Abstract
Ubiquitous learning is an innovative approach that combines mobile learning and context-awareness, detecting a user’s location and surrounding context to provide tailored learning materials. This research presents a context-aware, multi-agent-based mobile educational game designed for On-the-Job training (OJT). By utilizing a multi-agent architecture (MAA), the system remains lightweight and scalable for resource-constrained mobile devices. A pilot study involving 37 freshmen demonstrated that the game effectively helps users familiarize themselves with new environments, showing high usability and engagement.
Introduction: The Challenges of On-the-Job Training
Traditional On-the-Job Training (OJT) often relies on static manuals or the availability of human mentors, which can be inefficient. While mobile learning (m-learning) offers a potential solution, it faces significant hurdles:
- Indoor Positioning: GPS is unreliable inside buildings where OJT typically occurs.
- Resource Constraints: Mobile devices often have limited processing power and battery life.
- Content Overload: Generic systems fail to provide personalized learning paths based on a user’s specific role or immediate surroundings.
The Context-Aware Mobile Educational Game (CAMEG) addresses these issues by creating a lightweight system that uses 2D barcodes (QR codes) for indoor positioning and dynamically generates personalized "learning chains."
Methodology: Multi-Agent Architecture and Mathematical Filtering
The CAMEG system is built on two technical pillars:
- Multi-Agent Architecture (MAA): The application is decomposed into eight specialized agents (e.g., Player Agent, Position Locator). This modularity allows the mobile device to run only necessary processes, conserving battery and memory.
- Intelligent Task Prioritization: To determine the sequence of learning activities, the system employs Rough Set theory to filter irrelevant objects and Information Theory to weight tasks. It prioritizes "simplified" objects (those with fewer characteristics) before introducing "diversified" (complex) ones, a "simplicity first" heuristic designed to build user confidence.
Indoor positioning is achieved via QR codes placed on physical objects like printers or coffee machines. Scanning these codes provides the system with the user's exact location and context without requiring expensive hardware like RFID.
Knowledge Structure and Pilot Study
The system's intelligence relies on a detailed knowledge structure representing the physical environment. For this study, the researchers mapped the 11th floor of a university building.
As shown in the figure above, the environment is categorized into domains such as Tasks, People, Places, and Events, which are further linked to specific objects like workstations (WS_1128) and appliances (Coffee maker_1125).
A pilot study was conducted with 37 graduate students (25 male, 12 female) at the National Kaohsiung First University of Science and Technology (NKFUST) to evaluate the system's usability using ISO 9241-11 standards.
Key Results and Usability Findings
The evaluation revealed high usability across thirteen factors, with mean scores consistently above 3.5. Notable findings include:
- High Learnability: The system was perceived as easy to learn, particularly by female participants who rated this factor at 4.58/5.0 (compared to 3.76/5.0 for males).
- Strong Intention to Use: Female participants also showed a higher intention to use the system in the future (4.50/5.0) than male participants (3.68/5.0).
- Efficiency in Orientation: The system successfully generated helpful sequential learning chains, such as moving from finding a workspace to learning how to use communal office equipment.
Conclusion and Future Directions
While the CAMEG system proved effective for orientation and OJT, the researchers identified areas for improvement. Current limitations include the "scavenger hunt" nature of the game, which can become tedious, and a lack of narrative depth. Future work will focus on integrating storytelling strategies and a more robust reward system to enhance long-term motivation and engagement.
Summary updated