论文标题

部分可观测时空混沌系统的无模型预测

Two-photon real-time device for single-particle holographic tracking (red shot)

论文作者

Semmer, Florian, Emperauger, Marie-Charlotte, Lopez, Colin, Bogicevic, Christine, Treussart, François, Perronet, Karen, Marquier, François

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。

Three-dimension real-time tracking of single emitters is an emerging tool for assessment of biological behavior as intraneuronal transport, for which spatiotemporal resolution is crucial to understand the microscopic interactions between molecular motors. We report the use of second harmonic signal from nonlinear nanoparticles to localize them in a super-localization regime, down to 15 nm precision, and at high refreshing rates, up to 1.1 kHz, allowing us to track the particles in real-time. Holograms dynamically displayed on a digital micro-mirror device are used to steer the excitation laser focus in 3D around the particle on a specific pattern. The particle position is inferred from the collected intensities using a maximum likelihood approach. The holograms are also used to compensate for optical aberrations of the optical system. We report tracking of particles moving faster than 30 $μ$m/s with an uncertainty on the localization around 40 nm. We have been able to track freely moving particles over tens of micrometers, and directional intracellular transport in neurites.

扫码加入交流群

加入微信交流群

微信交流群二维码

扫码加入学术交流群,获取更多资源