| Summary: | 碩士 === 國立臺灣大學 === 資訊工程學研究所 === 100 === Wireless Sensor Networks (WSNs) are a type of wireless network systems which consist of a large number of wireless sensor nodes and a few gateways. Sensor nodes gather information about the environments and then forward it to the gateways, which in turn relay the information to a server for further processing and analysis. WSNs are initially developed for battlefield purposes by the U.S. military. In the past ten years, the technology continues to develop and begins to serve different purposes, such as scientific investigation, temperature and humidity control, fire forecasting, etc.
We observed that for a number of safety-related WSN applications, e.g., abnormal vibration or tension detection on the bridge, flammable gas density monitoring within the mine pit, the network topology consists of a few long and thin branches, or exhibits a “long-thin topology.” During the regular operation, the network only needs to obtain the sensor reading periodically, so that anomaly can be detected. However, if an accident happens, the network instead needs to enter a different operation mode where a much larger amount of data needs to be transferred. For example, when a gas explosion occurs in the mine pit, the sensor nodes transmit at a much faster rate, so that the positions of the trapped miners can be quickly determined, and the WSN can be used to relay voice and video transmissions from the miners to assist the rescue operation.
During the emergency operation, the WSNs require a large amount of throughput, but the common MAC protocols used in WSNs, usually a variant of the Time Division Multiple Access (TDMA) protocol or the Carrier Sense Multiple Access (CSMA) protocol, have the constraint that only one node can transmit within the receiving range of a node. The throughput is therefore greatly reduced. As the design objective for protocol during the regular operation is often optimized to reduce energy consumption, this does not pose a problem. However, the original protocol is obviously not feasible for the emergency purposes.
The operation of Code Division Multiple Access (CDMA) spreads the original message into a wideband signal by modulating it with a pseudo-noise (PN) code. The scheme removes the aforementioned constraint, and allows multiple nodes to transmit simultaneously within the receiving range of a node. As a result, with a CDMA-based protocol, the throughput could be significantly improved.
In this thesis, we investigate how to design a CDMA-based protocol for a dual-mode WSN system with long-thin topologies. The system would use a TDMA- or CSMA-based protocol during its regular operation and the CDMA-based protocol when switching to the emergency mode. Two most commonly used long-thing topologies, single-chain and multi-chain topologies, are studied in this thesis. For single-chain topologies, we propose an easy-to-implement heuristic power allocation scheme, which has low overhead and outperforms the full power allocation scheme in terms of throughput. For multi-chain topologies, we propose a heuristic scheduling principle, which can increase the number of simultaneous transmissions to the gateway in the same time slot and produces significantly higher throughput. Evaluation results suggest that the use of the CDMA-based protocol in WSNs with long-thin topologies approximately doubles the throughput compared to that of TDMA-based protocol.
Keywords: CDMA, TDMA, wireless sensor networks, long-thin topologies, power control, scheduling
|
|---|
