System Design – ND Wireless Institute
Skip to content

RadioHound: Distributed Spectrum Sensing

The heart of the RadioHound System is a low-cost, pervasive, persistent spectrum sensor that leverages low-cost SDRs suitable for deployment in mobile environments. RadioHound maps the entire spectrum from 25 MHz to 6 GHz, and will have the capability to support mmWave bands. RadioHound enables spectrum mapping in real time, providing a reliable indicator of spectrum utilization. It’s beacon capabilities provide a mechanism for measuring radio signal propagation between devices. Because of its portability and low cost, widespread deployment can give a good understanding of device-to-device and device-to-infrastructure connectivity, especially in crowded or dense environments.

The widespread deployment of RadioHound enables understanding of Wireless network architectures in new frequency bands, spectrum sharing possibilities and algorithm testing, and spectral efficiency and utilization measurements across large geographic areas. The overall cost of the RadioHound Node and the low-cost SDR (a commercial DVB-T receiver), but without the three antennas for the sub-GHz range or the mmWave front-end MMIC, is around $70. The eventual target is a sensor costing less than $10.

RadioHound System

The RadioHound System comprises three components: the RadioHound Sensor (Hardware), the RadioHound Cloud and the RadioHound App (Software). The hardware sensor includes a tri-band RF front end (including a tunable wideband receiver, tunable narrow-band beacon transmitters, and transverses for specialized frequency bands) and three antennas for the different bands: 25-100 MHz, 0.1-0.4 GHz, and 0.4-6 GHz respectively. The external operating environment of RadioHound requires a host to communicate with and allow data collection, management, and transmission to the RadioHound Cloud. The host provides geo-location capability, storage, and transmission to the central database, and can have different form factors, with the Raspberry Pi being the simplest and most nomadic. The RadioHound sensor derives its power from the host. The power consumption in the current Phase 1 (proof of concept) is 3W, with an envisioned target of 0.5W in future phases.

The RadioHound Cloud involves the scalable partitioning of functions into the following separate server instances: data warehouse for storage and retrieval of records, check in server for authorization of nodes and allows for future provisions for security. The RadioHound app provides two functions: send commands to the RadioHound sensors and enhanced visualization of the data stored in the Cloud through a GUI web app. The system software also includes a master controller that sends jobs to nodes and a state machine for data flow control amongst other functions.

Sponsors and Researchers

The RadioHound project was started in July 2015 and is sponsored by InterDigital and Nokia (previously Alcatel Lucent) via the BWAC I/UCRC, an NSF program. Currently, the funding is committed through Phase 3 towards the proof of concept. The future phases will focus on reducing the cost and power needs of the sensors, developing a cost-scalable Millimeter Wave MMIC, and the evolution of the software.

RadioHound is a Wireless Institute multi-PI project, with the hardware components developed by Jonathan Chisum and Bertrand Hochwald (Lead PI), and the software components by J. Nicholas Laneman and Aaron Striegel.

Some of the graduate students currently involved include Nik Kleber, Abbas Termos, Arash Ebadi Shahrivar, Lihua Wan, and Xiang Bo.

Related Publications

“RadioHound: A pervasive sensing network for sub-6 GHz dynamic spectrum monitoring,” https://arxiv.org/abs/1610.06212.

Distributed Spectrum Sensing

Low Complexity, Low Cost, Low Power RF

One of the key challenges of millimeter wave systems is the increased path loss at high frequencies compared with current sub-6 GHz wireless networks. A proposed solution is to employ high-gain antenna arrays which compensate for the path loss but have a corresponding narrow beamwidth. In order to maintain a wireless link with a mobile device, beam-steering is employed. Traditional beam-steering uses active phased arrays which incur a high cost and power dissipation. An alternative to active phased array beam steering antennas is the passive Luneberg lens which boasts zero power dissipation, low-loss, and high gain. However, the Luneberg lens is a 3D (spherical) gradient index lens traditionally requiring elaborate machining of concentric shells of dielectric. Due to the difficulties of fabrication this lens has traditionally only been realized in bands below 20 GHz.

