LGS Innovations CEO Kevin Kelly talks with MW&RF about the trends behind the increasingly crowded RF spectrum space, and what's being done to meet the demand.
|Download this article in .PDF format |
This file type includes high-resolution graphics and schematics when applicable.
The last 20 years have seen the massive adoption of mobile communications and data services by the global user community, a remarkable expansion in worldwide wireless network capacity, and a meteoric rise in the wireless equipment and enabling technology markets. This virtuous cycle has driven both technological innovation and a continuously increasing demand for a decidedly finite resource in this ecosystem: RF spectrum. Microwaves & RF talked with Kevin Kelly, CEO of LGS Innovations, about the evolving trends in this arena.
What are some of the specific challenges associated with today’s crowded RF spectrum?
The origins of today’s crowded RF spectrum are diverse, and pose real concerns. Spectrum is a finite resource, and not all frequencies are equally desirable. The most desirable frequency bands (VHF, UHF, and the low microwave bands) are already very crowded, and the explosive growth of cellular communications and the Internet of Things (IoT) are driving tremendous pressure for spectrum access.
At this month’s Consumer Electronics Show (CES), an estimated one of three displays pertained to IoT, with devices as diverse as automobiles, washing machines, and wearables featuring embedded wireless transmitters. Considering the possibility of nearly every person having multiple transmitters—many communicating nearly continuously and concurrently—it is easy to understand how quickly spectrum is becoming an even more valuable resource.
This resource needs to be utilized in an optimal manner in order to extract every bit of spectrum availability. It requires an approach that allows for the simultaneous dynamic sharing of this resource across the dimensions of frequency, time, physical layer code, multiple input, multiple output (MIMO) multiplexed channel, and/or geography.
More users accessing increasingly scarce spectrum yields an increase in overall RF noise and RF interferers. Nulling interferers and reducing noise becomes more difficult—both technically and from a policy perspective—in this rapidly evolving, heterogeneous spectral environment.
Can you explain some of the concerns of government agencies regarding spectrum usage?
Government agencies that once enjoyed seemingly limitless (and free) spectrum access must now learn to share bandwidth with industry and consumers. The challenges of sharing spectrum are clear, but not trivial. Sharing spectrum means that new RF interference will be present that could negatively impact an agency’s mission. Agency managers need to protect their ability to execute upon the mission while allowing other spectrum users to realize the value of their spectrum investment.
As the FCC, NTIA, and other regulatory bodies further outline government/industry spectrum-sharing opportunities and policies, government agencies will find themselves working with industry partners to establish a mutually beneficial partnership. Are their new partners viewing the government agency’s mission criticality through the same lens and operating in a manner that allows them to execute on their missions? How can they protect themselves if this is not the case? How will new types and levels of RF interference impact their systems? How can they tell who is interfering? Do they need to develop new internal expertise to monitor and manage RF interference?
Safe and effective spectrum sharing requires new systems and technologies. The government should be looking for technologies available to solve this problem, and how they can/should they procure this technology.
For our military services, compliance with U.S. spectrum allocations is only a small part of the problem. Military radars, communications, sensors, and weapon systems must maneuver within and around spectrum, which is allocated differently around the world.
For example, a Navy ship training off the U.S. coast must comply with U.S. spectrum allocations, but must also adapt to different allocations, laws, and policies (or the absence thereof) when operating in other parts of the world. A ship may operate in dozens of international spectrum-regulated areas on a single deployment, and may be limited or restricted in its use of sensors and systems that are essential for their missions and self-defense.
Moreover, these spectrum-management issues are mission-critical. Adversaries knowing our spectrum allocations or limitations can target their use of specific bands to detect, jam, spoof, or otherwise disable our defense systems.
Can you describe spectrum-management solutions?
Conventional spectrum management is generally a static process. Users are assigned exclusive rights to bands of frequencies to be employed in specific geographical regions. In this model, the main strategy for sharing spectrum is to partition the available spectrum into finer and finer spectral swaths. This forces industry to develop more efficient modulation schemes, where more bits of data can be transported per unit of frequency (hertz). Although many modulation techniques have been developed over the past 20 years, we are rapidly approaching the number of bits per hertz that can be transported efficiently.
