北极冰下自主/遥控水下机器人导航与轨迹跟踪研究

将水下机器人用于极地科考，可以通过其携带的多种传感器和设备进行大范围、长时间的冰下观测作业，并取得重要的极地科考资料，如冰下水纹，海冰厚度等。而这些观测数据必须与准确的浮冰位置信息相结合才有实用价值，但北极高纬度特性和长期覆盖的大范围海冰使得一些传统水下机器人导航技术难以有效实施，所以需要研究一种能在北极冰下具有良好性能和高可靠性的导航系统。同时，对北极海冰进行的一系列科学考察，需要载体能够沿着浮冰上预定轨迹航行，但由于海流等外界影响的存在，浮冰处于实时运动中，所以要顺利完成科考任务，必须有相应的轨迹跟踪算法作为支撑。 本文正是针对以上这两点需求，在充分考虑环境特殊性的情况下，研究了水下机器人在北极冰下的导航与轨迹跟踪问题，提出了基于GPS测向仪冰面修正的水下机器人自主导航技术和基于制导控制器的浮冰轨迹跟踪方法，并进行了仿真和实物测试，试验结果表明本文所研究的导航设计方案和浮冰轨迹跟踪方法合理，可以达到北极冰下科考的定位精度和作业要求。 主要工作包括：（1）ARV导航系统设计。详细研究了各导航传感器的输出信息；坐标系定义、坐标变换、相对于浮冰的航位推算以及绝对坐标的求解；最后详细研究了整个冰下导航系统。（2）ARV导航系统误差分析。首先对各传感器的误差进行了详细分析；接着通过分析整个导航算法，建立了浮冰坐标系下ARV位置误差传递方程；最后通过引入具体传感器的参数指标，数值计算了整个导航系统的精度。（3）浮冰轨迹跟踪系统设计。首先对水下机器人在冰下的航迹进行了描述；接着分析了两种冰下制导控制器，视线法制导和横向轨迹误差法制导；最后详细研究了整个ARV航迹控制系统。（4）浮冰轨迹跟踪算法仿真。详细讨论了北极“ARV”的水平面动力学模型。并结合模型对浮冰轨迹跟踪算法进行了Matlab和视景仿真，验证了算法的合理性。（5）湖试和海试结果分析。通过对ARV棋盘山湖试和北极冰下海试数据分析，定量说明了导航误差和轨迹跟踪性能。

2200万年来风尘沉积的粒度变化及古气候意义

Three eolian deposit formations, including Quaternary loess (QL, Liu et al.3 1985), Hipparion red earth (HRE, also called red clay, Liu et al., 1985) and Miocene loess (ML, Guo et al., 2002) constitute a set of unique paleoclimatic archives in northern China dated back to at least 22Ma ago. The Miocene loess is a recently discovered loess-soil sequence. Detailed investigation has been made on its origin, chronology and paleoclimatic significance (Guo et al., 2002), but further work is still needed to obtain detailed paleoclimate information, and mechanical links behind paleoclimatic changes. In this study, grain size analysis of QL, HRE and ML has been conducted on two sections: Qinan and Xifeng. The objective is focused on comparison of the grain size distribution characteristics (GSDC) among different eolian deposit formations, and reconstruction of the Asian monsoon circulation in the past 22 Ma. Results show that GDSC of ML resembles that of QL and HRE, and GDSC of ML is especially similar to HRE. Both ML and HRE contain a significant proportion of fine fraction, however, QL has a large amount of coarse sediments. This is mainly due to that the wind system transported aeolian dust was weaker in the late Tertiary than that in the Quaternary. Grain size difference between loess and paleosol in ML is much smaller than that in QL, indicating that the climatic fluctuations during the late Tertiary were much smaller than that happened in the Quaternary The grain size records of the past 22 Ma reveal several evolution phases of the Asia winter monsoon. -2.7 Ma BP is the most important boundary in the process of the winter monsoon evolution: the wind strength have significantly enhanced since 2.7 Ma ago. During a period between -22.0 and -3.6 Ma, three periods with relatively stronger winter monsoon are recorded in the QA-I section, between 21.2 and 19.9, and 16.0 and 13.3, and 8.7 to 6.9 Ma, respectively. From 3.6 to 2.7, the winter monsoon was enhanced gradually. In the Miocene time, the intensified winter monsoon phases (between 21.2 and 19.9, and 16.0 and 13.3, and 8.7 and 6.9 Ma) seemed to have a close relationship with the uplift of the Tibetan Plateau and/or the ongoing global cooling, but the forcing mechanism behind the Asia winter monsoon evolution need to be further investigated. During the Pliocene-Pleistocene time, the Asia winter monsoon strengthened at 3.6 and 2.7Ma ago are in good agreement with the ongoing global cooling and the Arctic ice sheet development. In the mean time, much evidence suggests that an intense uplift of the Tibetan Plateau occurred at ~3.6 Ma, which is synchronous with a major increase in Asia winter monsoon. Therefore, two major factors may be invoked to explain the winter monsoon enhancement: Arctic ice sheet development and Tibetan uplift. We propose that changes in location and intensity of the Siberian-Mongolian high that were caused by the Tibetan uplift and Arctic ice sheet development might be an important factor for Asian winter monsoon evolution in the Pliocene-Pleistocene.

A Dynamic Model For Ice-Induced Vibration Of Structures

A dynamic model for the ice-induced vibration (IIV) of structures is developed in the present study. Ice properties have been taken into account, such as the discrete failure, the dependence of the crushing strength on the ice velocity, and the randomness of ice failure. The most important prediction of the model is to capture the resonant frequency lock-in, which is analog to that in the vortex-induced vibration. Based on the model, the mechanism of resonant IIV is discussed. It is found that the dependence of the ice crushing strength on the ice velocity plays an important role in the resonant frequency lock-in of IIV. In addition, an intermittent stochastic resonant vibration is simulated from the model. These predictions are supported by the laboratory and field observations reported. The present model is more productive than the previous models of IIV.