Single Sensor Radio Scene Analysis for Packet Based Radio Signals Using 2nd and 4th Order Statistics

Invention Summary:

Most existing wireless systems use static allocation of the radio spectrum with each system operating in a predefined fixed frequency band. Since the allocated spectrum is not used at all locations and at all times, this static allocation leads to creation of spectrum holes, which implies inefficient utilization of the available radio spectrum. Dynamic spectrum allocation offers a much more optimal spectrum utilization scheme. However, dynamic systems must have a completely new feature embedded in them, spectrum sensing capability. Spectrum sensing can be defined as RF signal analysis with a goal to determine if an observed frequency band is occupied, and to possibly identify or characterize the signals present in the band. Most of the research in spectrum sensing relates to the emerging 802.22 standard, where the goal is to detect the presence of a primary user, which in this case is the digital TV signals in the 400-800 MHz frequency band.

Researchers at Rutgers University have devised a statistical algorithm which computes estimates of the spectra and activity sequences by optimizing certain criteria under application specific constraints. The algorithm uses RF signal analysis where one sensing node observes a frequency band possibly used by multiple packet based radio transmitters. Analysis of the received signal consists of two steps. 1) a spectrogram is used to perform temporal segmentation of the received nonstationary signal. This task is formulated as a clustering problem, and 2) A 2-D slice of the fourth order spectrum is computed for each of the segments found in the first step. These fourth order spectrum slices are arraigned in a three-dimensional array.

The goal of the second step is to use unique properties of the low rank decomposition of the three-dimensional array to recover spectra and associated activity sequences of individual components in the received signal. A numerical algorithm for the low rank decomposition is derived and estimates of the spectra and activity sequences are computed by optimizing certain weighted least squares criterion under application specific constraints. The approach is tested using simulation examples involving signals in the 802.11a/b/g and bluetooth networks.

The proposed algorithm can be used as a spectrum analysis tool, providing crucial information needed for achieving efficient utilization of radio spectrum and eliminating mutual interference between the coexisting systems.

Market Application:

Radio spectrum allocation.


  • Possible to identify spectra and activity sequences of individual signals forming the received signal
  • Method is blind in a sense that the knowledge of signal pulse shapes and their center frequencies is not assumed

Intellectual Property & Development Status:

Utility patent pending

Patent Information:
For Information, Contact:
Dan Benderly
Licensing Manager
Rutgers University