Abstract: An abnormal detection circuit is provided. The abnormal detection circuit includes a conversion circuit, a voltage detection circuit, and a warning circuit. The conversion circuit receives a three-phase alternating current (AC) power and converts the three-phase AC power into a driving power. The voltage detection circuit detects each phase of the three-phase AC power. When a voltage value of at least one phase AC power of the three-phase AC power is abnormal, the voltage detection circuit uses the driving power to output at least one control signal corresponding to the abnormality. The warning circuit is driven by receiving the driving power and outputs at least one warning signal corresponding to the abnormality in response to the at least one control signal.

Abstract: A MIMO symbol detection and search method, a decoding circuit and a receiving antenna system are provided. The signal detection and search method includes the following steps. A symbol search tree is obtained, and a plurality of candidate symbols at each layer of the symbol search tree are sorted. The candidate symbols are traversed in sequence at each layer of the symbol search tree. At each layer of the symbol search tree, if a cumulative partial Euclidean distance is greater than or equal to a threshold, un-scanned candidate symbols are excluded. If the cumulative partial Euclidean distance is less than the threshold, the threshold is updated by the cumulative partial Euclidean distance. When all of the candidate symbols have been calculated, an estimated symbol combination is outputted, and the scan of the symbol search tree is terminated.

Abstract: A data compression method, a data compression system and an operation method of a deep learning acceleration chip are provided. The data compression method includes the following steps. A filter coefficient tensor matrix of a deep learning model is obtained. A matrix decomposition procedure is performed according to the filter coefficient tensor matrix to obtain a sparse tensor matrix and a transformation matrix, which is an orthonormal matrix. The product of the transformation matrix and the filter coefficient tensor matrix is the sparse tensor matrix. The sparse tensor matrix is compressed. The sparse tensor matrix and the transformation matrix, or the sparse tensor matrix and a restoration matrix, are stored in a memory. A convolution operation result is obtained by the deep learning acceleration chip using the sparse tensor matrix. The convolution operation result is restored by the deep learning acceleration chip using the restoration matrix.

Abstract: An orientation predicting method, adapted to a virtual reality headset, comprises obtaining an orientation training data and an adjusted orientation data, wherein the adjusted orientation data is obtained by cutting a data segment off from the orientation training data, wherein the data segment corresponds to a time interval determined by an application latency; training an initial neural network model based on the orientation training data and the adjusted orientation data corresponding to the time interval; retrieving a real-time orientation data by an orientation sensor of the virtual reality headset; and inputting the real-time orientation data to the trained neural network model to output a predicted orientation data. The present disclosure further discloses a virtual reality headset and a non-transitory computer-readable medium.

Abstract: An orientation predicting method, adapted to a virtual reality headset, comprises obtaining an orientation training data and an adjusted orientation data, wherein the adjusted orientation data is obtained by cutting a data segment off from the orientation training data, wherein the data segment corresponds to a time interval determined by an application latency; training an initial neural network model based on the orientation training data and the adjusted orientation data corresponding to the time interval; retrieving a real-time orientation data by an orientation sensor of the virtual reality headset; and inputting the real-time orientation data to the trained neural network model to output a predicted orientation data. The present disclosure further discloses a virtual reality headset and a non-transitory computer-readable medium.

Abstract: In a training phase of a facial image reconstruction system, a thermal facial image is input into a convolutional encoder and a pre-trained thermal facial recognition model for feature fusion. The fusion result is input into a convolutional decoder for visible facial image reconstruction. A hybrid loss function is used in the training phase. In the training phase, the reconstructed visible facial image and its corresponding real visible facial image are input into the pre-trained visible facial recognition model and the pre-trained facial landmark detection model to calculate the hybrid loss function for model learning and parameter update.

Abstract: An integrated circuit and a method are provided. An integrated circuit comprises a first circuit, with a first character and at least one external control signal, and a character control unit. The character control unit controls the at least one external control signal and has a second circuit, with a second character essentially proportional to the first character, a character adjuster for adjusting the at least one external control signal, and a character monitor for monitoring the operation behavior of the second circuit to control the character adjuster to adjust the at least one external control signal accordingly.

