Packet Size Optimization in Wireless Sensor Networks for Smart Grid Applications


Kurt S., YILDIZ H. U. , Yigit M., TAVLI B., GÜNGÖR V. Ç.

IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, cilt.64, ss.2392-2401, 2017 (SCI İndekslerine Giren Dergi) identifier

  • Cilt numarası: 64 Konu: 3
  • Basım Tarihi: 2017
  • Doi Numarası: 10.1109/tie.2016.2619319
  • Dergi Adı: IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
  • Sayfa Sayısı: ss.2392-2401

Özet

Wireless sensor networks (WSNs) are envi-sioned to be an important enabling technology for smart grid (SG) due to the low cost, ease of deployment, and versatility of WSNs. Limited battery energy is the tightest resource constraint on WSNs. Transmission power control and data packet size optimization are powerful mechanisms for prolonging network lifetime and improving energy effi-ciency. Increasing transmission power will reduce the bit error rate (BER) on some links, however, utilizing the high-est power level will lead to inefficient use of battery energy because on links with low path loss achieving low BER is possible without the need to use the highest power level. Utilizing a large packet size is beneficial for increasing the payload-to-overhead ratio, yet, lower packet sizes have the advantage of lower packet error rate. Furthermore, trans-mission power level assignment and packet size selection are interrelated. Therefore, joint optimization of transmission power level and packet size is of utmost importance in WSN lifetime maximization. In this study, we construct a de-tailed link layer model by employing the characteristics of Tmote Sky WSN nodes and channel characteristics based on actual measurements of SG path loss for various envi-ronments. A novel mixed integer programming framework is created by using the aforementioned link layer model for WSN lifetime maximization by joint optimization of trans-mission power level and data packet size. We analyzed the WSN performance by systematic exploration of the parameter space for various SG environments through the numer-ical solutions of the optimization model.