【作者】 李烁；

【导师】 彭军；

【作者基本信息】 中南大学 ， 交通信息工程及控制， 2014， 博士

【摘要】 摘要：能效优化是无线传感器-执行器网络(Wireless Sensor and Actuator Network, WSAN)的核心问题之一。近年来,以动态拓扑控制和节点分布式协作为基本策略的协作数据汇聚是提升WSAN能量效率的重要研究方向。本文以重载桥梁和路基长期安全状态监测的WSAN为研究背景,采用增强学习、协作通信和博弈论等理论和方法,研究执行器路径规划、多执行器负载平衡和节点干扰抑制等问题,提出WSAN中时延约束下的能效优化机制。主要工作包括以下几个方面：论文从应用于监测重载桥梁和路基的WSAN实例出发,分析了时延约束下多执行器WSAN能效优化的重点和难点,提出了基于移动数据汇聚和协作通信干扰抑制的WSAN系统架构,首先构建具有能量感知的动态拓扑,以此为基础选择轮询点,再规划执行器采集信息的路径；然后针对多执行器情况,采用在线模糊Q学习方法对网络进行自适应分区,保证多个执行器间的通信负载和能耗均衡；针对执行器采集信息过程中相邻节点间存在的数据干扰,提出基于协作通信和博弈论的干扰抑制方法。为实现低数据汇聚时延、高能量效率、高可靠性的WSAN提供一套有效的解决方案。提高网络能效,优化网络生存时间是WSAN的核心目标。论文针对多跳路由数据汇聚的负载均衡问题,选择网络数据汇聚轮询点。首先根据传感器节点的剩余能量构造网络动态生成树；在全局拓扑信息可得的场景下提出基于入度优先的轮询点选择算法,在只能获得局部拓扑信息的场景下提出基于节点通信负载的轮询点选择算法。选择的轮询点作为局部网络数据的汇聚点以分散化、区域化的方式限制网络的路由跳数,均衡网络能耗。针对WSAN的实时性要求,本文提出时延约束下的移动执行器的赛道寻优路径规划算法。以轮询点为执行器路径规划停留的参考位置,将执行器的路径规划问题转换为旅行商问题；考虑轮询点通信半径,采用基于赛道寻优的最近邻点启发式算法,求解执行器遍历所有轮询点的最优移动轨迹；同时将网络数据汇聚的时延约束转换成执行器的移动距离约束,作为算法的迭代收敛条件,保证网络数据汇聚实时性,并通过建立和求解空间域最优规划模型提供算法性能评估。对多执行器场景,论文提出一种在线模糊Q学习的能耗均衡自适应分区算法。将每个执行器作为具有学习和决策能力的智能体,实时获取网络中传感器节点的能量和分布状态信息；再使用模糊推理机制离散化位置和能量分布状态,构建Q学习的状态空间和Q函数；根据当前的网络能量分布状态选择具有最大回报值的分区中心位置,并产生对应的Voronoi划分,实现能耗均衡的分区自适应调整。WSAN中突发数据导致的相邻节点间的数据发送干扰是影响数据传输效率的重要问题。本文提出了一种基于协作数据汇聚的干扰抑制机制。首先将协作通信技术引入WSAN,通过协作中继转发和协作干扰转发的方式提升节点的抗干扰能力；然后考虑到节点自私特性,引入频谱共享机制激励中继节点参与协作干扰抑制；采用斯坦伯格博弈方法对协作干扰抑制过程进行分析；再通过求解博弈均衡来选择最优的策略进行协作信息传输,得到一种干扰情况下的协作数据传输方法。该方法可抑制节点间的干扰影响,确保数据公平、合理的协作传输。仿真实验验证了算法的有效性。更多还原

【Abstract】 Abstract:The optimization of energy efficiency has been a key issue in wireless sensor and actuator networks (WSAN). Recent research has indicated that using dynamic topology control and distributed cooperation in WSAN to achieve collaborative data gathering can greatly increase the network energy efficiency. Under the background of heavy bridges and roadbed safety condition monitoring by WSAN, this paper investigates the multi-actuator load balance, path planning of actuators and transmission interference suppression and so on. Reinforcement learning, cooperative communication and game theory are adopted to propose an energy optimization scheme for WSAN with time delay constraints. The main work is as follows:Considering the WSAN applied in heavy bridge and roadbed safety condition monitoring, this paper analyzes the characteristics and problems of energy optimization, and proposes the system framework based on mobile data gathering and cooperative communication interference suppression. Reinforcement Learning is firstly adopted to divide the network into partitions to balance communication and energy load between actuators. Then the dynamic topology with energy sensing is generated in each partition. Based on it, the polling points selection and actuator path planning strategy are proposed. Finally, to suppress the interference from the neighbor nodes, a cooperative communication and game theory based interference suppression strategy is presented in this paper. The proposed scheme will provide an efficient solution for WSAN with low data gathering delay, high energy efficiency and high reliability.A core aim of WSAN is to improve the energy efficiency and to optimize the network life. This paper investigates the load balance of multi-hop data gathering problems and choose polling point of network data. First, a network dynamic spanning tree is constructed based on the residual energy of wireless sensors. Then in-degree priority polling point selection algorithm is proposed with global topology information and communication load based polling point selection algorithm is proposed with local topology information. The selected polling point restricts the number of hops, balancing the energy cost decentraly and regionally as a local network data gathering point.To satisfy the real-time requirements of WSANs, a heuristic actuator path planning algorithm with delay constraints is proposed. The algorithm first converts mobile actuator path planning problem into TSP optimization problem; then considering communication radius of polling point, the Race Search based nearest neighbor heuristic algorithm is adopted to get the optimal actuator movements, which can traverse all the polling points. Finally, the algorithm converts the time-delay constraints into distance constraints, and uses it in iterative process to obtain the optimal actuator path solution. In order to evaluate the performance of our algorithm, this paper proposes linear programming model describing the actuator path planning for actuators in space domain. The1-ε optimal algorithm is used to achieve the theoretical optimal solution of mobile data gathering. In addition, Matlab tool is used to assess the effectiveness of our algorithm in large scale WS AN.This paper proposes an online fuzzy Q learning based energy balancing adaptive partition algorithm for multi-actuators. Every actuator is treated as an agent with learning and decision abilities, and is used to obtain the energy and state information of sensors. Then fuzzy reasoning is used to discretize the location and energy state. And Q learning state space and Q function are constructed. Then the partition center location is selected based on the current network energy distribution state. Corresponding Voronoi partition is produced to adaptively adjust the partition.Interference among the adjacent nodes caused by the unexpected data is the significant problem, which impacts the efficiency of the data transmission. A collaborative data transmission mechanism based on interference suppression is proposed. Firstly, the cooperative communication is introduced into WSAN, and the ability of anti-interference can be improved by relay transmission and cooperative interference. Considering the selfishness of the nodes, a spectrum sharing mechanism is introduced to motivate the relay node to participate in the cooperative interference suppression, and the Stackelberg game is used to analyze the process of the cooperative interference suppression. Then, a data transmit method with interference is obtained by solving the equilibrium to select the optimal strategy for cooperative information transmission. The method can suppress the interference among nodes, and can guarantee the fairness and rationality of the cooperative data transmission. The simulation results verify the effectiveness of the proposed algorithm.更多还原