Abstract: A channel estimation apparatus includes a de-map circuit to map multiple preambles corresponding to multiple channels into real part (I) and imaginary part (Q) in QPSK coordinate. The costas channel estimation circuit receives the real part (I) and the imaginary part (Q) to operate corresponding to one of the BPSK structure and the QPSK structure to proceed calculation of (Q?I)*Sign (I+Q) or Q*Sign (I)?I*Sign (Q) for outputting an output value, wherein Sign(x) represents a sign circuit taking a sign of the x value. The loop filter circuit filters the output value under time domain. The smooth circuit receives the output of the loop filter circuit to perform smooth processing under frequency domain and then feeds back the same to the de-map circuit to continue channel estimation of a next loop, wherein the phases of the preambles are adjusted according to a direction of the output value approaching to zero.

Abstract: A finite impulse response (FIR) filter and a corresponding filtering method are provided. The FIR filter receives an input sequence. The input sequence includes a plurality of input values. The FIR filter includes at least one first adder, at least one multiplier, and a second adder. Each first adder performs multiple addition operations simultaneously in parallel. Each addition operation outputs a sum of two of the input values. Each multiplier performs multiple multiplication operations simultaneously in parallel. Each multiplication operation outputs a product of one of the sums and one of a plurality of coefficients of the FIR filter. The second adder outputs a total sum of the products.

Abstract: An integrated circuit and a method are provided. An integrated circuit comprises a first circuit, with a first character and at least one external control signal, and a character control unit. The character control unit controls the at least one external control signal and has a second circuit, with a second character essentially proportional to the first character, a character adjuster for adjusting the at least one external control signal, and a character monitor for monitoring the operation behavior of the second circuit to control the character adjuster to adjust the at least one external control signal accordingly.

Abstract: A pre-decoded tail-biting convolutional code (TBCC) decoder and a decoding method thereof are provided. The decoder includes a pre-decoder, a storage module, and a control module. The pre-decoder receives a current state, a neighboring state, and a current path status corresponding to sequential data encoded in TBCC, generates predicted decoded bits, and determines whether states corresponding to minimum path metrics of neighboring stages are in continuity according to the current state, the neighboring state, and a current path status. The storage module is connected to the pre-decoder and stores the predicted decoded bits. The control module is connected to the storage module and the pre-decoder. In addition, the control module selects to output the decoded bits from the storage module when the continuity between the states corresponding to the minimum path metrics of the neighboring stages reaches a truncation length.

Abstract: A channel estimation method is provided. The method includes the following steps of: receiving an input symbol of an input signal and obtaining several pilot channel gains through calculation; executing an operation of interpolation on the pilot channel gains by a Wiener filter to obtain several data channel gains through calculation; calculating an adaptive alteration for the first and second multi-path statistical characteristic parameters according to the data channel gains and the pilot channel gains, and accordingly having the first and the second multi-path statistical characteristic parameters adjusted; generating an updated Wiener filter according to the adjusted first and second multi-path statistical characteristic parameters to execute an operation of channel estimation on a next input symbol of the input signal.

Abstract: A channel estimation method is provided. The method includes the following steps of: receiving an input symbol of an input signal and obtaining several pilot channel gains through calculation; executing an operation of interpolation on the pilot channel gains by a Wiener filter to obtain several data channel gains through calculation; calculating an adaptive alteration for the first and second multi-path statistical characteristic parameters according to the data channel gains and the pilot channel gains, and accordingly having the first and the second multi-path statistical characteristic parameters adjusted; generating an updated Wiener filter according to the adjusted first and second multi-path statistical characteristic parameters to execute an operation of channel estimation on a next input symbol of the input signal.

Abstract: A pre-decoded tail-biting convolutional code (TBCC) decoder and a decoding method thereof are provided. The decoder includes a pre-decoder, a storage module, and a control module. The pre-decoder receives a current state, a neighboring state, and a current path status corresponding to sequential data encoded in TBCC, generates predicted decoded bits, and determines whether states corresponding to minimum path metrics of neighboring stages are in continuity according to the current state, the neighboring state, and a current path status. The storage module is connected to the pre-decoder and stores the predicted decoded bits. The control module is connected to the storage module and the pre-decoder. In addition, the control module selects to output the decoded bits from the storage module when the continuity between the states corresponding to the minimum path metrics of the neighboring stages reaches a truncation length.