On the Usability of Wireless Sensor Networks in Smart Farming Applications

Abstract

Wireless communication has become a vital part of modern life. In recent years, mobile networks have undergone a rapid transformation from analog telephony to digital data networks, thereby opening up completely new possibilities for private and commercial use. The crucial factor for the robustness of these networks, however, is expanding mobile radio stations to cover as many areas as possible. Given that this expansion is being conducted in Germany by private commercial enterprises, it is not yet economically feasible to implement nationwide coverage. This leads to the situation where populated areas are served first, and rural areas are served less or not at all. The focus on inhabited areas poses some challenges for Wireless Sensor Networks (WSNs) because the expansion of mobile communications in rural or cultivated areas is usually not possible, or it is very inadequate because of cost reasons. For applications in the context of Smart Farming or Precision Farming, however, a way is needed to transmit the measured sensor data or status messages to cloud services. One method to realize this data transmission is to use Delay Tolerant Networking. This dissertation examines the use of such networks by considering two different agricultural scenarios. In one scenario, the use of WSNs in a large agricultural area is evaluated, where a heterogeneous wireless network of four different types of sensor nodes is considered. This network includes battery-powered and solar-powered sensor nodes, agricultural machines, and mobile devices. In the second scenario, a communication network is installed in a food storage warehouse, more specifically, a potato warehouse. The high water content of the potato creates a challenging environment for wireless communication between sensor nodes. A suitable radio for wireless communication of the measured data is determined, and a communication protocol is developed. The combination of both allows reliable data transmission between the sensor nodes. Finally, a routing protocol adapted to a special type of sensor node is presented. These sensor nodes essentially consist of two very contrasting units, an energy-saving part and a high-performance part. The presented protocol, Adaptive Wake-Up Routing, can abstract both parts in such a way that for other nodes, this double function is not noticeable and appears as one node. In addition, factors such as battery charge levels, time of day, and the amount of data to be sent are considered.