Lost node recovery in a city-wide wireless sensor network using adaptive preselect filtering

In this paper, we describe our efforts to combat lost nodes in wireless sensor networks. Often times, the wireless sensor network is limited by interference. Specifically, in a practical setting it was found that lost nodes were caused by location of the sensor nodes in relative proximity to cellular broadcasting stations. This effect was modeled and combated using adaptable front end filtering to show the impact of an intelligent front end system as opposed to the more traditional static preselect filtering commonly applied. Using real system parameters, greater than 10% improvement in the packet reception rate compared to a static filter was achieved by adjusting the filtering to the surrounding environment. Only slight tuning was needed to be able to achieve the result and this indicates the traditional method of using a static preselect filter is suboptimal.