This remained effective for multiple wavelengths. Without requiring prior knowledge of the complex EM environment, the proposed SELAF system can adaptively converge incident waves to a user-defined position, even in the presence of multiple wave sources or scattering obstacles. In this work, inspired by the accommodation of human eyes, we propose the concept of supervised evolving learning (SEL) and demonstrate an SEL-driven adaptive focusing (SELAF) neuro-metasurface. Thereafter, the control circuit applies the generated decisions to the metasurface to adjust the states of the diodes, thereby changing the entire transmitted EM spectrum for subsequent focusing (as shown in the bottom half of the diagram). Then, a focus controller generates voltage regulation decisions based on both the steering signals and prior knowledge.
The detected signals subsequently pass through the convolutional layers and fully connected layers of a focus steering network and are thus converted into focus steering signals. Analogously, a metasurface first modifies the transmissive phases and amplitudes of incident waves, and an array of monopole probe detects these EM waves. Then, the contraction of the ciliary muscles alters the focal distance, controlling the ability to focus nearer or farther images on the retina (as shown in the upper half of the diagram). Neural signals, originating from the oculomotor nerves, travel through the parasympathetic nervous system and travel via the short ciliary nerves. Through the optic nerve, the electrical signals stimulated by visible radiation are transmitted to the occipital lobe, where they are interpreted as vision 33. In that case, this will significantly simplify the design of the device and facilitate its application (Fig. Suppose a naturally intelligent focusing system can be designed to automatically converge the EM waves (light) in different environments, such as the human eye. Therefore, the visual system is highly adaptive and responsive to environmental changes 32, 33. In the eye, the lens primarily focuses the incident light onto the retina and achieves a wide focal range by adjusting its shape via accommodation 31. The human eye is a perfectly focused system more than 80% of the information about our surroundings is acquired through our eyes 30. Given these factors, a practical approach to avoid time-consuming searches and develop self-adaptability is highly sought after, although challenging. Deep learning may fail when confronted with rapidly changing environments during a focusing task.
Despite fruitful studies on adaptive optics by empowering artificial intelligence (deep learning) on metasurfaces 22, 23, 24, 25, 26, 27, 28, 29, the success of deep learning relies heavily on the quantity and quality of available training data and requires prior information about the environment. Therefore, an intelligent adaptive strategy is required. Most metalenses work only in predefined environments, which makes it challenging to cater to the ever-changing application scenarios and incident waves. In the past decade, we have witnessed numerous studies on metalenses using various metasurface structures with the goal of achieving broadband, achromatic performance, high efficiency, and more 14, 15, 16, 17, 18, 19, 20, 21. Compared to conventional bulky lenses, which specifically rely on the polished surface profile of transparent optical materials to attain the required gradual phase change, metalenses can focus incident light with more compact dimensions. Among many functional devices, a metalens is one of the most thought-provoking. By carefully designing a subwavelength element and spatiotemporal layout, many attractive optical responses can be customized at will 7, 8, 9, 10, 11, 12, 13. Metasurfaces are artificially engineered structures composed of subwavelength scatterers with periodic or quasi-periodic arrangements 3, 4, 5, 6. Significant effort has been made with the advent of metasurfaces to miniaturize and integrate optical lenses. They are typically manufactured on various bulky substrates, generating user-defined beam intensity profiles or beam shapes by rationally designing microstructures 2. For example, digital cameras perceive image information by relying on in-built complex optical lenses systems 1. Optical focusing lenses are ubiquitous in our daily lives and modern optical laboratories. Light focusing has been a long-standing research topic for centuries.
0 kommentar(er)
