Papyrus is a software platform that implements frequency agile wireless transceivers using SDRs. Papyrus addresses two key issues at the physical layer. First, SDR devices must be able to quickly sense specific spectrum ranges that are unused and available. Second, available spectrum is likely to be in fragments of all sizes. To support bandwidth hungry applications, SDR devices must be able to combine multiple spectrum fragments into single channels for data transmissions. On top of these two issues, SDR devices must also be able to operate in a decentralized manner without relying on any central control.
Non-contiguous Frequency Access via Decentralized OFDMA
The spectrum band is divided into a large set of frequency subcarriers. Each Papyrus sender can transmit on any subset of the subcarriers, either contiguously or non-contiguously aligned in frequency. Multiple transmissions can occur simultaneously on isolated frequency subcarriers without mutual interference. Papyrus offers this level of flexible frequency access by implementing decentralized orthogonal frequency division multiple access (OFDMA). OFDMA is a multi-user version of the orthogonal frequency-division multiplexing, a digital modulation scheme. By dividing the frequency band into a large set of subcarriers and letting transmissions operate on isolated subset of subcarriers, OFDMA allows several conflicting users to transmit simultaneously.
Figure 1. Papyrus enables non-contiguous frequency access via decentralized OFDMA
Implementing OFDMA on distributed networks, however, is difficult because of several significant challenges. Existing designs in centralized networks rely on global synchronization to maintain subcarrier orthogonality, so that transmissions on different subcarriers do not interfere with each other. In distributed networks, where global synchronization is infeasible, this subcarrier orthogonality no longer exists, and transmissions interfere with each other. Papyrus overcomes this problem using advanced signal processing techniques, essentially reengineer receivers to “filter” out unwanted signals and restoring the desired transmission properties.
Usable Frequency Detection
When sharing spectrum with peers, wireless devices must sense the spectrum accurately and quickly to identify locally usable frequency. Papyrus configures its OFDMA-based transceiver to listen to the entire spectrum span supported by the underlying radio hardware and produce a power spectral density (PSD) map that measures the received energy level across the frequency band. Papyrus then uses the PSD map to identify usable frequency by detecting “busy” subcarriers.
Papyrus exploits a unique property of radio transmissions in the frequency domain. Regardless of energy levels, all active transmissions display a pair of rising and falling edges in the PSD map. By identifying these edges, Papyrus can reliably identify active transmissions and usable subcarriers. To do so, Papyrus first applies preprocessing to smooth the PSD map and filter out most noise before trying to locate edges. Papyrus then applies a search-based edge detection mechanism similar to those used in image processing. A frequency block with a rising edge to its left and a falling edge to its right is busy, and the rest are free. While this new method still requires a threshold to detect edges, our experimental results confirm that its sensitivity to both noise and threshold choice is much smaller than that of the energy detector.
Figure 2: A sample PSD map and its first order derivative. Papyrus identifies occupied frequency blocks using edge detection. While the absolute signal strength varies significantly across the frequency, the rising/falling edges are easier to detect.