Asymmetric Visibility via Designer Obscurants and Computational Photonics

This DARPA-STTR program aims to research and develop a computational and theoretical framework to assess the extent of asymmetric visibility through an externally tunable nanoparticle-based obscurant cloud, and how much such asymmetry can be enhanced by using a computational backend. Essentially, we will explore how the previous knowledge of the statistics of the obscurant can help the computational reconstruction and can be exploited to enhance the asymmetry. We will identify fundamental limits in progressively sophisticated schemes, ranging from simple nanoparticle geometrical anisotropy to switching-based coded aperture schemes in tandem with computational reconstruction. We will develop fundamental limits in optics-only scenarios of passive, active, and nonreciprocal nanoparticles, as well as optics-plus-computational-imaging scenarios for which there is little understanding of the limits to what is possible. Information theoretic metrics will be used to identify such limits as well as design the system. A complete end-to-end design framework will be established to co-optimize the obscurant with the computational backend. This framework will employ a forward model consisting of multi-scale electromagnetic simulation coupled with computational reconstruction. The inverse design process will employ an automatic differentiation approach. We will perform experimental measurements testing the theoretical predictions, with a key goal being the measurement of computation-based visibility enhancements beyond the optics-only fundamental limits. A successful program would identify optimal long-term pathways for one-way visibility, the possible material systems, control schemes, metrics, and fundamental principles for achieving them. Three classes of NPs exhibiting anisotropic scattering patterns will be studied: one with geometrical anisotropy, which can be actively tuned by photoinduced structural changes, such as plasmon-enhanced volume changes. We will determine fundamental bounds that will likely tradeoff between asymmetric visibility and total transmissivity. In the second class, we will consider nanoparticles that support (reciprocal) gain, whereupon stimulated emission can be activated to generate asymmetric visibility that goes beyond the passivity bounds. Finally, going beyond the reciprocal passive and gain cases, we will consider particles with nonreciprocal responses, induced by nonlinear and/or magneto-optical effects, and develop the theory for ultimate bounds on nonreciprocity-based asymmetric visibility.