Development of a novel intelligent wireless sensor actor network for agricultural applications

Abstract

Deployment of a successful wireless sensor actor network requires intelligent processing schemes, effective power consumption and reliable communication links. A novel intelligent wireless sensor-actor network (IWSAN) based on Fuzzy Inference System for agricultural greenhouse climate control is presented in this research. The two most important mutual effects agricultural greenhouse climate parameters are considered which are the temperature and humidity during diurnal and nocturnal time. Design, simulation and implementation of IWSAN for agricultural greenhouse climate control based on Fuzzy Inference System is outlined both for off the shelf and handmade platforms. The integration of artificial intelligence with WSAN proves superior features in comparison with traditional wired control systems and simple wireless monitoring and acting network. IWSAN proves high efficiency of controlling task of ±0.5 and ±1 tolerance of setting points, scalability, mobility and cost effective of handmade platform, beside flexibility of using the system in various geographical areas besides the capability of tuning the whole system for other agricultural tasks. New Dynamic Check-In Interval technique (DCI) is modelled and developed to prolong the lifetime of network sensor nodes. It is implemented as software embedded in sensor nodes while a proposed hardware solution is conducted. The DCI technique of software implementation offers maximum of 10days while 367days for proposed hardware solution with a threshold value of 0.5. IWSAN of non beacon start network topology provides a 1.24 year lifetime for sensor nodes based 1 minute sleep period and powered by 210mAh battery. In agricultural greenhouse, unreliable link quality of IWSAN may raise unpredictable packet loss that is considered as a factor of more noteworthy effect on the performance of the control system. The system reliability of IWSAN in the greenhouse is high with minimum success rate of packet transmission of 85% and achieves settled performance during one week of operation. The unpredictability in the system performance is minimized when one or more sensor nodes are failed. Vegetation path loss in the greenhouse is modelled as a function of antenna heights, separation distance and various foliage depths, simulated with square grid distribution scheme and programmed to assist nowadays well known WSN simulators. The maximum partitioning distance for reliable communication in vegetation field is determined where network connectivity based on MED vegetation model shows perfect connectivity of 100% of foliage depth less than 50m while ITU model exposes less connectivity of about 88%.