Abstract
As mobile learning (m-learning) becomes ubiquitous, verifying the identity of remote learners remains a significant challenge. Traditional username and password systems are vulnerable to credential sharing and fraud. This paper introduces the Android Virtual Learner Identify (AVLI) system, a multi-factor verification framework that leverages smartphone hardware and sensors. By collecting real-time facial images, GPS coordinates, and unique device identifiers (IMEI, MAC addresses), AVLI creates a robust "learner profile" stored within a Virtual Learning Environment (VLE) database, enabling educators to reliably identify students in a virtual classroom.
Introduction
In traditional educational settings, physical presence provides an inherent layer of accountability. However, the shift toward virtual learning environments has created an "identity gap." Without face-to-face interaction, teachers cannot easily verify if the person logged into an account is the actual student. Standard authentication methods are insufficient for high-stakes scenarios like remote examinations. The AVLI system addresses this by transforming the smartphone into a sophisticated identity verification tool, ensuring that student identities are tied to verifiable physical and environmental data.
The AVLI System Architecture
The AVLI prototype consists of an Android client application integrated with a PHP-based server and a Moodle VLE. The system's core functionality revolves around a multi-factor data collection process designed to verify the learner's identity and location.
As shown in the figure above, the client sends a comprehensive data packet to the Server API. This packet includes:
- Authentication: User and password credentials.
- Network Data: IP and MAC addresses.
- Geographic Data: Precise GPS latitude and longitude.
- Hardware Identity: The device's unique IMEI number.
- Visual Identity: A live facial image captured via the device camera.
Security and Data Integrity
To ensure the reliability of the collected data, the authors implemented several technical safeguards. A critical innovation is the "memory variable" approach for image handling. Instead of saving captured photos to the phone's file system—where they could be intercepted or replaced with fraudulent images—the application processes the camera data directly in RAM.
The image data follows a strict pipeline: from the camera to a bitmap object, then into a byte array, and finally encoded as a Base64 string for secure transmission. This ensures the image is a "live" capture. Furthermore, passwords are secured using MD5 hashing combined with a unique "salt" to prevent rainbow table attacks, while image compression is optimized at a 90% ratio to balance visual clarity with mobile bandwidth constraints.
Experimental Evaluation
The system was evaluated using a Lenovo P70 smartphone (Android 2.3) and a custom Moodle 1.9 environment.
- Performance: Testing indicated that a 90% compression rate offered an optimal upload time of approximately 1103ms, compared to 1143ms for uncompressed images.
- Efficiency: The Android client was designed to be extremely lightweight, with an installation package size of only 69 KB.
- Reliability: The use of GPS was found to be significantly more dependable than IP-based geolocation, which can be easily masked or altered by network routing.
Discussion and Future Work
While AVLI provides a robust data collection framework, it currently relies on manual verification by educators who must review the stored profiles. Future iterations could incorporate automated facial recognition to streamline this process. Additionally, while the system secures data transmission, the collection of sensitive identifiers like IMEI and GPS raises significant privacy considerations that warrant further ethical review. As mobile technology evolves, updating the prototype for modern Android permission models and deeper Google Maps API integration remains a priority.
Summary updated