Reliable human-centered automation remains challenging to deploy in everyday settings because camera-based sensing raises privacy and bystander-consent concerns, while wearable-based sensing depends on long-term user compliance. CogniWave is a practical framework that integrates contactless FMCW millimeter-wave radar with an uncertainty-aware cognitive decision module to enable privacy-preserving and safe intelligent automation. The framework is organized into three tiers: (i) a Perception Layer that converts raw radar IQ into compact, machine-consumable features a micro-Doppler motion embedding, 3D trajectory kinematics, a respiration variability index (RVI), and a radar reliability score; (ii) a Cognitive Interpretation Layer that maps these features through CogiNet into activity probabilities, structured human-state tokens, and an uncertainty signal; and (iii) a Human-Centered Automation Layer that enforces an explicit act abstain clarify policy with user control (override/opt-out) and transparency logging. We focus on an eight-class activity set—empty, presence-still, sitting, standing, walking, sit-to-stand, fall, and inactivity anomaly—selected to balance daily automation triggers with safety-critical events. In experiments across subject and environment-disjoint splits, the dataset comprises 4000 trials collected from 50 participants, yielding 328 min of radar recordings and 37,128 segmented windows. CogniWave achieves 94.1
This letter proposes a framework for multi-parameter recognition of overlapping spectrum signals (OSS) using unmanned aerial vehicle assisted radio monitoring systems. It integrates signal category identification, signal count, position estimation, and power transmission estimation. We merge the signal-separation and location-power estimation tasks into a coupled global formulation and design the global joint non-negative matrix factorization recognition (GJNR) method for OSS multi-parameter estimation. Under the adopted line-of-sight -dominant large-scale fading model, simulations show that GJNR improves signal category recognition accuracy by at least 32.8
Current noise reduction algorithms for cochlear implants (CIs) adopt a binary speech–noise discrimination mechanism, which tends to erroneously suppress background musical signals. Meanwhile, conventional acoustic features suffer insufficient robustness in real complex soundscapes such as indoor reverberation and amplitude perturbation, severely degrading musical perception for implant recipients. To address this problem, this paper explores the inherent frame-wise distribution law of audio bandwidth per frame and proposes the acoustic feature of Discrete Frame Bandwidth Concentration (DFBC), enabling accurate identification and preservation of musical signals under complex acoustic conditions. First, spectrogram and temporal feature analyses are performed to clarify frequency-domain discrepancies among speech, music and ambient noise, verifying that musical signals possess exclusive stepwise discrete bandwidth variation characteristics. Spectral centroid trajectory typing for music is accomplished using the GTZAN dataset, which validates that the discrete concentration property of bandwidth per frame is immune to musical genres and spectral patterns, with superior stability compared with spectral centroid. A single-threshold DFBC binary classification model for music/non-music discrimination is constructed, with Mel-Frequency Cepstral Coefficients (MFCC) set as the baseline for comparative experiments under four practical CI operating conditions: amplitude fluctuation, sampling rate distortion, additive noise and indoor reverberation. The experimental results demonstrate that DFBC achieves a classification accuracy of 80.07
This paper investigates the problem of reconstructing band-limited signals from their samples. The concepts of ill-posed problems and regularization operators are introduced, and the ill-posedness of the reconstruction problem is analyzed in the presence of noisy sampling data. A regularized reconstruction algorithm is proposed by modifying the sinc function through the removal of discontinuities in its Fourier transform. The convergence properties and error estimates of the proposed algorithm are established theoretically for noisy sampling data. Numerical experiments are presented to compare the proposed method with several existing reconstruction algorithms, demonstrating its improved performance in the presence of noise.
High-precision indoor positioning remains sensitive to non-line-of-sight (NLOS) propagation and multipath interference. This study investigates cross-environment adaptation of Encoder-only Transformer (ET) models for raw channel impulse response (CIR) sequences. The adaptation configuration combines optional Adapter insertion after self-attention with selective updates of existing transformer modules and is selected separately for NLOS classification and ranging-error regression under limited target-domain supervision. TPE and hyperband are used for configuration search and early pruning. Experiments on a public dataset collected in four indoor environments show that, under the evaluated settings, ET obtains an average classification accuracy of 84.94
