| Wireless sensor network(WSN), or sensor network, has broad applications in military affairs, civil affairs, environment monitoring and city management etc. The initial wireless sensor network is static wireless sensor network(SWSN) in which the sensors are not movable once they have been deployed. With the research of sensor network broaden, dynamic wireless sensor network(DWSN) in which the sink can be movable and opportunistic wireless sensor network(OWSN) in which all the sensor nodes can be movable are proposed.Data query of wireless sensor network is one of the important ways for user to actively obtain and analyze the sensing data. With the research of sensor network broadened, many new challenges emerged for data query process in wireless sensor network. On one hand, the query type is broadened from simple aggregation query to complex query. Top-k query and skyline query etc. in wireless sensor network are proposed. How to resolve the weakness of the common query process mechanism and propose the energy-efficient query process mechanism for special query operator in wireless sensor network is new challenges. Additionally, because of the node's mobility in DWSN and OWSN, how to plan the sink's moving route in DWSN to improve the performance of query process and how to propose the query process mechanism for OWSN are all new problems. This thesis focuses on data query in wireless sensor network and researches the key issues on data query in different WSN model. The main work and contributions are presented in the following aspects:(1) Aiming at the higher energy cost of TAG algorithm which is a gerneral exact and snapshot Top-k query in SWSN, this thesis proposes an exact and snapshot Top-k query process algorithm QVTOP which is based on query-verify scheme and span-tree optimized. Firstly, according to the properties of Top-k query operator and the communication character of WSN, a Top-k query process algorithm QVTOP-Base which is based on query-verify scheme is proposed. QVTOP-Base firstly queries the nodes with higher probability to be the Top-k nodes and gets a limit value. Then QVTOP-Base verifies the other nodes with the limit values so as to reduce the communication cost effectively. Based on QVTOP-Base algorithm, with optimizing the span-tree structure to reduce the energy cost spent in querying and verifying phase, QVTOP-Improve algorithm is proposed. Theory analysis and simulation experiment results show for the application in which the sensing data vary smoothly, QVTOP-Base and QVTOP-Improve algorithms are more energy-efficient.(2) Aiming at the skyline query application in which location attribute is often used as a decision-making attribute in SWSN, this thesis proposes the KNN_Skyline query problem in SWSN which combines skyline query and KNN control relation and an exact and snapshot KNN_Skyline query algorithm Ring_Skyline(RS) which is based on the idea of partitioning query area. Firstly, we propose KNN_Skyline query in SWSN. Based on the properties of KNN_Skyline query operator, we propose KNN_Skyline query algorithms RS_Serial and RS_Parallel which are based on the idea of partitioning query area. The main idea is to divide the query area into several annular sub-areas centered on the query position. The nodes in every annular sub-area are combined with a link-cluster structure.The KNN_Skyline query is processed with a dissemination-aggregation scheme in every annular sub-area. The KNN_Skyline queries in different annular sub-areas can be processed with serial or parallel ways. Theory analysis and simulation experiment results show that compared to the central query process algorithm and TAG algorithm, RS_Serial and RS_Parallel can progressively output the query result and reduce the query cost and the query delay effectively.(3) The route of mobile sink must be planned in DWSN before query process. But the present work on route-plan of mobile sink is mainly focused on setting the related mobile parameters of mobile sink and the constraint of the moving area for the mobile sink is less considered. This thesis proposes the problem of route-plan for mobile sink on the case of which the moving area of mobile sink is constrained. With the assumption that the sensor network is a high density network and the buffer-assisted data gathering scheme is adopted for query results collection, for the case of which the global information of sink moving area is known, this thesis proposes a route-plan approach based on Voronoi partition with the theory model of enertgy cost and node load. For the case of which only the local information is known, this thesis proposes a heuristic approach of which the main idea is the longest route dividing the monitoring area into two equal sub-areas is selected as the approximate best one. The simulation experiments results show the feasibility and effectivity of proposed approaches.(4) Aiming at the problem of how to disseminate the query plan in OWSN with higher reliability, lower delay and lower energy cost, this thesis proposes a query dissemination protocol which is based on fine-grained evaluation of node dissemination ability for cycle-model OWSN. Firstly time-probability graph and transmit state-space graph is adopted to describe the cycle-moving model and how to evaluate the dissemination ability of a node based on these two graphs is researched. Then three query dissemination protocols called QADD_d, QADD_r and QADD_c, which are respectively objected to delay optimization, reliability optimization and communication cost optimization, are proposed. The main idea is when two nodes meet each other, every node evaluates its expected dissemination ability according to its state-space graph and then the two meeting nodes dispose and exchange the dissemination plan so that every node has the maximal expected dissemination ability to its keeping dissemination sub-plan. Simulation experiment results show compared to the multiple unicast protocol, QADD_d and QADD_r protocols have less dissemination energy cost, dissemination delay and higher dissemination reliability. |
PDF Full Text Request