Once allocations are set, spectrum management is composed of verification and enforcement of allocations in a rather manual fashion. One example of this could be a government agency employing a manual, human-intensive RF interference hunting approach using only the most basic tools, such as spectrum analyzers and large form-factor directional finding tools. Their goal would be to identify and locate violators of spectrum access agreements, and then apply appropriate regulatory methods to protect their spectrum from interference.
This approach does not scale well with large numbers of competing users interfering with each other, and would render reliable communication of data impractical.
A better, more sustainable solution for spectrum management should feature highly concurrent, dynamic, seamless, and at least semi-autonomous (if not fully self-governing) allocation of the spectrum for multiple competing users with different data-capacity and data-traffic requirements. The cognitive sensing of actual spectrum usage in real time will prove a key enabler of autonomous network awareness and bandwidth management.
Proactive and dynamic spectrum awareness and management approaches such as this will require three components:
• First, new methods are needed to monitor spectrum utilization with enhanced resolution over multiple resource dimensions (e.g., frequency, time, space, and signal analysis) in order to provide dynamic awareness of the RF environment.
• Next, dynamic spectrum access and monitoring techniques will enable multiple competing users to efficiently and safely exchange data.
• Third, techniques and algorithms are needed to facilitate multi-dimensional, autonomous resource allocation.
By applying these new methods for spectrum monitoring, dynamic spectrum access, and resource allocation, a scaled solution for spectrum management can avoid excessive “human-in-the-loop” requirements while allowing for enforcement of spectrum access agreements.
Tell us a little about the software that enables spectrum management.
Spectrum management covers a diverse set of applications and requires an equally diverse set of software technologies. To succeed in a crowded RF environment, a spectrum-management system will need to assess the behavior of relevant emitters. This involves use of signal processing to identify and characterize various RF sources; eliciting the frequency usage and temporal patterns of those sources; and perhaps demodulating and decoding the signals to extract information useful for understanding the environment.
The increasing complexity of RF systems requires spectrum-management systems become increasingly intelligent to detect and track the behavior of the systems emitting in the environment. This “pattern of life” analysis enables prediction of system behavior and determination of how best to achieve mission objectives. Rules-based processing, efficient database-management techniques, and machine-learning methods will be necessary to accomplish this goal.
In addition to monitoring emitters in the environment, many spectrum-management applications (i.e., aboard ship or aircraft) might carry responsibility for managing their own emitters. This can require scheduling algorithms, optimization techniques, and in some cases, collaboration with other systems in the environment. Generally, spectrum-management systems have to be adaptive, flexible, robust enough to react to unexpected changes in the environment, and able to work in real time.
Spectrum-management software solutions must be modular, scalable, and distributable, with the capacity and flexibility to support and integrate with small, embedded sensor systems as well as large, dispersed “systems of systems” involving thousands of complex sensors. Similarly, spectrum-management systems will need to interface with current and future standards-based infrastructure. Consequently, many systems-level features must be built into a spectrum-management solution, including the ability to tailor and adapt the user interface/user experience and ensure compliance with security and secure networking protocols and standards.
In terms of spectrum sharing, what challenges do you think will arise due to the IoT and the potentially massive number of wireless devices?
The IoT has been rightly touted as the next revolution in the mobile ecosystem, although it presents many challenges. As vendors demand lower-priced IoT components, the IoT will be populated by a massive number of low-cost, thin-margin, spectrum-using devices manufactured by vendors new to the wireless space. Many of these offerings will be anomalous, non-compliant devices that use out-of-spec frequencies, bandwidths, and power levels.
Various well-known protocols have been developed over the years to support IoT, but none of them have gained mass-market penetration. One limiting factor to their adoption has been a lack of ubiquitous coverage. Cellular networks provide adequate coverage, but the deployed protocols have proven to be inefficient for many IoT uses. As a result, power-consumption expenses are too high for many lower-power devices.
