Current Proceedings on Technology

Yazarlar: Tuncay Yigit, Mevlut Ersoy

Konular:-

Anahtar Kelimeler:Wireless Local Area Network,Artificial Neural Network,WLAN performance test

Özet: Mobile devices can be connected to the Internet via wireless local area networks. These devices need to obtain sufficient signal strength at every point in an indoor area to maintain the Internet connection. This condition determines the performance of the wireless network. However, the performance of indoor WLAN, which was built with traditional methods, is not at the required level. The wireless network coverage is determined by the access points' location in the building. The wireless network coverage of the indoor areas is lower than expected due to the complex structure. In this study, we developed performance analyzer software to test the coverage of indoor WLAN that was established with method of trial and error or the network administrator’s experience. Back Propagation Neural Network algorithm is used for performance testing of indoor WLAN. For this purpose, we measured the signal strength of indoor WLAN in Suleyman Demirel University Ertokus Bey Classrooms in the 3rd floor and analyzed the performance. We evaluated the signal strengths of three Access Points (APs) that exist in the indoor WLAN. Signal strength measurements of two Access Points form the Artificial Neural Network (ANN) training set, and signal strength measurement of one Access Point forms the ANN test set. Our analysis shows that indoor WLAN coverage rate is approximately 88 percent and this rate is sufficient for the chosen indoor WLAN to operate. By using our performance analyzer software, network administrators will be able to setup an indoor WLAN that has better coverage.

---