| In this paper, a comprehensive evaluation of time-frequency analysis techniques is presented as applied to the detection of finite duration communication signals. This class of signals includes both phase and amplitude modulated, pulsed signals. Time-frequency analysis based on quadratic distributions are of particular importance in analyzing these signals since this class covers the most common types of time-frequency analysis techniques including the standard, windowed spectrogram as well as more general types, e.g., the Wigner-Ville distribution and the B-distribution introduced by Barkat and Boashash. Typical performance criteria used in the design of time-frequency distributions include high frequency resolution as well as suppression of cross modulation products that are present in a multi-component signal environment. In addition to these criteria, we are also interested in the time resolution properties of these algorithms. An interesting feature of the B-distribution is the sharpness parameter (power coefficient) that controls the time-frequency resolution. In this paper, we will evaluate the performance sensitivity of the B-distribution to this parameter with respect to short-time signal feature analysis. We will also include in this analysis, subband methods based on polyphase Fourier transforms. Combining subband decomposition with the discrete wavelet transform will also be evaluated as another tool for detecting and analyzing finite duration communication signals. The ultimate goal of this work is to determine the best time-frequency algorithms for a given application based on both performance and computational considerations. |
| Edgar Satorius. Principal Member of Technical Staff, Flight Communications Systems Section of the Jet Propulsion Laboratory, Pasadena California. Technical Manager and lead analyst for US Navy research in the areas of adaptive modulation and channel modeling.
Summary Credentials: Ph.D. Elec. Eng. (1975), California Institute of Technology. Naval Ocean Systems Center, 1975 – 1979, Dynamics Technology 1979 – 1983, Gould NavCom 1983 – 1986, Jet Propulsion Laboratory, 1986 – present.
Relevant Research and Experience: Dr. Satorius has over 90 publications and two patents in the field of signal processing and is an Adjunct Associate Professor at the University of Southern California where he teaches digital signal processing courses. He has extensive experience in the general fields of adaptive signal processing and communications systems. He performs systems analysis in the development of digital signal processing and communications systems with specific applications to blind demodulation; digital direction finding and digital receivers. Specific programs have included the development of novel signal processing algorithms in support of NASAs Mars Exploration Rover program; FPGA-based IF equalizers for the Air Force’s Advanced Range Telemetry (ARTM) program and high speed, FPGA-based receiver architectures in support of JPL’s Electra digital transceiver. |