Skip to main content
Lasers

Lasers

Chaos starts to communicate

01 Apr 1998

A laser is a device that exemplifies the notions of stability, coherence and purity. Yet it is now well known that the behaviour of a laser can easily become unstable and chaotic. While most research has been devoted to eliminating such behaviour and establishing lasers as uncontaminated light sources, the 1990s have seen increasing interest in the idea of using chaotic laser signals in communications.

Gregory VanWiggeren and Rajarshi Roy of the Georgia Institute of Technology in the US have now used chaotic lasers to generate the complex signals of the type needed in many of today’s communications systems (G D Van Wiggeren and R Roy 1998 Science 279 1198). The chaotic laser system can be used to synchronize a transmitter and receiver, and to transmit encrypted data.

So what made optical scientists decide to exploit, rather than prevent, chaos in lasers? To understand this goal, we must first recognize that chaos can be useful. Although the dynamics of chaotic systems are continually unstable, they are also bounded. This causes the variables in the system to oscillate in a noisy but deterministic way. It is therefore difficult – or in some cases impossible – to predict the future states of the system.

The result is that the signals produced by a chaotic system are highly complex and contain a broad range of frequencies. Such signals are similar to those used in “spread-spectrum” communications and related fields, in which information is hidden within noisy signals. Current methods for generating chaotic signals generally use simple electrical circuitry to create the background noise, but it is difficult to push the circuitry to the high frequencies needed for many communication channels. But a physical system, such as a laser, produces the noise automatically.

The potential of chaotic signals in communications was first investigated in 1990, when Thomas Carroll and I at the US Naval Research Laboratory used chaotic circuitry to synchronize transmitters and receivers, and to send hidden or encrypted messages. In 1993 Kevin Cuomo and Alan Oppenheim of the Massachusetts Institute of Technology published the first details of a successful transmitter-receiver circuit.

There have been two main problems with this approach. First, most of the circuitry has been in the audio range or just above, which limits the rate at which messages can be transmitted to between 0 and 10 kHz. In principle, there is no reason why radio-frequency (rf) circuitry should not be used, with frequencies in the megahertz or even gigahertz range, but this would add an extra element of complexity.

Another problem is that almost all of the chaotic systems used so far have been low-dimensional. In other words, the number of variables needed to describe the system, such as currents and voltages, have been less than 10, and often less than 5. Many communications schemes require more complex signals with higher dimensions.

The chaotic laser system developed by VanWiggeren and Roy overcomes both of these problems. The chaotic signals from the system are around 100 MHz, which yields a data rate comparable with that used in rf communications. Perhaps more importantly, the method used to generate the chaos in the laser system can lead to high-dimension laser dynamics. The actual dimension is not yet accurately known, but it could be 50 or more. Mathew Kennel and Henry Abarbanel of the University of California at San Diego (UCSD), who originally showed that such a laser system could be used for synchronization, are now calculating the dimension of the system.

At the heart of the laser system is a transmitter consisting of a ring laser made from erbium-doped fibre. An optical signal is generated by an erbium-doped fibre amplifier (EDFA), and is reinjected into the EDFA after circling the ring once. This means that the laser is driven by its own output but at some time delay, which leads to chaotic and high-dimensional behaviour. This type of response is common to time-delayed dynamical systems of any kind.

The message to be transmitted is converted into an optical signal with another EDFA. This optical information is coupled into the fibre ring of the transmitter, and is injected into the laser together with the time-delayed laser signal. This means that the information signal also drives the laser and so becomes mixed with the dynamics of the whole transmitter. Such an information-dynamical mixing technique was first suggested by Alexander Volkovskii and Nikolai Rulkov at UCSD.

As the combined information/laser signal travels around the transmitter ring, part of it is extracted and transmitted to the receiver. At the receiver the signal is split into two. One part is fed into an EDFA almost identical to the one in the transmitter, which ensures that the signal is synchronized with the dynamics of the ring-fibre laser in the transmitter. It is then converted into an electrical signal by a photodiode, providing a duplicate of the pure laser signal at some time delay. The other part is fed directly into another photodiode, which provides a duplicate of the laser-plus-information signal. After taking account of the time delays, the chaotic laser signal can be subtracted from the signal containing the information, removing the chaos and leaving the initial message.

The information signal in this experiment was kept simple – a pure square wave with a period of about 100 ns. No structure is observed when the transmitted signal is plotted in three-dimensional phase space, consistent with the high-dimensional nature of the chaos. The signal recovered by the receiver matches the transmitted signal reasonably well. Although the extracted signal does not reproduce the fine structure, the receiver can easily detect the 1s or 0s in a typical digital signal.

Several issues need to be resolved before such laser systems can be engineered into communications systems. The first is the information rate. Although these experiments have proved that data rates in the rf range can be achieved, higher data rates will be needed to compete with current communication schemes. In more recent work, the researchers have increased the information rate to 150 MHz.

The main bottleneck seems to be the use of electro-optical components to convert optical signals to electrical ones. If all of the signal processing could be performed optically, with no limitation set by the detection electronics, the technique would only be constrained by the dynamics of the laser system. In principle, the method could be extended into at least the gigahertz range.

Another problem with simply adding a chaotic signal to the information is that it is possible to identify information signals with a distinctive frequency content. For example, the periodic square wave used in the experiments appears as a spike in the spectrum of the transmitted signal. An intruder would be able to filter this frequency and isolate the information signal, although some typical filters applied by VanWiggeren and Roy were unable to achieve this. This problem could be overcome by mixing the information and chaos in some nonlinear way, and the researchers say that some work with nonlinear mixing has been successful. This work is sure to continue.

Despite the remaining problems, the experiments show that physically chaotic systems, rather than chaotic circuits, could overcome the problems that have stymied researchers and engineers in this field. Indeed, this might be the best message of this new research. Many materials and optical systems are nonlinear and can display chaotic behaviour, suggesting that an abundance of chaotic sources are available. The only problem is finding suitable chaotic systems, and VanWiggeren and Roy have provided us with an impetus to do just that.

Related events

Copyright © 2024 by IOP Publishing Ltd and individual contributors