Non Line of Sight

Non-line-of-sight (NLOS) imaging and communication aim to reconstruct scenes or track objects hidden from direct view by using indirect reflections or signals. Current research heavily utilizes machine learning, particularly convolutional neural networks (CNNs), autoencoders, and transformer-based architectures, often coupled with techniques like independent component analysis (ICA) and Kalman filtering, to improve accuracy and efficiency in diverse applications. These advancements are crucial for improving the reliability of positioning systems (e.g., GNSS) in challenging environments and enabling new capabilities in areas such as autonomous driving, search and rescue, and robotics.