Abstract: Disclosed is an apparatus for estimating the features of a radio frequency signal including a reference signal generator configured for generating a reference signal from a complex signal and one or more feature estimation units configured to generate one or more feature estimates from the reference signal and the complex signal. Techniques for creating reference signals from both signals spread by a scrambling code and FM signals are discussed.

Abstract: An apparatus and method for signal separation and SINR improvement is disclosed. In particular, disclosed is an iterative technique whereby a noisy signal, that may include signals from one or more distinct emitters, is first improved through an SINR improvement technique. Next, one signal from the noisy signal is identified and a reference signal substantially free of noise and interference is generated for this signal, and then this reference signal is subtracted from the noisy signal. This process is repeated until no other signals may be identified. In this way, the signal with the highest SINR may be identified first and then removed from the original signal. As such, the identified signal is no longer present as interference for the remaining signals, thus improving each of the remaining signals' respective SINRs.

Abstract: An apparatus for sensing and classifying radio communications is provided. The apparatus includes one or more sensor units configured to detect radio frequency signals, a signal classifier configured to classify the detected radio frequency signals into a classification, the classification including at least one known signal type and an unknown signal type, a clustering learning algorithm capable of finding clusters of common signals among the previously seen unknown signals. The system is then further configured to use these clusters to retrain the signal classifier to recognize these signals as a new signal type.

Abstract: A apparatus configured to detect the presence of one or more unknown signals in the presence of noise and/or interference. The apparatus includes an input configured to receive an input data from a sensor, an integrator configured to integrate the power of the input data over a period of time, and a comparator configured to compare the integrated power to a threshold to determine if the input data contains at least one signal, or if the input data contains noise only. The apparatus may further include a mean estimator configured to estimate the mean of the power integrated input data over a period of time, a variance estimator configured to estimate the variance of the power integrated input data over a period of time, and a threshold calculator configured to calculate the threshold value based on the estimated mean or estimate mean and estimated variance.

Abstract: An apparatus for identifying a specific emitter in the presence of noise and/or interference is disclosed. The apparatus includes a sensor configured to sense radio frequency signal data, the signal data containing noise and signal from at least one emitter, a reference estimation unit configured to estimate a reference signal relating to the signal transmitted by one emitter, a feature estimation unit configured to generate one or more estimates of one or more feature from the reference signal and the signal transmitted by that particular emitter, and an emitter identifier configured to identify the signal transmitted by that particular emitter as belonging to a specific device using one or more feature estimates. The emitter identifier identifies the signal transmitted by that particular emitter as belonging to a specific device using Gaussian Mixture Models and the Bayesian decision engine.