Designing Immersive Science Learning for Low-Resource Secondary Classrooms: The VLEARN Platform

Immersive virtual reality (VR) has repeatedly been shown to support conceptual understanding in scienceeducation, but most of the systems described in the literature were built for well-resourced settings, and veryfew are aligned to a particular national curriculum. In Rwanda, the immersive-technology initiatives that doexist tend to focus on tertiary or vocational training, serve teachers rather than learners, or deliver immersivecontent with no instructional support inside the environment itself. How to design curriculum-aligned,learner-facing immersive science learning under real resource constraints is therefore still an open question.This paper describes the design and implementation of VLEARN, a standalone VR learning platform builtaround the Rwandan Competence-Based Curriculum at secondary level, covering Physics, Chemistry, Biology,and Geography. The contribution is a design case: an account of the instructional and technical decisionsmade while building the system, the constraints behind them, and the reasoning that connects them toestablished principles of multimedia and experiential learning.Three decisions receive particular attention. First, where the behaviour of a system is itself the thing beingtaught, the simulation is actually computed rather than scripted. The electric circuits module embedsSpiceSharp, a C# implementation derived from Berkeley SPICE 3f5, so circuits built by learners are solved bynodal analysis and can genuinely surprise them. Second, the atom-construction module enforces the KLMNshell-capacity model instead of merely depicting it: configurations that violate the model are rejected, andelement identity is resolved dynamically from particle counts. Third, the platform includes a conversationaltutor character whose main limitation, namely that she has no perceptual access to what the learner is doingin the scene, is treated here as a substantive open problem for embodied tutoring rather than a footnote.We report implementation-quality measures and describe a formative session with 17 participants (13 learnersand 4 instructors), which is reported qualitatively. Quantitative learning-outcome data from that session isnot available, and no claims of learning gain are made. The controlled evaluation the platform now needs isspecified in detail.Keywords: virtual reality; science education; instructional design; educational technology; low-resourcecontexts; Rwanda; secondary education; design case