Optimum Configuration and Measurement Scheme for Dynamic Autonomous Distributed Cooperative Networks

针对应用于移动ad hoc/mesh网络、包括家庭网络在内的传感器网络以及P2P自治网络，获得了以下研究成果。(1)移动自组网（ad hoc/mesh）网络：针对多目标传输相同信息的组播通信，提出了一种在静态节点环境和移动节点环境下，通过控制中继节点的位置和数量，实现高可靠性和高性能通信的新算法。针对多媒体通信，提出了一种采用前向纠错（FEC）机制的新算法，通过动态改变恢复频率和冗余度，大大提高了传输率。(2)传感器网络：针对有望作为传感器网络部署的ZigBee，提出了一种新颖的算法，在给定传感器放置位置的情况下，将汇聚节点（协调器）、中继节点（路由器）和传感器（终端设备）的总功耗最小化。提出了传感器网络作为家庭网络时，传感器网络与家电网络之间网关的节能软件架构。(3) P2P自治网络：研究了高效的网络广播或泛洪方案和通信节点间通信流量交换的估计方案。针对高效的网络广播或泛洪问题，提出了将双向链路在发送前分解为单向链路的方案，以及每个网络节点计算由相邻节点根的最短路径树并将树信息提前存储为路由表的方案。在通信流量交换估计方案中，将估计问题形式化为确定性优化问题或概率优化问题，并给出了每个问题的估计解。

The following research results were obtained aiming at being applied to mobile ad hoc/mesh networks, sensor networks including home networks, and P2P autonomous networks. (1) Mobile ad hoc/mesh networks: In multicast communication in which the same information is transmitted to multiple destinations, a novel algorithm, which enables highly reliable and high performance communication by controlling the placement and the number of relaying nodes, was proposed in both static node environment and mobile environment. Regarding multimedia communication, a new algorithm, which employs Forward Error Correction (FEC) mechanism and highly improves the delivery ratio by dynamically changing the recovery frequency and redundancy, was developed. (2) Sensor networks: Concerning ZigBee which is expected to be deployed as sensor network, a novel algorithm, which minimizes the total power consumption of a sink node (coordinator), relaying nodes (routers) and sensors (end devices) when placement of sensors are given. Energy-efficient software architecture of a gateway between a sensor network and a home appliances network is also proposed when the sensor network is used as a home network. (3) P2P autonomous networks: Research was conducted on efficient network broadcast or flooding scheme and estimation scheme of communication traffic exchange between communication nodes. Regarding the efficient network broadcast or flooding, a scheme in which bidirectional links are decomposed into unidirectional links before sending, and a scheme in which each in-network node calculates the shortest path tree rooted by neighboring nodes and stores the tree information in advance as routing tables were proposed. In the estimation scheme of communication traffic exchange, the authors formalize the estimation problem as a deterministic or probabilistic optimization problem, and give the estimation solution for each problem.