A critical issue in optical chaos-based communications is the possibility to identify the parameters of the chaotic emitter and, hence, to break the security. In this paper, we study theoretically the identification of a chaotic emitter that consists of a semiconductor laser with an optical feedback. The identification of a critical security parameter, the external-cavity round-trip time (the time delay in the laser dynamics), is performed using both the auto-correlation function and delayed mutual information methods applied to the chaotic time-series. The influence on the time-delay identification of the experimentally tunable parameters, i.e., the feedback rate, the pumping current, and the time-delay value, is carefully studied. We show that difficult time-delay-identification scenarios strongly depend on the time-scales of the system dynamics as it undergoes a route to chaos, in particular on how close the relaxation oscillation period is from the external-cavity round-trip time.

We investigate theoretically the possibility of retrieving the value of the time delay of a semiconductor laser with an external optical feedback from the analysis of its intensity time series. When the feedback rate is moderate and the injection current set such that the laser relaxation-oscillation period is close to the delay, then the time-delay identification becomes extremely difficult, thus improving the security of chaos-based communications using external-cavity lasers.

Optics InfoBase is the Optical Society's online library for flagship journals, partnered and copublished journals, and recent proceedings from OSA conferences.

Terahertz sparse deconvolution based on an iterative shrinkage algorithm is presented in this study to characterize multilayered structures. With an upsampling approach, sparse deconvolution with superresolution is developed to overcome the time resolution limited by the sampling period in the measurement and increase the precision of the estimation of echo arrival times. A simple but effective time-domain model for describing the temporal pulse spreading due to the frequency-dependent loss is also designed and introduced into the algorithm, which greatly improves the performance of sparse deconvolution in processing time-varying pulses during the propagation of terahertz waves in materials. Numerical simulations and experimental measurements verify the algorithms and show that sparse deconvolution can be considered as an effective tool for terahertz nondestructive characterization of multilayered structures.

We propose a secure communication scheme based on anticipating synchronization of two chaotic laser diodes, one subject to incoherent optical feedback and the other to incoherent optical injection. This scheme does not require fine tuning of the optical frequencies of both lasers as is the case for other schemes based on chaotic laser diodes subject to coherent optical feedback and injection. Our secure communication scheme is therefore attractive for experimental investigation.