Smart Login System with Behaviour Based Security | IJCT Volume 13 – Issue 3 | IJCT-V13I3P39
International Journal of Computer Techniques
ISSN 2394-2231
Author
Siddhi Chavan, Dhanush Prithiviraj, Krishna Patil, Yash Choudhary
Abstract
Mobile devices today are only as secure as the moment they are unlocked. Once past the lock screen, nothing challenges the active session—regardless of who is actually holding the phone. AEGIS Shield addresses this gap through a multi-level security mesh (v4.0) that combines a three-stage authentication pipeline with continuous post-login behavioural monitoring. The system runs users through Identity Vector Analysis, OTP Decryption via intercepted access keys, and a Biometric Core Verification knowledge challenge before granting Terminal Access. Inside the session, a live dashboard monitors power source, compute heap, bridge network status, and core uptime in real time. Four background processes—SECURE_GATEWAY, BIO_SENSOR_CORE, NEURAL_MESH_SYNC, and THREAT_REACTION—feed the Decision Engine (System Analyst V2), which flags behavioural anomalies and can terminate all linked sessions simultaneously. Testing demonstrated 14.8ms system latency with stable encrypted 4G connectivity, confirming that continuous authentication at this level is practically deployable on standard mobile hardware.
Keywords
Behavioural Authentication, Multi-Layer Security, Biometric Verification, OTP, Keystroke Dynamics, Anomaly Detection, Mobile Security, Neural Mesh
Conclusion
AEGIS Shield demonstrates that a phone can remain secure after it is unlocked—not just at the moment it opens. The three-stage login establishes identity on solid ground; the behavioural monitoring layer keeps reassessing it throughout the session. At 14.8ms system latency with no perceptible impact on device performance, the framework shows that continuous authentication of this depth is practically viable on standard mobile hardware, not just in a controlled lab environment.
Work remains. Behavioural baselines drift as operators naturally change how they interact with their devices over time, and the system needs smarter recalibration logic to handle that without creating new vulnerabilities. The NEURAL_MESH_SYNC memory footprint needs to be trimmed for lower-end devices. And the anomaly detection has not yet been tested against adversaries specifically attempting to mimic a target’s typing patterns—that is the most important stress test still ahead. The core architecture is sound; the remaining problems have clear paths forward.
References
[1][1] A. Amraoui and B. Zouari, “Machine learning-based behavioural security for smart environments,” J. Network Security, vol. 23, no. 4, pp. 891–907, 2024. [2][2] H. Watanabe, K. Tanaka, and R. Fujimoto, “Immunity-based continuous authentication on smartphone platforms,” IEEE Trans. Mobile Comput., vol. 23, no. 8, pp. 4102–4117, 2024.
[3][3] D. Progonov, O. Kovtun, and I. Opirskyy, “BehaviorID: Context-dependent user authentication via neural pattern recognition,” Proc. ACM CCS, pp. 1543–1558, 2024.
[4][4] M. Alawami, I. Khalil, and A. Abutaleb, “MotionID: Practical behavioural biometrics for heterogeneous mobile hardware,” Proc. NDSS Symp., pp. 1–16, 2025.
[5] A. Buriro, B. Crispo, and Y. Zhauniarovich, “Risk-driven one-shot-cum-continuous authentication for mobile platforms,” J. Netw. Comput. Appl., vol. 218, p. 103702, 2024.
How to Cite This Paper
Siddhi Chavan, Dhanush Prithiviraj, Krishna Patil, Yash Choudhary (2026). Smart Login System with Behaviour Based Security. International Journal of Computer Techniques, 13(2). ISSN: 2394-2231.
