Described a system for anomaly detection using anomaly cueing. In operation, an input image having two-dimensional (2D) image mixtures of primary components is reformatted into one-dimensional (1D) input signals. Blind source signal separation is used to separate the 1D input signals into separate output primary components, which are 1D output signals. The 1D output signals are reformatted into 2D spatially independent component output images. The system then calculates all possible pair product images of the 2D spatially independent component output images and corresponding signal-to-noise ratios. A pair product image is selected based on the peak signal-to-noise ratio and thresholded to identify anomalies in the pair product image. Several types of devices can then be controlled based on the identified anomalies in the pair product image.

Disclosed are implementations, including a method that includes obtaining measurement samples relating to electrical operation of an electric motor drive providing power to an electric motor, deriving, based on the samples, instantaneous estimates for parameters characterizing speed and/or position of the motor according to an optimization process based on a cost function defined for the samples, and applying a filtering operation to the instantaneous estimates to generate filtered values of the motor's speed and/or position. The filtering operation includes computing the filtered values using the derived instantaneous estimates in response to a determination that a computed convexity of the cost function is greater than or equal to a convexity threshold value, and/or applying a least-squares filtering operation to the derived instantaneous estimates and using at least one set of previous estimates derived according to the optimization process applied to previous measurement samples.

A tap-coefficient control circuit sets the tap coefficient converged by the second tap coefficient updater as an initial value of the tap coefficient in the first digital filter which is to be updated by the first tap coefficient updater, arranges the tap coefficients converged by the second tap coefficient updater in descending order of contribution degree to the convergence operation of tap coefficient update in the first tap coefficient updater, judges the tap coefficient not less than upper specified number to be valid and the tap coefficient less than the specified number to be invalid, and sets the tap coefficient of the first digital filter corresponding to the tap coefficient judged to be invalid to zero not to be used in a calculation of the first tap coefficient updater until a next judgment result is made.

A precision frequency monitor provides a precision frequency monitor value (PFM) indicative of the precision of the frequency or period of an input reference signal. A first averaging module is responsive to the input reference signal to find an average frequency or period during successive predetermined time periods defining operational cycles. A second averaging module is responsive to an output of the first averaging module to average the output of the first averaging module over N operational cycles, where N is an integer, and output an updated PFM value every N operational cycles. An infinite impulse response (IIR) filter is responsive to the output of the first averaging module to filter the output of the first averaging module to output interim updated PFM values within each sequence of N operational cycles.

There is proposed a method of estimating a navigation state with several variables of a mobile carrier according to the extended Kalman filter method, comprising the steps of:—acquisition of measurements of at least one of the variables,—extended Kalman filtering (400) producing a current estimated state and a covariance matrix delimiting in the space of the navigation state a region of errors, with the help of a previous estimated state, of an observation matrix, of a transition matrix and of the measurements acquired, the method being characterized in that it comprises a step (310, 330) of adjustment of the transition matrix and of the observation matrix before their use in the extended Kalman filtering in such a way that the adjusted matrices satisfy an observability condition which depends on at least one of the variables of the state of the carrier, the observability condition being adjusted so as to prevent the Kalman filter from reducing the dimension of the region along at least one non-observable axis of the state space, in which the observability condition to be satisfied by the adjusted transition and observation matrices is the nullity of the kernel of an observability matrix associated therewith and in which the adjustment comprises the steps of:—calculation (301) of at least one primary basis of non-observable vectors with the help of the previous estimated state—for each matrix to be adjusted, calculation (306, 308) of at least one matrix deviation associated with the matrix with the help of the primary basis of vectors, shifting (330) of each matrix to be adjusted according to the matrix deviation associated therewith so as to satisfy the observability condition.

Techniques are described for implementing automated control systems to control operations of specified physical target systems. In some situations, the described techniques include forecasting future values of parameters that affect operation of a target system, and using the forecasted future values as part of determining current automated control actions to take for the target system—in this manner, the current automated control actions may be improved relative to other possible actions that do not reflect such forecasted future values. Various automated operations may also be performed to improve the forecasting in at least some situations, such as by combining the use of multiple different types of forecasting models and multiple different groups of past data to use for training the models, and/or by improving the estimated internal non-observable state information reflected in at least some of the models.

From a mixed signal in which a first signal and a second signal are mixed, the second signal is removed at low processing cost and without delay. As a result, an estimated first signal which has low residue of the second signal and low distortion is obtained. An estimated first signal is generated by subtracting a pseudo second signal which is estimated to be mixed in a first mixed signal in which a first signal and a second signal are mixed from the first mixed signal. The pseudo second signal is obtained by a first adaptive filter using a second mixed signal in which the first signal and the second signal are mixed in a different proportion from the first mixed signal. A coefficient update amount of the first adaptive filter is made smaller as compared with a case when the estimated first signal is smaller than the first mixed signal, in case the estimated first signal is larger than the first mixed signal.

A digital filter circuit is described. The digital filter circuit includes a digital filter input, at least two finite impulse response (FIR) filter circuits, and a connection circuit. The digital filter input is configured to receive a digital input signal set having a data parallelism. The at least two FIR filter circuits are configured to process the digital input signal set at least partially. The at least two FIR filter circuits include a pre-adder sub-circuit, a convolution sub-circuit, and a post-adder sub-circuit, respectively. The connection circuit is configured to selectively connect the at least two FIR filter circuits based on the data parallelism of the digital input signal set.

An acoustic response monitor. The acoustic response monitor includes a speaker, a microphone, and a response analyzer in electrical communication with the speaker and the microphone. The speaker is configured to generate a sound in response to an excitation signal. The microphone is configured to generate a microphone signal in response to a sound. The response analyzer is configured to generate an adaptive filter to minimize a difference between the excitation signal as modified by the adaptive filter and the microphone signal. The response analyzer may be configured to determine a difference between the adaptive filter and a previously generated adaptive filter. The response analyzer may be configured to trigger an alarm if the difference exceeds a predetermined threshold.

Methods and systems for pruning a convolutional neural network (CNN) include calculating a sum of weights for each filter in a layer of the CNN. The filters in the layer are sorted by respective sums of weights. A set of m filters with the smallest sums of weights is filtered to decrease a computational cost of operating the CNN. The pruned CNN is retrained to repair accuracy loss that results from pruning the filters.