Search

Abstract: Event detection is a central component in numerous wireless sensor network (WSN) applications. Nevertheless, the area of event description has not received enough attention. The majority of current event description and detection approaches rely on using precise values to specify event thresholds. However, we believe that crisp values cannot adequately handle the often imprecise sensor readings. In this paper we demonstrate that using fuzzy values instead of crisp ones significantly improves the accuracy of event detection. We also show that our fuzzy logic approach provides higher event detection accuracy than two well-established classification algorithms. A disadvantage of using fuzzy logic is the exponentially growing size of the fuzzy logic rule-base. As sensor nodes have limited memory, storing large rule-bases could be a challenge. To address this issue, we have developed a number of techniques that help reduce the size of the rule-base by more than 70%, while preserving the event detection accuracy. [Copyright &y& Elsevier]

The importance of real-time wireless data transfer is rapidly increasing for Internet of Things (IoT) applications. For example, smart glasses worn by a doctor need to transmit real-time data to a hospital information system, which performs face detection and recognition, for real-time interaction with recognized patients within a certain deadline, which is ideally a few hundred milliseconds. Other emerging IoT applications, e.g., structural health monitoring, clinical monitoring, and industrial process automation, also require real-time wireless data transfer. Those applications have critical demands for real-time and energy-efficient communication through wireless medium. However, it is very challenging to support stringent timing constraints energy-efficiently through wireless medium due to its inherent unreliability and timing-unpredictability. Fortunately, heterogeneous radios are becoming increasingly available in modern embedded devices, offering new opportunities to use multiple wireless technologies to accommodate the needs of real-time applications. In this paper, we formulate the runtime radio selection and data partitioning for real-time IoT applications as an Integer Linear Programming (ILP) problem and present an optimal algorithm that makes quick and optimal decisions when selecting between two radios, a heuristic algorithm for the platforms with more radios, and a runtime algorithm that reduces deadline miss ratio when facing tight deadlines. [ABSTRACT FROM AUTHOR]