UCLA researchers have unveiled a new optical computing method that can replace traditional techniques for performing permutation operations. The proposed technique effectively performs an optical permutation leveraging only passive hardware, removing the need for advanced signal processing and potentially saving power.
The UCLA optical computing system works as an “over-the-air” permutation source, allowing messages to be secured for applications such as telecom or encryption. Image used courtesy of Laser and Photonics Reviews
Optical computing is not an inherently new concept, with researchers in academia and industry working toward more efficient and high-performance optical devices. The UCLA device, however, uses multiple layers of passive diffraction materials to encrypt a signal before transmission.
UCLA Team Optimizes Diffraction Features
The UCLA researchers designed multiple layers of diffraction materials, each optimized with different phase values to focus the light at distinct points. The resulting output beams appeared “shuffled” compared to the input, requiring an inverse operation before the data could be used.
In addition to its optimized diffraction materials, the device enabled each layer to be rotated, further increasing the number of permutation states available. As a result, incident light rays could be permuted without consuming any electronic computing resources.
The optimized diffraction materials (left) effectively created a predefined permutation matrix, allowing the receiver to use inverse operation for optical data permutation and recovery. Image used courtesy of Laser and Photonics Reviews
A similar concept exists in quantitative phase imaging, where image information is stored in the phase instead of the amplitude of the light rays. Applied to encryption, the optimized UCLA devices can not only shuffle data but also recover that same data by applying the inverse transform upon reception.
Shuffling Optical Data
A permutation operation takes an ordered set of data, like an image or bitstream, and shuffles its members into a new set of data. While the new set still contains the same information that can be easily recovered by applying an inverse permutation, it becomes extremely difficult to recover any useful information without knowing the details behind the original permutation.
The permutation operation “shuffles” input data according to a predefined permutation matrix. The original data still remains, but is secured against unauthorized access. Image used courtesy of Springer Nature
In an image, the permutation operation will effectively shuffle each individual pixel while remaining within the same bounds of the image. So, while each pixel remains at the same magnitude and color in a different location, the resulting image after permutation appears to contain no useful information. This is a basic form of encryption and allows secure messages to be sent between parties who know the details behind the permutation.
Traditionally electronic hardware or software achieve this permutation. In the case of software, the CPU can manually shuffle data given the permutation matrix, all the while consuming power and occupying a CPU core. Hardware-based permutation is much faster, but still takes some time to shuffle the input data. A method of performing a permutation in the optical domain, as the UCLA researchers pursued, can save power and speed up the operation.
The Growing Utility of Optical Computing
Although the UCLA device is still in its infancy, the preliminary results show potential use cases, from telecommunications to data security. Furthermore, if the inverse permutation can be applied in the optical domain as well, the UCLA diffraction-based permutation could enhance security as a physical encryption/decryption key pair.
While optical computers may not replace traditional computers in every setting, with the help of researchers like those at UCLA, optics-based computers are slowly becoming more applicable to engineering applications.
