Search for: All records

The growing quantity of wireless network activity generated every second of every day creates challenges for network operators, such as detecting anomalies and providing sufficient capacity. This same network activity also creates opportunities for Smart and Connected Systems (SCSs) to adapt to changing population dynamics, detect and proactively adapt to unexpected events such as public safety threats, traffic jams, or adverse weather events, for example. The GHOST project is researching the challenges of modeling, analyzing, and generating patterns of network activity. The GHOST project has demonstrated that Nonnegative Matrix Factorization (NMF) provides a robust mechanism for modeling network activity patterns that can be used to generate realistic network activity. The GHOST team has further demonstrated the capability for injecting programmed activity patterns into a live, functioning wireless network. 

Since the advent of mobile communication, the growth in demand for wireless communication devices and associated spectrum needs has been unstoppable. As a result, due to limited spectrum availability and historically inefficient management of assigned frequencies, spectrum sharing has steadily grown in importance and become a necessary solution to various capacity constraints. To support new developments in spectrum sharing, research in spectrum monitoring and spectrum utilization have become most valuable. GNU Radio offers a compelling opportunity to quickly develop and prototype new research in spectrum monitoring, sharing, and related radio frequency research that can support future deployments. GNU Radio’s packaged capabilities combined with its compatibility with a multitude of Software Defined Radio (SDR) hardware OEMs allow spectrum sharing research to be conducted nimbly and rapidly. To improve spectrum sharing and management, this research used GNU Radio in conjunction with Ettus USRP SDRs to collect I/Q data across the CU Boulder campus in regular intervals over 4 weeks, to monitor changes in the power levels recorded across 1 indoor and 10 outdoor locations. The results show that a simple sensor consisting of an SDR and a Raspberry Pi is capable of tracking changes in Wi-Fi signal strengths measured in outdoor environments. With calibration and careful hardware design such a platform could also be used for broader spectrum monitoring applications. 

Abstract—Demand for wireless communication devices has been growing continuously since the advent of mobile communication. Even though spectral efficiency and throughput keep increasing, consumer demand continues to seemingly outpace that growth. Spectrum sharing is becoming a more attractive solution to solving various capacity constraints as the resulting perceived spectrum scarcity can mostly be attributed to inefficient spectrum management. However, increasingly complex sharing arrangements come with an increased risk of interference. This makes it necessary to address such events in a timely manner. At the same time, research into using machine learning for solving issues such as signal classification, decision-making processes, and anomaly detection in wireless communication has been growing. To support machine learning research in anomaly detection for wireless communications, this research uses IQ data to train two autoencoders for anomaly detection in shared spectrum: a Long Short-Term Memory (LSTM) and a Deep Autoencoder. These algorithms are used to successfully identify anomalies in the time and frequency domain of recorded IQ data in the form of unauthorized LTE transmissions on top of Wi-Fi communication. 

As radio spectrum becomes increasingly scarce, coexistence and bidirectional sharing between active and passive systems becomes a crucial target. In the past, spectrum regulations conferred radio astronomy a status on par with active services, thereby protecting their extreme sensitivity against any harmful interference. However, passive systems are likely to lose exclusive allocations as capacity constraints for active systems increase. The resulting increase in ambient radio frequency noise from various terrestrial and non-terrestrial emitters can only be mitigated with informed collaboration between active and passive users. While coexistence using time-division spectrum access has been proposed in the past, a more dynamic approach following the CBRS sharing principle promises greater spectral occupancy and efficiency, enabled by a spectrum access system capable of constantly monitoring the ambient RF environment. Instead of simply minimizing the potential for any ”harmful” interference to passive users, the goal is to use good engineering to enable sharing between active and passive users. To this end, this research created a Software Defined Radio (SDR)-based testbed at the the Hat Creek Radio Observatory to quantitatively characterize the radio-frequency environment, and flag potential sources of radio frequency interference in the vicinity of the Allen Telescope Array. Sensor validation was carried out via data analysis of I/Q data collected in well-characterized RF bands. Results so far from ground and drone-based surveys are consistent with the expected sources of interference, based on both the deployment of static RF transmitters in the Hat Creek/Redding area as well as the interference detected in telescope observations themselves.