| In order to better understand and interact with the world we are living, we need to process large quantities of data collected from various sources in fine grain. This requires a large number of sensors densely deployed to provide rich data over time, both spatial and temporal, about the physical surroundings. Despite new technology break-throughs in chip manufacturing, power sources and charging methods, the overall capital and operating expense are still key considerations when it comes to the adoption of such pervasive sensing technology. Accordingly, low cost and low-power wireless sensors are needed for massive deployment.;All these factors have led us to envision a scenario, where there are a significant number of sensors deployed in our surroundings, either as a built-in part of the existing infrastructure or part of some specific deployed applications. These small sensors are cheap and require low power, but have moderate computing and storage capability. They detect various types of events with different levels of importance in the physical world, and deliver the data to the nearby sink (s), from where the data will be digested and information/knowledge will be generated.;In this research, we address the problem of how to efficiently aggregate the sensed data from a significant number of wireless sensor nodes densely deployed nearby, paying special attention to reducing the overall cost while providing a satisfactory performance to the applications that use the collected data to enhance the civil life.;This research presents a series of studies on the densely deployed wireless sensor network systems with function reduced sensor nodes, i.e. transmit-only sensor nodes, as key components. Firstly, three studies are presented for the single-hop wireless sensor network systems, in which three different architectures based on different levels of trade-off between cost (in this work we only consider the deployment cost as the overall system cost) and performance, are examined and the corresponding protocol designs are discussed. The results show a promising flexibility of WSNs in achieving different combinations of cost and performance when using the transmit-only nodes. Further, the hybrid WSN cluster architecture and its corresponding protocol design are identified as the most capable solution that worth further investigation. Secondly, various practical implementation issues are discussed and the enhancement schemes are presented and evaluated to demonstrate the effectiveness of the improved solution. Thirdly, the study is extended from the single-hop scenario to the multi-hop one. An efficient approach for multiple hybrid WSN clusters to cooperate on the same task is proposed and evaluated. |