Recently, the method of transformation optics has been presented as a means of physically distorting an electromagnetic structure and maintaining its functionality by spatially varying the permittivity throughout the structure. This has enabled the design of flat lenses but they require so-called gradient index optics. Over the past year, we have developed a process for manufacturing gradient index optics targeting the millimeter wave bands from 30-150 GHz. The approach is known as perforated media.

Crowded Spectrum: Dynamic Sensing and Sharing

Efficient use of spectrum is key to allowing government and commercial wireless technologies to flourish in the future. The RF spectrum below 6 GHz, in particular, has become extremely valuable as wireless communications technologies supplant wired solutions. As everyone looks for “1000x” improvements over the existing generation of wireless communications, one of the main challenges has become how to better utilize frequencies at 6 GHz and below.

Notre Dame’s Wireless Institute is actively engaged in researching and developing system models, sensing algorithms, and signaling schemes for utilizing spectrum more efficiently. The vision is to create localized spectrum efficiency maps, possibly in coordination with national geo-location databases, to enable wireless devices to dynamically find under-utilized spectrum, while minimizing interference.

To accomplish this, the RadioHound project has been initiated to develop a new spectrum efficiency monitoring platform – a system with a network of high-resolution, low-cost distributed overlay and compare local, real-time spectrum usage and efficiency to the FCC/NTIA radio spectrum sensors, a central data store, and an interface for analytics and visualization that could frequency allocation chart. The spectrum monitors would also be sufficiently sensitive, portable, programmable and low cost to provide an invaluable resource for mapping out spectrum efficiency across the United States. A complete understanding of the space-time efficiency utilization of spectrum is the key to improving it and will enable the regulated sharing of any inefficiently utilized spectrum resources.

Distributed Spectrum Sensing

802.11ax: AN Uplink MU-MIMO SDR Testbed

The proliferation of smartphones and growth of user-generated content has led to substantial increase in uplink traffic in use by mobile devices. Multi-input multi-output (MIMO) technology has been successfully adopted in past 802.11 standards, leading to high throughput and robust performance. Downlink MU-MIMO (DL MU-MIMO) was introduced in 802.11ac, where an access point (AP) with multiple transmit antennas is able to transmit multiple data streams to multiple spatially distributed stations (STAs) at the same time. To address the challenge of burgeoning uplink traffic is one of the primary objectives of the new IEEE 802.11ax. Uplink multi-user MIMO (UL MU-MIMO) can improve uplink capacity by enabling multiple spatially separated clients to access the channel at the same time and is especially useful in scenarios where the STAs have limited number of antennas (e.g. typical smartphones). The focus of this project is to build a software-defined radio (SDR) testbed for prototyping UL MU-MIMO and studying related design trade-offs. The project uses the Wireless Institute’s Software Defined Radio (SDR) Lab, namely the National Instruments (NI) USRP RIOs, the LabVIEW Communications System Design Suite, and NI’s 802.11 Application Framework.

The project is anticipated to have several phases. As the first step, we have already implemented two algorithms for UL MU-MIMO: Carrier Frequency Offset (CFO) correction at STAs, and multiuser detection (MUD) of users at the AP. The prototyping setup includes two NI USRP RIOs associated with their host computers, where the AP is implemented in one USRP and two stations are implemented in the second USRP. In comparison with the SISO system, we are able to get doubled system throughput. In the next step, we plan to add transmit power control functionality to the current framework to study the UL performance for the situations in which distances of stations to the AP are not similar. In addition, we will study the effect of asynchronous UL MU transmissions on the system performance. 

The 802.11ax project was started in 2015 and is sponsored by InterDigital and National Instruments via the BWAC I/UCRC, an NSF program. J. Nicholas Laneman is the PI for this project.