A System Level Solution for DSA Systems: From Low-Cost Sensing to Spectrum Database

Abstract

In this paper, we present and implement a low-cost yet effective architecture that enables dynamic spectrum access (DSA) for any type of network. Our intention is to break the cost-scalability barrier and show that a complete system level solution for a database-assisted DSA system can be implemented with standard servers and inexpensive software configurable RF chips, thereby achieving economics of scale. First, we present the overall architecture that is capable of providing networks of any size to perform in-band and out of band channel access in a dynamic manner. The two main components of this architecture are the received signal strength indicator (RSSI) sensing network and the DSA server. For the RSSI sensing network, we built wired and wireless spectrum sensors that operate on 280-930 MHz using low-cost off the shelf software configurable RF (SCRF) chips. To get the RSSI values on a set of bands, we use generic micro-controllers to program the operating parameters (scan range, center frequency, bandwidth resolution, demodulation scheme and scan rate) of the SCRF chips. The wireless sensors transmit the sensed RSSI values to the nearest Ethernet-enabled hub using a light-weight communication protocol. The hub aggregates the data from multiple sensors and streams to the DSA server using UDP over IP. On receiving the real-time RSSI values from various sensors, the DSA server stores them in database engine with other meta data. Entries from the database are used by the channel allocation service that finds the best channel for the inquiring DSA nodes. To demonstrate the efficiency of the implemented database-assisted DSA system, we compare it to a legacy system and show the benefits in terms of resilience to jamming, channel relinquishment on primary arrival, and best channel determination and allocation. We also show the performance gains in terms of frame error rate (FER) and spectral efficiency. Finally, we compare the RSSI sensitivity of the low-cost sensors to that of a professional spectrum analyzer.