Evaluation of the Training Efficiency of AR-Based Versus Paper-Based Instruction Methods with Different Levels of Industrial Maintenance Task Complexity

The objective of this paper is to evaluate the training efficiency of AR-based versus paper-based instruction methods with different levels of task complexity in terms of knowledge retention (i.e., short-term memory and long-term memory). Two maintenance operations were selected (i.e., low-demand (check all seals of the pump housing) and high-demand maintenance tasks (repair a gearbox of the piston pump and check all seals of the gearbox)). Twenty-eight healthy males were recruited and randomly divided into two groups. Each participant performed the maintenance task with assistance (i.e., AR-based or paper-based instruction) and repeated it without instructions twice (30 min apart and then two weeks later). Knowledge retention was measured using the total task completion time without guidance, error number, the number of self-corrected errors (SCEs), and the number of times help was sought from the researcher. The results show that using AR-based instruction enhanced the participants’ performance, especially in the highly demanding task, by reducing task completion time and errors in both short- and long-term retention tasks. Participants who were guided through a paper-based instruction method showed more errors, more instances of seeking help from the researcher, and a longer time to complete the long-term retention task, compared with those guided through the AR-based instruction method. Overall, using AR instruction to guide maintenance workers increased training transfer by 33.61%.