Subject Area
Computer Science
Abstract
In this work, we propose a multimodal biometric authentication system leveraging sig nals from inertial sensors and audio-based haptic responses collected through standard gaming controllers. To enhance the biometric representation, we introduce a haptic-enhanced chirp signal and design a feature fusion strategy that integrates inertial and MFCC-based audio features. Our approach employs a novel LSTM-based architecture with multi-head self-attention, dual cross-modal attention, dynamic feature fusion, and residual correction, significantly improving model performance over traditional method without haptic input. Experimental evaluations demonstrate that the proposed method achieves a substantial improvement in authentication accuracy, achieving an accuracy of 98.5% and reducing Equal Error Rate (EER) to 0.4%. Robustness and generalization tests, conducted under varying environmental conditions and background noise scenarios, further validate the model’s resilience to real-world disturbances. Moreover, the proposed method provides practical advantages for deployment, offering a seamless, low-cost, and non-intrusive authentication experience suitable for immersive gaming environments.
Degree Date
Spring 2025
Document Type
Thesis
Degree Name
M.S.
Department
Computer Science and Engineering
Advisor
Chen Wang
Number of Pages
45
Format
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License
Recommended Citation
Ji, Rongwei, "PlaySense-ID: Haptic-Enhanced Multimodal Biometric Authentication for Gaming Controller Users" (2025). Computer Science and Engineering Theses and Dissertations. 49.
https://scholar.smu.edu/engineering_compsci_etds/49
