Abstract: A non-linear signal classifier (26) includes a polynomial expansion unit for expanding signal feature vectors determined by a feature extraction unit (25) for a received signal. The expanded signal feature vectors are each combined with a plurality of signal classification models that are stored in a model memory (76). The signal classification models are each associated with a particular signal type that is recognized by the non-linear signal classifier. A scoring unit (72) generates a score for each of the signal classification models based on the result of the combination. The scores are analyzed by a selection unit (74) which determines which of the signal classification models (i.e., which of the signal types) most likely represents the received signal. Training equipment (60) is also provided for training the non-linear signal classifier (26) to recognize new signal types.

Abstract: An improved method and system for training and classifying using a low complexity and high accuracy multiresolutional polynomial classifier (412) is presented. A method of training an multiresolutional polynomial classifier which reduces the complexity of existing classifiers allows models representing subgroups of classes to easily be created. The models which represent subgroups of classes are applied to an unidentified input to produce a coarse classification of the unidentified input using a low order classifier. Once a coarse classification of the unidentified input is performed, a more detailed classification is performed using another low complexity classifier.

Abstract: An adaptive arrangement and method for the coded digital transmission of images includes an adaptive transmitter (101) having an image coder (109) which is operable at multiple image coding rates, a channel coder (111) which is operable at a plurality of channel coding rates, and is further operable to provide a plurality of power outputs, baud rates, and a plurality of image delivery rates. The transmitter (101) further includes a channel status monitor (115) which monitors the communication channel (105) connecting the transmitter (101) to a receiver (103). The channel status monitor (115) responds to changes in the quality of the communications channel by varying one or more of the image coding rate, the channel coding rate, the power, the baud rate, and the image delivery rate.

Abstract: The invention relates to a communication system (300) having a receiver (304) that is capable of performing targeted interference suppression. An interference classifier (314) within the receiver (304) analyzes a signal received from a channel (306) and identifies and classifies interference components within the signal. An interference suppressor (316) then suppresses the interference components in the signal based on interference type. In one embodiment, the interference suppressor (316) includes a plurality of interference suppression modules that are each optimal for suppressing certain interference types. The interference suppressor (316) selects one of the interference suppression modules based on the type of interference present in the received signal. In another embodiment, a hybrid interference mitigation system (10) is provided by combining targeted interference suppression, frequency hopping adaptation, and processing gain adaptation.

Abstract: The invention relates to a communication system (300) having a receiver (304) that is capable of performing targeted interference suppression. An interference classifier (314) within the receiver (304) analyzes a signal received from a channel (306) and identifies and classifies interference components within the signal. An interference suppressor (316) then suppresses the interference components in the signal based on interference type. The interference suppressor (316) includes a plurality of interference suppression modules that are each optimal for suppressing certain interference types. In one embodiment, one or more of the interference suppression modules perform interference suppression using modal moment estimates to identify different modes within the signal.

Abstract: The invention relates to a communication system (300) having a receiver (304) that is capable of performing targeted interference suppression. An interference classifier (314) within the receiver (304) analyzes a signal received from a channel (306) and identifies and classifies interference components within the signal. An interference suppressor (316) then suppresses the interference components in the signal based on interference type. In one embodiment, the interference suppressor (316) includes a plurality of interference suppression modules that are each optimal for suppressing certain interference types. The interference suppressor (316) selects one of the interference suppression modules based on the type of interference present in the received signal. In another embodiment, a hybrid interference mitigation system (10) is provided by combining targeted interference suppression, frequency hopping adaptation, and processing gain adaptation.

Abstract: The present invention relates to a communications system (100) that is capable of adapting to an unknown or varying spectral environment in a channel (14) between two communications units (12, 16). The system (100) maintains a spectral profile of the channel (14) and uses the spectral profile to determine appropriate transmit parameters for the system (100). The system (100) can be programmed to provide an optimal transmit signal for achieving a predetermined performance goal (such as, for example, maximum data rate at a given bit error rate (BER)) in light of the spectral environment in the channel (14). In one embodiment, the spectral profile is maintained in a spectrum table memory (27) that is periodically (or continuously) updated.

Abstract: An improved method of training a SISRS uses less processing and memory resources by operating on vectors instead of matrices which represent spoken commands. Memory requirements are linearly proportional to the number of spoken commands for storing each command model. A spoken command is identified from the set of spoken commands by a command recognition procedure (200). The command recognition procedure (200) includes sampling the speaker's speech, deriving cepstral coefficients and delta-cepstral coefficients, and performing a polynomial expansion on cepstral coefficients. The identified spoken command is selected using the dot product of the command model data and the average command structure representing the unidentified spoken command.

Abstract: A system is disclosed for an adaptive rate voice system to provide improvements in coded operation over changing communication channel (50) conditions. This adaptive rate system efficiently determines optimal voice/channel coding rates , coding strategies and modulation/demodulation for optimum voice quality and intelligibility. A system state estimator (100), channel status estimator (110) and channel status monitor (120) provides feedback in the system to optimize the communication channel. The system maintains a continuous link despite changing channel conditions and minimizes delays through the system. Even though simple in design, it provides relatively low complexity and powerful channel coding operation. Operating conditions are thus extended for CDMA and portable communication systems. Voice intelligibility is preserved in extremely noisy or even hostile channel conditions.

Abstract: A speaker identification system (10) employs a supervised training process (100) that uses row action projection (RAP) to generate speaker model data for a set of speakers. The training process employing RAP uses less memory and processing resources by operating on a single row of a matrix at a time. Memory requirements are linearly proportional to number of speakers for storing each speakers information. A speaker is identified from the set of speakers by sampling the speaker's speech (202), deriving cepstral coefficients (208), and performing a polynomial expansion (212) on cepstral coefficients. The identified speaker (228) is selected using the product of the speaker model data (213) and the polynomial expanded coefficients from the speech sample.