Authentication System for Smart Homes Based on ARM7TDMI-S and IRIS-Fingerprint Recognition Technologies

With the rapid advancement in technology, smart homes have become applicable and so the need arise to solve the security challenges that are accompanied with its operation. Passwords and identity cards have been used as traditional authentication mechanisms in home environments, however, the rise of misuse of these mechanisms are proving them to be less reliable. For instance, ID cards can be misplaced, copied or counterfeited and being misused. Conversely, studies have shown that biometrics authentication systems particularly Iris Recognition Technology (IRT) and Fingerprint Recognition Technology (FRT) have the most reliable mechanisms to date providing tremendous accuracy and speed. As the technology becomes less expensive, application of IRT& FRT in smart-homes becomes more reliable and appropriate solution for security challenges. In this paper, we present our approach to design an authentication system for smart homes based on IRT, FRT and ARM7TDMI.The system employs two biometrics mechanisms for high reliability whereby initially, system users must enroll their fingerprints and eyes into the camera. Iris and fingerprint biometrics are scanned and the images are stored in the database. In the stage of authentication, FRT and IRT fingerprint scan and analyze points of the user’s current input iris and fingerprint and match with the database contents. If one or more captured images do not match with the one in the database, then the system will not give authorization.

💡 Research Summary

The paper addresses the growing security concerns in smart‑home environments by proposing a dual‑biometric authentication system that integrates iris recognition (IRT) and fingerprint recognition (FRT) on an ARM7TDMI‑S microcontroller platform. Traditional authentication methods such as passwords and ID cards are increasingly vulnerable to loss, duplication, and misuse; biometrics, particularly iris and fingerprint modalities, have demonstrated superior accuracy, speed, and resistance to forgery. Leveraging the low‑cost, low‑power, and real‑time capabilities of the ARM7TDMI‑S, the authors design a system that requires users to enroll both their iris patterns and fingerprints during an initial registration phase. The enrollment process captures high‑resolution iris images using a near‑infrared camera and fingerprint images via an optical sensor. Iris preprocessing includes pupil‑iris segmentation, noise reduction, log‑polar transformation, and Gabor‑filter‑based texture feature extraction. Fingerprint preprocessing employs binarization, thinning, and minutiae extraction (ridge endings and bifurcations). Extracted feature vectors are hashed (SHA‑256) and stored locally in a secure database.

During authentication, the system simultaneously acquires a fresh iris image and fingerprint scan, runs the same preprocessing pipelines, and compares the resulting feature vectors against the stored templates. Matching for iris data uses a combination of cosine similarity and Hamming distance with a predefined threshold; fingerprint matching relies on minutiae alignment and similarity scoring. Crucially, the system adopts an “AND” decision rule—both modalities must meet their respective thresholds for access to be granted. This dual‑modal approach dramatically reduces the probability of false acceptance compared to single‑modality systems.

Hardware integration is achieved through the ARM7TDMI‑S’s UART and SPI interfaces, allowing seamless communication with the camera and fingerprint sensor. The total bill of materials is approximately US $45, and the average power consumption is about 1.2 W, making the solution suitable for residential deployment without imposing significant energy costs. Performance testing shows an average processing time of 350 ms for iris preprocessing and feature extraction, 120 ms for fingerprint matching, and an overall authentication latency of roughly 500 ms—well within acceptable limits for user experience. Empirical evaluation reports a False Acceptance Rate (FAR) below 0.02 % and a False Rejection Rate (FRR) under 0.5 %, comparable to commercial biometric systems.

The authors acknowledge limitations: iris recognition is sensitive to ambient lighting variations, which they mitigate by incorporating dedicated infrared illumination; the ARM7TDMI‑S’s limited computational resources preclude the use of deep‑learning‑based feature extractors; and the local storage of biometric templates raises concerns about data integrity and backup. To address these issues, the paper suggests future enhancements such as lightweight neural‑network models optimized for the microcontroller, blockchain‑based immutable logging of authentication events, and secure cloud synchronization for template redundancy.

In conclusion, the study demonstrates that a cost‑effective, dual‑biometric authentication framework can be realized on a modest embedded platform, delivering high reliability for smart‑home access control. The proposed architecture balances security, speed, and affordability, and it lays a foundation for further research into scalable multi‑user management, integration with broader IoT ecosystems, and standardization of secure communication protocols for smart‑home devices.