In recent years, 3rd Generation Partnership Project (3GPP) standards such as LTE-M (LTE modified for IoT) and Narrowband-IoT (NB-IoT) have been modified to support IoT devices more efficiently. Nationwide IoT networks utilizing LTE-M are just starting to be deployed, and will support a variety of services including utility meters, vending machines, etc. This will result in a massive number of devices sending small bursts of traffic, which will increase the overall utilization of the spectrum.
While these new cellular standards and technologies will expand the number of IoT devices, they won’t be the only protocols in use. Existing protocols will continue to be used where ubiquitous coverage is not needed, leveraging unlicensed spectrum in many cases. Regardless of the protocol, efficient spectrum sharing will keep being challenged by increased utilization of the spectrum, the distributive nature of the IoT devices, and more short bursts of traffic, which will make it harder to sense and predict the spectrum utilization. Advanced machine-learning algorithms will be required to detect the patterns in traffic and appropriately determine opportunities to share the spectrum.
What spectrum-management challenges do you expect to see as a result of 5G?
The introduction of 5G has the potential to revolutionize a number of industries by providing ultra-reliable, high-throughput, low-latency communication links. This will support a diverse set of applications, from autonomous vehicles to IoT network traffic.
To meet this vision, the network will have to be very dynamic. Multiple deployment paradigms will be supported, including licensed spectrum, shared spectrum, and unlicensed spectrum. Sharing and aggregation will be done both within the same band, as well as between bands having very different propagation characteristics, like sub-6-GHz and millimeter-wave.
The concepts pioneered in LTE, such as heterogeneous networks and inter-cell interference coordination, and the anticipated sharing between access and backhaul, will allow LTE deployment to reach full potential. Taken together, these advanced 4G features will create new and unique spectrum-management challenges.
During 5G network operation, not only will RF carriers be set up and torn down quickly, but the transmission formats used on those carriers will change rapidly in order to optimize waveforms to meet the application’s needs. This will require adaptive algorithms that can sense the spectrum and share it appropriately. To boost network capacity, 5G will also have to make extensive use of MIMO techniques to simultaneously send and receive multiple data signals over the same channel. This will make sensing the network’s spectrum use more challenging, due to the high directivity of MIMO signals. Sensing the network in one location may not reveal potential interference in another location.
5G will also support very short (<1 ms) transmission intervals and fast switching between uplink and downlink transmissions, which will create greater challenges in sensing the network. Traditional survey methods will be insufficient with 5G. Instead, a network of sensors will have to be utilized that continuously monitors the network in order to gain a thorough understanding of spectrum usage.
Lastly, what is your vision of the RF spectrum and spectrum sharing in the future?
Spectrum is a finite natural resource. Advanced wireless signal-processing and optimized dynamic-allocation techniques will continue to enable increasingly efficient use of this finite natural resource. However, vastly scaled user demand for both military and commercial applications such as IoT, machine-to-machine (M2M) communications, and advanced multimedia data services will outstrip the gains enabled by improved spectral efficiency alone.
To address the growing need and attendant financial opportunities posed by these applications, spectrum substrata must be allocated more narrowly and maximized through spatial system-level techniques such as very small cells with their extremely high density footprints.
To support 5G protocols, highly adaptive heterogeneous wireless networks and standards will be needed to achieve optimal spectrum utilization throughout the network.
Beyond these breakthroughs in making effective use of bandwidth, our ability to dynamically sense the wireless-network environment and adapt to it will be paramount. This cognitive-sensing component will be critical to the design of end-user devices and to the evolution of the wireless infrastructure itself. Only through cognitive sensing will we be able to monitor and enforce spectrum-usage rules, maintain our awareness of network anomalies and spectrum interference, enable spectrum sharing, and deploy the wireless data analytics needed to simultaneously serve many mission needs.
Prior to LGS Innovations, CEO Kevin Kelly held senior positions within General Dynamics Advanced Information Systems (GD-AIS) and Lockheed Martin. He has served in board and advisory positions with the LGS Innovations Board of Directors, Innovative Technologies Council of INSA, AFCEA, IEEE, the National Advisory Council for GWU SEAS, and several other firms. Kelly is also a member of the Engineering Hall of Fame at GWU SEAS. He holds a bachelor’s degree in electrical engineering from Penn State University and a master’s degree in engineering management/systems engineering from George Washington University.