Diane G. Mills

Recent advances in cognitive communications

This article describes recent advances in cognitive communications. We combine the concepts of signal processing, communications, pattern classification, and machine learning to make dynamic use of the spectrum, such that the emanated signals do not interfere with the existing ones. Unlike other programs such as neXt Generation communications of the Defense Advanced Research Projects Agency, where radio scene analysis is performed to find the spectrum holes or the white space, we make use of the white, as well as the gray space for non- interfering signal transmission. We examine the possibility of employing machine perception and autonomous machine learning technologies to the autonomous design and analysis of air interfaces. The underlying premise is that a learning module will facilitate adaptation in the standard classification process so that the presence of new types of waveforms can be detected, features that best facilitate classification of the previously and newly identified signals can be determined, and waveforms can be generated by using the basis-set orthogonal to the ones present in the environment. Incremental learning and prediction allows knowledge enhancement as more snapshots of data are processed, resulting in improved decisions. Some of the contributions of this project include technological advances in signal detection, feature identification, signal classification, sub-space tracking, adaptive waveform design, machine learning, and prediction.

Machine Learning based Cognitive Communications in White as Well as the Gray Space

This paper describes new ideas and results on machine learning based cognitive communications in White as well as the Gray space. We combine the concepts of signal processing, communications, pattern classification and machine learning to make a dynamic use of the spectrum such that the emanated signals do not interfere with the existing ones. Unlike other programs such as the neXt Generation (XG) communications program of the Defense Advanced Research Projects Agency (DARPA), where radio scene analysis is carried out to find the spectrum holes also known as the White space, we make use of the White as well as the Gray space for non-interfering signal transmission. Our assumption is that a learning module will facilitate adaptation in the signal classification process, so that the presence of new types of waveforms can be detected, features that best facilitate classification of the previously and newly identified signals can be determined, and waveforms can be generated by using the basis-set orthogonal to the ones present in the environment. Incremental learning and prediction allow knowledge enhancement as more snap-shots of data are processed, resulting in improved decisions. Use of non-competitive policy set results in zero interference with the already existing signals with a modest increase in the White and Gray space utilization. On the other hand competitive policy set utilizes machine learning to predict the future behavior of the signals which results in more than 90% utilization of spectrum at an expense of some interference due to errors in prediction.

A multiple access differential frequency hopping system

Military communication systems require waveforms that are resilient in the presence of jamming signals (i.e. are antijam (AJ)), that have a low probability of detection (LPD) by unintended receivers, that have a low probability of intercept (LPI) by hostile receivers, and that operate well in the presence of many authorized users. Current spread spectrum communication techniques such as standard frequency hopping (FH) and direct sequence spread spectrum (DSSS) do not provide signals that simultaneously exhibit AJ, LPD, and LPI properties, allow conferencing, and are easily implemented. A newly proposed multiple access differential frequency hopping (DFH) communications system demonstrates improved AJ and LPD performance when compared to existing spread spectrum techniques. In addition, it successfully detects and decodes multiple (interfering) users in an architecture that is ad hoc and is easily implemented. Preliminary analysis and simulation results demonstrating the improved performance are provided.

A performance comparison of differential frequency hopping and fast frequency hopping

This paper discusses the multiple access performance of the spread spectrum modulation called differential frequency hopping (DFH) and its associated signal processing algorithms, and compares DFH performance to the performance of fast frequency hopping (FFH) multiple access systems. To fairly compare FFH and DFH, it is necessary to carefully choose the parameters for each system, using the concept of efficiency. The single-user efficiency provides a fair way to set parameters when comparing FFH and DFH systems when there is no multiuser interference (MUI) present. We develop the multi-user efficiencies for the two waveforms, as a mechanism to determine comparable system parameters for FFH and DFH multiple-access systems. It is shown that, for the same bits/symbol, the data rate of a single-user DFH is larger than the FFH data rate if the hop dwell time is the same DFH is shown to be significantly more MUI-tolerant than the comparable FFH and FFH systems with interference cancellation, particularly at low and mid-level SNRs. Similarly, the theoretical same-efficiency DFH outperforms the non-optimal FFH at most SNRs and outperforms optimal FFH with collision resolution at SNRs above 7 to 11 dB, depending on the number of simultaneous users. The theoretical and simulation results consistently indicate that DFH is more MUI-tolerant than FFH for the same efficiency, which is desirable in ad hoc networks with multiple users.

Remotely deployed autonomous sensor/processing network for surveillance uses

Adding remote deployment capability to unattended surveillance systems provides the capability to unobtrusively monitor activities in hostile or neutral areas where in-situ placement is not practical. Targeted surveillance activities include tracking force movements, monitoring keep-out areas, and detecting the presence of assets hidden underground or in buildings. Key technologies to achieve this capability include localization of deployed nodes, field optimization and control techniques, multi- sensor fusion, and low cost miniaturized sensors.