A High-Performance Training-Free Pipeline for Robust Random Telegraph Signal Characterization via Adaptive Wavelet-Based Denoising and Bayesian Digitization Methods

Random telegraph signal (RTS) analysis is increasingly important for characterizing meaningful temporal fluctuations in physical, chemical, and biological systems. The simplest RTS arises from discrete stochastic switching events between two binary states, quantified by their transition amplitude and dwell times in each state. Quantitative analysis of RTSs provides valuable insights into microscopic processes such as charge trapping in semiconductors. However, analyzing RTS becomes considerably complex when signals exhibit multi-level structures or are corrupted by background white or pink noise. To address these challenges and support high-throughput RTS characterization, we propose a modular, training-free signal processing pipeline that integrates adaptive dual-tree complex wavelet transform (DTCWT) denoising with a lightweight Bayesian digitization strategy. The adaptive DTCWT denoiser incorporates autonomous parameter selection rules for its decomposition level and thresholds, optimizing white noise suppression without manual tuning. Our Bayesian digitizer formulates RTS level assignment as a probabilistic latent-state inference problem incorporating temporal regularization without iterative optimization, effectively resolving binary trap states even under residual notorious background pink noise. Quantitative benchmarking on large synthetic datasets with known ground truth demonstrates improved RTS reconstruction accuracy, trap-state resolution, and dwell-time estimation across diverse noise regimes and multi-trap scenarios, while achieving up to 83x speedups over classical and neural baselines. Qualitative validation on experimental RTS data when no ground truth is available illustrates practical usability and flexibility for real-time or large-scale analysis in real measurement settings.

💡 Research Summary

This paper addresses the challenging problem of accurately characterizing random telegraph signals (RTS) that are corrupted by both white and pink (1/f) noise and may contain multiple discrete levels (multi‑trap scenarios). The authors propose a fully modular, training‑free processing pipeline that couples an adaptive dual‑tree complex wavelet transform (DTCWT) denoiser with a lightweight Bayesian digitizer.

The DTCWT denoiser exploits the shift‑invariance and directional selectivity of the complex wavelet transform to preserve abrupt transitions while suppressing background fluctuations. Crucially, the method automatically selects the decomposition depth and threshold values based on signal length and estimated noise power, using a SURE‑based rule that eliminates the need for manual tuning. This adaptive scheme yields up to a 12 dB improvement in signal‑to‑noise ratio for white noise and effectively attenuates low‑frequency components of pink noise.

After denoising, kernel density estimation (KDE) extracts candidate amplitude levels. The Bayesian digitizer then treats level assignment as a latent‑state inference problem. For each time point, the posterior probability of belonging to a particular level combines a Gaussian observation model (centered at the KDE‑derived means) with a Markov‑type prior that encodes expected dwell‑time statistics. Because the inference is performed in a single forward‑backward pass, there is no iterative EM or MCMC overhead, resulting in very low computational cost. The approach robustly distinguishes true levels from spurious peaks introduced by pink noise, achieving trap‑count errors below 5 % even in three‑trap scenarios with high noise levels.

The authors benchmark the pipeline on a synthetic dataset of 1,800 RTS traces (each 100 k samples long) covering 1‑, 2‑, and 3‑trap configurations, white‑noise amplitudes η∈