Spectrum Observatory


TV White Space (TVWS) has become a new opportunity for unlicensed wireless networks also known as cognitive radios. However, it meets several challenges such as spatial and temporal variation due to the requirements to protect primary users and the nature of unlicensed bands. Two methods, geo-location database and spectrum sensing, are commonly used to find unused TV spectrum. UW EE's FuNLab has implemented a system named SpecObs that utilizes both techniques to improve white space prediction and to more accurately compute the protection region of primary transmitters and available capacity. This research will be extended to other frequency bands that are being opened for unlicensed operation.

The system consists of a spectrum sensor node and a server. Sensor nodes periodically sense broadband spectrum and detects occupancy. After that, sensing data is uploaded to the server tagged with other information such as location and sensing time. A number of sensor nodes can be deployed in various locations to cover a geographic area.

System Architecture
Fig 1. Overall System Architecture
System Components
Fig 2. Specobs System Components
Server Components
Fig 3. Specobs Server Components

The server supports two main functions. The first function manages and analyzes sensing data. The server reads uploaded files from sensor nodes and computes statistical characteristics of channel availability. Another function models the activity of each licensed owner of the spectrum in the form of a protected region and estimates white space availability. This is accomplished by using a path-loss model (currently the Longley-Rice propagation model in Area and Point to Point mode) and real terrain data. To achieve this, the server downloads technical specification from the FCC CDBS database and its computing engine updates the analyzed data on a daily basis.

The server also provides web-based GUI tools. Therefore, users are able to not only see protection regions of each primary transmitter, but also white space channels, predicted noise floor and capacity for each channel in their location. Also, overall available capacity is visualized in a color map for the U.S. region which highlights the dynamics of available throughput in different markets.

SpecObs Website
Fig 4. SpecObs Website
Capacity Color Map
Fig 5. Capacity Color Map

The SpecObs system can be associated with secondary wireless networks operating in TV bands. Hence, performance of these networks can be significantly improved by providing the best white space channels and precise network planning according to their locations. Future efforts will focus on developing enhanced detection algorithms with spectrum sensing and modeling white space availability in other bands.

Current Research