Sorry, you need to enable JavaScript to visit this website.

This paper presents a new adaptation of a Gaussian echo model (GEM) to estimate the distances to multiple targets using acoustic signals. The proposed algorithm utilizes m-sequences and opens the door for applying other modulations and signal designs for acoustic estimation in a similar way. The proposed algorithm estimates the system impulse response and uses the GEM to limit the effect of noise before applying deconvolution to estimate the time of arrival (TOA) to multiple targets with high accuracy.

Despite the success of sequence-to-sequence approaches in automatic speech recognition (ASR) systems, the models still suffer from several problems, mainly due to the mismatch between the training and inference conditions. In the sequence-to-sequence architecture, the model is trained to predict the grapheme of the current time-step given the input of speech signal and the ground-truth grapheme history of the previous time-steps. However, it remains unclear how well the model approximates real-world speech during inference.

We propose a novel automatic shrinkage tuning technique for the adaptive proximal forward-backward splitting (APFBS) algorithm. The shrinkage tuning aims to choose an appropriate value of the shrinkage parameter and achieve minimal system mismatch as possible. The system mismatch is approximated based on time-averaged second-order statistics. Numerical examples show that the proposed method achieves performance fairly close to that with a manually chosen shrinkage parameter for colored input signals at some signal to noise ratio (SNR).

Knowledge of second-order statistics of channels (e.g. in the form of covariance matrices) is crucial for the acquisition of downlink channel state information (CSI) in massive MIMO systems operating in the frequency division duplexing (FDD) mode. Current MIMO systems usually obtain downlink covariance information via feedback of the estimated covariance matrix from the user equipment (UE), but in the massive MIMO regime this approach is infeasible because of the unacceptably high training overhead.