05 / FROM INDUSTRIAL SYSTEM TO RESEARCH INSTRUMENT
Build independence, not prompt dependence.
Adaptive guidance in industrial VR training — a research extension grounded in Hari-developed VRJ.
Existing VRJ + new browser policyVR integration of fading: planned · Learner evaluation: not conducted
Check the workspace
Try the choices below. Two consecutive successes reduce full prompts; four reduce support to on-request.
010203040506
Current instruction
Confirm the workspace is ready
Safety information is never faded. This simplified browser task is not a field procedure, qualification, or instruction for work on electrical equipment.
Changing the condition or assessment mode starts a fresh trace.
An inspectable support policy
A deterministic rule makes the experimental manipulation reproducible.
- 01 / Full prompt
Fewer than two consecutive successful actions: display the complete instruction.
- 02 / Brief cue
Two or three successes: retain a short cue instead of the complete instruction.
- 03 / On request
Four or more successes: withhold automatic instruction. Help remains available.
Errors and help reset the success streak. Independent mode disables automatic prompts. Safety information is always retained. The browser represents instruction levels as text; it does not modify VRJ audio.
Inspect this session’s event trace (0)
No actions recorded yet.
The existing industrial foundation
VRJ — developed by Hari Ragavan, not a newly invented portfolio concept.
Existing VRJ footage · 05:46 · Silent recording. The source contains voice playback even though this recording has no audio track.
View the full VRJ project description ↗What is built, and what is next
- Existing VRJ
- Unity/C# procedural interactions, visual and voice instructions, multiple training and assessment modules, plus a companion admin dashboard with progress, results, feedback and reports.
- Built here
- Fixed/adaptive policy comparison, error and help recovery, independent-mode representation, and an exportable interaction trace.
- Next engineering step
- Connect the proposed policy to VRJ’s instruction/audio controllers and validated step events. Add explicit help and error telemetry without changing the established procedure or assessment rubric.
- Not claimed
- No participant study, retention advantage, real-world transfer, or adaptive-policy deployment in VRJ.
A study that could answer the question
Proposed protocol — not results from a study already conducted.
Compare fairly
Randomly assign fixed or adaptive guidance for the same procedure, practice exposure and rubric. Record prior experience. If using a within-person design, counterbalance equivalent procedures to limit carry-over.
Measure independence
Use procedural errors in a delayed unguided assessment as the primary outcome. Also record help, time, omissions and safety-critical mistakes. Fewer help requests alone do not prove learning.
Respect the boundary
Obtain consent and appropriate ethics approval before collecting learner data. Physical transfer needs a supervised, de-energised mock-up; VR completion does not establish professional competence.
Research precedent: Aoyagi et al. (2019), fading feedback and motor retention. That task differs from cable jointing; it motivates testing, not assuming transfer.