自适应海洋采样是我国海洋资源开发与走向深蓝进程中一个重要的研究课题。为实现水下移动传感网络的协作数据采样，传感器节点平台-自主水下航行器（AUV）需运动到采样信息最丰富的观测配置上。为此，本项目深入开展以下内容研究：（1）提出一种海洋环境时空数据模型建模与分布式估计方法，实现对温度、盐度、热液羽流等一类采样对象的建模与估计；（2）研究设计一种事件触发的逻辑通讯机制，在各AUV进行局部、间歇性通讯情形下，获得对环境参数的协同感知。（3）从理论上分析观测配置与采样性能之间的关系，同时考虑通讯时延、数据丢包等各类约束，以获得最优采样为目的，建立多AUV协作采样性能评价模型。（4）基于自适应动态规划理论对AUV的航迹进行优化与控制，驱动各AUV保持对采样环境的最优观测。通过本项目的研究，构建一种由采样数据信息驱动的多AUV协作、认知与控制框架，为自适应海洋采样提供新的理论和方法。

Adaptive ocean sampling is a challenging research topic to China's ocean resources development,exploration and going to deep blue. To meet the goal of designing and proving an effective and reliable underwater mobile sensor network for collecting the richest data set in an uncertain environment given limited resources, the following issues will be considered: (1) The modeling and distiributed estimation algorithm will be proposed to characterize ocean spatio-temporal random fields, under which, the temperature, salinity, hydrothermal plumes and other interest ocean fields could be modeled and estimated. (2) An event-triggered communication logic will be poposed to ensure all the sensor nodes have coordinated cognition for the ocean fields, when the communication between AUVs can only be locally and intermittently.(3) The relationships between the deployment of sensors and sampling performance will be analyzed theoretically,and the corresponding performance metric constrained by communication delays, transmission loss will be proposed to derive the optimal paths for the network of mobile sensors.(4) The paths of AUVs will be optimized and controlled to take measurments of ocean spatio-temporal random fields and collect the richest data sets based on adaptive dynamic programming theory. An expected result of this work is to construct a cooperative,cognitive and control framwork for the multiple AUVs, which can act in response to changing conditions as measured data druing the sampling. We wish the aforementioned researches could present a novel theory and method for adaptive ocean sampling networks.