REAP: Robust, Energy Efficient Wireless Sensor Networks for Outdoor Scenarios by Adaption of Operation Parameters

点击翻译按钮获取中文摘要

The opportunities for Wireless Sensor Networks are as well various and challenging. In agriculture, e.g., distributed sensors can be used to measure plant growth or for soil analysis. Here, robustness of single nodes and the whole network is of major importance for successful projects or smart farming approaches. Almost every environmental impact has also direct or indirect influence on the lifetime of the nodes or the network, whereby also huge deviations in environmental conditions occur. Provoked by sun exposure, a single sensor node can show a difference in temperature of 56° C in just one day. Additionally, there is an intense heterogeneity in temperature between the particular nodes within the network.To enable a long-term operation of sensor networks under such conditions they should be very robust against such influences and energy supply has to be ensured. Besides other things, robustness can be achieved by reducing the energy requirements of the node, respectively by adapting the requirements to the generation or vice versa. Additionally, some tasks (e.g., record, store, or process data) can be postponed or relocated in order to adapt them in the temporal or the spatial dimension, on a single node and/or within the network, and to adjust them to the specific environmental or energy conditions.So, especially in outdoor scenarios we have the situation that a.) Some nodes may work more energy efficient than others - due to their temperature and their individual characteristics (heterogeneity within the network) and b.) Changing environmental conditions (over time - e.g. direct sunlight / shadow or day / night) lead to dynamic shift of the energy efficiency of single nodes or the whole network. Opportunities for Energy Harvesting share this dynamics, as well.Monitoring the changing environmental conditions allows to benefit from the systems dynamics. Realistic models for energy and reliability will be derived from these measurements. Combined with the knowledge about the characteristics of the nodes in the network, this information can, e.g., be used for routing decisions, task scheduling, and processing within the network. By predicting the environmental variables, a suitable time for performing intensive computations or forwarding data in delay tolerant networks can be determined. Specific scheduling and routing protocols will be developed to achieve a robust wireless sensor network under dynamic conditions. The investigated relations and the implemented algorithms will be evaluated continuously in a real environment.