Citation

Sungheetha, Akey and R., Rajesh Sharma and S, Xavier Samuel (2026) Seismic Analysis of Mars Subsurface Liquid Water Detection: Computational Intelligence Approach for Planetary Characterization. Procedia Computer Science, 282. pp. 2805-2816. ISSN 18770509

Text 1-s2.0-S1877050926016807-main.pdf - Published Version Restricted to Repository staff only Download (461kB)

Abstract

This work presents the Enhanced Adaptive Stream Fusion (EASF) framework, a 28-phase theoretical seismic analysis pipeline for Mars subsurface water detection through computational signal processing and deep learning. The framework processes synthetic 300-second seismic waveforms sampled at 100 Hz, applying bandpass filtering in the 0.05–20 Hz range to extract P-wave and S-wave signatures. Feature extraction employs statistical descriptors encompassing mean values from -0.05 to +0.05 m/s and kurtosis from 2.0 to 12.0, combined with 2048-point Fast Fourier Transform analysis resolving a dominant 2.3 Hz subsurface resonance and Daubechies-4 wavelet decomposition across eight levels. A five-layer convolutional neural network with filter counts escalating from 64 to 512 classifies features with dropout regularisation at rates 0.50, 0.40, and 0.30. Subsurface characterisation employs iterative closest point procedures converging within 18 iterations to alignment error 4.3 × 10-4 s2 followed by Bayesian velocity inversion recovering a low-velocity anomaly of 4.2 km/s at 38–52 km depth. Multi-criteria decision fusion combining normalised amplitude (Anorm = 0.24, weight 0.40), velocity anomaly (Vanom = 0.11, weight 0.35), and attenuation ratio (Qratio = 0.58, weight 0.25) yields water indicator W = 0.71 exceeding threshold 0.65. Bootstrap aggregation over 1000 realisations confirms 95% confidence interval [0.63, 0.79]. The system achieves true positive rate 94.7%, false positive rate 5.3%, and area under the ROC curve 0.975, representing statistically significant improvements over conventional detection approaches with mean difference 16.4 percentage points (p < 0.001).

Item Type:	Article
Uncontrolled Keywords:	Seismicanalysis, Subsurface water detection, Computational intelligence, Neuromorphic computing, Planetary characterization, Mars exploration
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines
Divisions:	Faculty of Applied Communication (FAC)
Depositing User:	Ms Suzilawati Abu Samah
Date Deposited:	04 Sep 2026 02:13
Last Modified:	04 Sep 2026 02:13
URII:	http://shdl.mmu.edu.my/id/eprint/16678

Downloads

Edit (login required)