论文标题
部分可观测时空混沌系统的无模型预测
Minimoon still on the loose
论文作者
论文摘要
储层计算是预测湍流的有力工具，其简单的架构具有处理大型系统的计算效率。然而，其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间，然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用，该系统模拟了湍流的若干特征。我们表明，使用径向基函数作为非线性投影器，即使只有部分观测并且不知道控制方程，也能稳健地捕捉复杂的系统非线性。最后，我们表明，当测量稀疏、不完整且带有噪声，甚至控制方程变得不准确时，我们的网络仍然可以产生相当准确的预测，从而为实际湍流系统的无模型预测铺平了道路。
On Aug 22, 2016, a bright fireball was observed by the Desert Fireball Network in South Australia. Its pre-atmosphere orbit suggests it was temporarily captured by the Earth-Moon system before impact. A search was conducted two years after the fall, and a meteorite was found after 6 days of searching. The meteorite appeared relatively fresh, had a mass consistent with fireball observation predictions, and was at the predicted location within uncertainties. However, the meteorite did show some weathering and lacked short-lived radionuclides ($^{58}$Co, $^{54}$Mn). A terrestrial age based on cosmogenic $^{14}$C dating was determined; the meteorite has been on the Earth's surface for $3.2\pm1.3$ kyr, ruling out it being connected to the 2016 fireball. Using an upper limit on the pleistieocene terrain age and the total searched area, we find that the contamination probability from another fall is $<2\%$. Thus, the retrieval of the "wrong" meteorite is at odds with the contamination statistics. This is a key example to show that fireball-meteorite pairings should be carefully verified.
