Research On Some Key Problems Of Location Awareness In Wireless Sensor Networks

A wireless sensor network involves micro-elector-mechanism system, sensing technology, embedded computing, distributed information processing and wireless communication. It brings up a new pattern for information obtaining and processing. Location data is very important to wireless sensor networks in that contexts captured by wireless sensor networks would be much less valuable without relative location data. In this dissertation, we perform research on essential problems of location awareness in wireless sensor networks.Firstly, we analyze existing node self-localization algorithms in wireless sensor networks, and propose a new one which provides higher accuracy. In a centralized application system, our algorithm utilizes Time-Of-Arrival (TOA) measurements and constrained optimization method to minimize location errors, and applies multiplier method to compute the optimal node locations. Notice that location computations are performed on sink nodes, rather than sensors. By doing this, our algorithm dramatically reduces computational costs on wireless sensor nodes. Finally, we provide simulation results of our localization algorithm. It shows that, compared to traditional TOA and DV-distance locating methods, our algorithm not only achieves higher location accuracy but also requires less sensor loads, which is a key requirement in wireless sensor networks.Most existing work on routing protocol design in wireless sensor networks focus on minimizing energy consumption while preserving the network connectivity. We propose to work on a centralized routing algorithm which is energy-efficient. Given locations of all sensors nodes, our algorithm performs routing calculating, updating and dispatching on sink nodes and can be applied to centralized wireless sensor networks. Based on sensor locations and energy consumption, sink nodes computes an optimal path using Dijkstra algorithm, passes next-hop information to each network node, and then starts a new round of data transmission. Our algorithm performs most of computation and dispatching on sink nodes. Hence, it reduces workload and energy consumption on sensors and thus prolong network lifetime.We consider node self-localization and routing algorithms in wireless sensor networks. Although numerous work has been done in these two individual areas, as far as we know, problems in the combined area have not yet been studied well. In this dissertation, we discuss problems in the combined area and present a cooperative design. We consider the self-localization and routing algorithms in the same manner and using the same parameters and rules. Furthermore, we erase the redundant steps in calculating node localization and location-based routing, which then dramatically reduces unnecessary energy consumption in wireless sensor networks. We also provide simulation results to evaluate our cooperation design. It shows that our design can minimize energy consumption on sensors, reduce system workload, and thus improve the efficiency and prolong the network lifetime. In addition, its performance can satisfy the demand of basic applications. This cooperation design can be used as the basis of a location-aware application system.Nowadays, people spend more time on social activities, such as conferences, exhibitions, etc. When people go to a strange city where local service is insufficient, it would take them a lot of time and effort to find the exact location. In such situations, if we could provide a system that accurately predicts customers’ destinations, we could offer customers more timely services based on the information. With location-aware technology in wireless sensor networks, we present a location prediction algorithm in indoor environments. Our algorithm abstracts the topology of rooms and corridors into a map. Based on this map, it introduces historical and current contexts into a regression model and then predicts customer’s next position using discriminant functions. We also provide some application examples of this prediction model and analyze the scalability, customization and flexibility of this model.In this dissertation, we combine node self-localization, location-based routing strategy, indoor object localization and location prediction technology, and propose an indoor location prediction service model in wireless sensor networks. We then introduce an application example, a wireless sensor network in a hospital, to describe the process of network nodes deployment, node self-localization, location-based routing protocol, user registration, user services and exit. We provide technical and theoretical guidance for indoor location-aware prediction application systems.