments, it is necessary to determine the location of assets that do not carry any signal transmitters on them. Hence, in the systems without mobile devices, the mobility in the environment is estimated by analyzing fluctuations in the radio signals. In these systems, there are some static devices at various points in the environment that send radio signals periodically and can receive signals sent by other devices. If an object moves in the environment, the mobility and density of mobile objects in the environment can be detected by analyzing the fluctuations in signal strength. However, in some cases, temporary but relatively violent fluctuations, which are not related to any movement, may occur in radio signals. These fluctuations reduce the sensitivity and accuracy of the system by causing false detections. Thanks to the success of machine learning techniques in revealing hidden patterns and complex relations among data, human detection systems based on machine learning techniques offer higher accuracy estimation ability than traditional methods. Therefore, in this study, machine learning algorithms are utilized for human detection in indoor environment. Within the experimental studies, 10 different traditional (Naive Bayes, Multilayer Perceptron (MLP), Support Vector Machines (SVM), and K-nearest Neighbors (K-NN)) and decision tree-based (C4.5, Random Forest, Random Tree, REPTree, Decision Stump, and HoeffdingTree) classification algorithms were applied separately on a data set consisting of 23585 records, obtained from 3 different wireless sensor nodes in the indoor environment, and were compared according to accuracy rate and model build time performances. When the obtained results are examined, it is observed that all algorithms applied in the study offer over 80% success performance in the detection of human in indoor environments and the most successful algorithm is Random Forest with an accuracy rate of 99.68%. In addition, when the traditional and decision tree-based classification algorithms compared according to the average accuracy rates they provided, it is seen that decision tree-based algorithms present higher estimation ability with 95.78%.
