Insect-Brain inspired Neuromorphic Nanophotonics

我们提出了纳米光子片上设备集成传感和神经计算，灵感来自昆虫的大脑。这将独特地结合联合收割机四条研究线：1）在理解昆虫神经生物学方面的进展，提供经过验证的电路设计，以解决自主导航等重大问题; 2）先进的III-V族半导体纳米线技术，利用光以极低的功耗获得大量互连; 3）可用于新型存储组件的光学高效稳定分子染料; 4）为量子计算开发的电路技术。作为概念的证明，我们的目标是从昆虫眼睛中的偏振光传感到内部罗盘和记忆电路的完整路径，通过这些电路，这些信息被集成在一个连续准确的位置估计中。基于经过验证的模型，我们将首先证明计算原理可以在纳米级系统中使用重叠的光信号来实现，具有高容错性和比现有技术更好的能量和空间足迹数量级。然后，我们将探索从纳米电子和分子染料的神经形态记忆功能，在昆虫大脑中的记忆基板的更深入的调查平行。用于计算的相同纳米结构可用于光学传感，我们将开发一种集成传感器和信息处理阵列，以从偏振天空中提取全局方向信息。直接的结果将是一个节能，强大的芯片，使自主车辆导航没有GPS，具有许多潜在的应用;但更重要的是，我们将开发的新型神经组件可以重新组装成一个广泛的电路，以模仿昆虫大脑中的其他计算。该技术平台可以与硅技术集成，我们将展示升级和商业化的途径。

We propose nanophotonic on-chip devices for integrated sensing and neural computation, inspired by the insect brain. This will uniquely combine four lines of research: 1) progress in understanding insect neurobiology that provides proven circuit designs to solve significant problems such as autonomous navigation; 2) advanced III-V semiconductor nanowire technology that exploits light to obtain a large number of interconnects with extremely low power consumption; 3) optically efficient stable molecular dyes that can be used for novel memory components; 4) circuit technology developed for quantum computing. As proof of concept, we target the complete pathway from polarised light sensing in the insect eye to the internal compass and memory circuits by which this information is integrated in a continuous accurate estimate of location. Building on verified models, we will first demonstrate that the computational principles can be implemented using overlapping light signals in a nanoscale system, with high error tolerance and orders of magnitude better energy and spatial footprint than present technologies. We will then explore neuromorphic memory functionalities from nanoelectronics and molecular dyes, in parallel with deeper investigation of the memory substrates in the insect brain. The same nanostructures used for computing can be used for optical sensing, and we will develop an integrated sensor and information processing array to extract global orientation information from polarised skylight. The direct outcome will be an energy efficient, robust chip enabling autonomous vehicle navigation without GPS, with many potential applications; but more importantly, the novel neural components we will develop can then be re-assembled into a wide spectrum of circuits to mimic other computations in the insect brain. The technology platform can be integrated with silicon technology and we will demonstrate the pathway to upscaling and commercialization.