Network issues for 3D wireless sensor networks.

摘要

Wireless sensor networks (WSN) give the opportunity to monitor the environment by performing sensing tasks in places that are difficult to reach or dangerous for humans. Nevertheless, topographical characteristics of such places and the sensor node's limitations introduce new issues in WSN performance. Additionally, in scenarios where sensors are moving or in rugged terrain, there is a high chance for them to be out of communication range, causing network connectivity problems. Hence, solutions have to take into consideration the aspect of the topography consisting of its three dimensional characteristics, namely, type of terrain, terrain unevenness, and obstacles.;This dissertation discusses several research topics addressing issues relevant to WSN connectivity and area coverage problems. First, changes in sensor communication range are studied by varying sensors' heights relative to the surface. A novel communication technique that relies on the jumping capabilities of sensors is proposed. While the jumping sensor robots are airborne, the change in elevation enhances their ability for a short time to successfully communicate with other sensors that are out of communication range at the ground level. Field experiments were conducted and results show a considerable improvement in wireless communication ranges.;Second, the impact of network connectivity and area coverage in a jumping sensor network is further studied. A Hopping Sensor Network Model is defined to increase sensing area coverage along with the enhancement of network connectivity. A Hopping Sensor Routing Protocol is designed from the model that balances the energy consumption on active jumping sensor nodes. Results from simulations show the increase in area coverage obtained from jumping sensor networks, and the effectiveness of the routing protocol to optimize communication paths while balancing energy depletion in the network.;Third, a distributed wireless sensor network organization to establish a functional network, without requiring initial topology information, is presented. Two decentralized algorithms that use the jumping capabilities of sensors are designed for the discovery of isolated sensors. Simulation results show the success of the algorithms to enhance base station reachability. Additionally, cluster to cluster (C-to-C) packet forwarding schemes relying on boundary jumping sensor gateways are defined and analyzed, showing remarkable savings in network energy consumption.;Fourth, in order to have a functional network, it is important to address connectivity issues in an application oriented manner. This work presents an efficient node redeployment-decision process to produce a functionally heterogeneous (jumping and non-jumping sensors) WSN with a performance guarantee. Network performance is defined as a network fitness formula considering the network Quality of Connectivity (QoC). Decision making algorithms for node relocation and topology defragmentation are presented, along with a discussion of their performance.;Fifth, a multi-step procedure to produce a direction oriented jumping sensor network is presented. A jumping sensor robot approach is introduced for collecting and processing signal strength data into relative geographical orientation information. A directional-orientation decision algorithm is defined to process the orientation information. Furthermore, an error identification and correction procedure is established. This has proven to accurately fix the true orientation of the nodes by using only a pair of location aware beacon nodes.