Mobile expert system for arm muscle training using forward chaining and best-first search

Abstract

Purpose:  to develop a mobile-based expert system to provide personalized arm muscle exercise recommendations, addressing the limited availability of professional fitness instructors.

Method: this study used the waterfall development model, including analysis, design, implementation, and testing, and utilized J2ME and MySQL. Expert knowledge was obtained through interviews and literature review then formalized into if–then rules. The system applies forward chaining with best-first search optimization and was evaluated using black box testing and expert validation.

Findings: the mobile-based expert system successfully generates personalized arm muscle training recommendations based on user anthropometric data, providing accurate recommendations for exercise type, sets, repetitions, and load.

Implications: mobile-based expert systems can provide accessible, personalized, and reliable resistance training guidance, bridging the gap between beginner users and professional fitness expertise while demonstrating the potential of AI-based health technology in resource-limited environments.

Originality: combining forward chaining and best-first search in a mobile expert system to provide personalized arm training recommendations.