Abstract
This study examines three instructional modalities—laser projection, video, and pictorial guidance—in a VR-based manufacturing assembly task. Twenty-five participants (M = 24.5 years) completed a 20-block layout task under easy (no-task) and hard (1-back) conditions. Cognitive workload (ocular metrics, NASA-TLX) and task performance (completion time, accuracy) were measured. Laser projection showed the lowest blink rate (M = 14.7 blinks/min) and fastest times (M = 241 s) versus video (M = 18.6 blinks/min, M = 353 s) and pictorial (M = 22.2 blinks/min, M = 276 s), suggesting less cognitive strain. Pupil dilation was higher for laser (M = 2.82) and video (M = 2.96) than pictorial (M = 2.28), likely due to VR light reflexes. Accuracy was similar across modalities. Hard conditions increased workload, reducing performance. Laser projection enhances efficiency for complex assemblies, with implications for smart manufacturing.
Introduction
Cognitive load, the mental effort required for task management, is critical for workplace safety and efficiency (International Organization for Standardization, 2016). Cognitive overload impairs performance, slows reaction times, and reduces accuracy (Biondi et al., 2021). In manufacturing, elevated cognitive load prolongs assembly times and increases errors (Cohen et al., 2017). Technological advancements have increased task complexity, necessitating advanced guidance systems (ElMaraghy et al., 2021). Augmented reality (AR), such as laser projection, overlays instructions onto the workspace, potentially reducing cognitive strain by minimizing attention switching (Vanneste et al., 2020). This study compares laser projection, video, and pictorial guidance in a virtual reality (VR) based assembly task to assess their impact on cognitive workload and performance.
Methods
Participants
Twenty-five volunteers (6 females, M = 24.5 years, SD = 3.5) from the University of Windsor participated, receiving a $20 Amazon gift card. Participants had normal or corrected-to-normal vision, no motion sickness, and limited VR experience (M = 2.3 on a 1–5 scale). The study complied with APA ethics and was approved by the University of Windsor Research Ethics Board (#24-135).
Design
A 3 × 2 factorial design was used with instructional modality (laser projection, video, pictorial) and cognitive task difficulty (easy: no-task; hard: 1-back) as within-subject factors. Dependent variables included standardized pupil size, blink rate, task accuracy, completion time, and NASA-TLX scores.
Equipment
The HTC VIVE Pro Eye VR headset (1440x1600 pixels/eye, 90 Hz, Tobii eye tracker) was used. The task involved arranging 20 uniquely coded blocks on an 86 cm × 86 cm virtual table (see Figure 1 and Figure 2). Laser projection used a green laser beam (#41FF00FF, 27.5 cd) to guide placement. Video guidance was displayed on a 24.8-inch virtual tablet, and pictorial guidance showed the final layout on a 100 cm × 100 cm board. The 1-back task required recalling the second-to-last digit in a sequence presented every 2 seconds.
Laser projection (a), video guidance (b), pictorial guidance (c).
Experimental setup. (a) represents the VR workspace, and (b) represents participants using the VR.
Procedure
Participants signed consent forms, underwent eye-tracking calibration, and completed a practice session. Six conditions (3 modalities × 2 difficulties) were counterbalanced using a Latin square design. Each trial lasted 3–5 min, followed by NASA-TLX ratings. The experiment took up to 60 min.
Data processing
Pupil size was standardized using baseline means and standard deviations. Blinks were detected as interruptions in pupil tracking (>75 ms). Task accuracy was scored out of 40 points (2 per block: 1 for position, 1 for orientation). Completion times were recorded manually. NASA-TLX scores were collected via a digital interface in VR.
Data analysis
Repeated-measures ANOVA was conducted in R Studio, with Mauchly’s tests for sphericity and Greenhouse-Geisser corrections as needed. Bonferroni-corrected post-hoc tests and Cohen’s d effect sizes were calculated (alpha = .05).
Results
Cognitive Workload
Blink rate varied by modality, F(2, 48) = 32.438, p < .05. Laser projection (M = 14.7 blinks/min, SD = 6.73) yielded lower rates than pictorial (M = 22.2, SD = 8.61; t(24) = −8.15, p < .05, d = −1.63) and video (M = 18.6, SD = 8.53; t(24) = −3.75, p < .05, d = −0.75) (see Figure 3). Pupil diameter also differed, F(2, 48) = 18.086, p < .05, with laser (M = 2.82, SD = 1.42) and video (M = 2.96, SD = 1.16) higher than pictorial (M = 2.28, SD = 1.22; t(24) = 4.22, p < .05, d = 0.84) (see Figure 4). NASA-TLX frustration scores were lower for laser projection (M = 8.12) than pictorial (M = 9.42, p < .05, d = −0.59) and video (M = 10.1, p < .05, d = −0.86).
Average blink rate (blinks/minute) across condition by task difficulty. Error bars represent standard error.
Standardized pupil diameter across conditions by task difficulty. Error bars represent standard errors.
Task Performance
Completion times varied by modality, F(2, 48) = 31.472, p < .05. Laser projection (M = 241 s, SD = 101) was faster than video (M = 353 s, SD = 104; t(24) = −7.46, p < .05, d = −1.49) and pictorial (M = 276 s, SD = 102; t(24) = −5.87, p < .05, d = −1.17). Accuracy showed no modality effect, F(1.58, 38.04) = 2.982, p > .05 (laser: M = 0.965; pictorial: M = 0.948; video: M = 0.935) (see Table 1).
Task Completion Times and Accuracy Scores.
Effect of Cognitive Difficulty
Hard conditions increased blink rates (M = 20.7 vs. 16.3 blinks/min, F(1, 24) = 35.620, p < .05, d = -0.83), reduced accuracy (M = 0.921 vs. 0.977, F(1, 24) = 27.507, p < .05, d = 0.65), and prolonged completion times (M = 316 s vs. 265 s, F(1, 24) = 14.333, p < .05, d = −0.43). Pupil diameter showed no main effect (p >.05) but interacted with modality, F(2, 48) = 7.468, p < .05, with laser projection showing higher dilation under hard conditions (t(24) = −3.05, p < .05, d = −0.61).
Discussion
Laser projection reduced cognitive workload, with lower blink rates (M = 14.7 blinks/min) and faster completion times (M = 241 s), supporting its efficiency in minimizing attention switching (Vanneste et al., 2020). Higher pupil dilation in laser and video conditions may reflect VR light reflexes (Lee et al., 2024). Accuracy was comparable across modalities, suggesting all are viable for simple tasks. Hard conditions increased workload and impaired performance, consistent with prior studies (Biondi et al., 2021). Laser projection’s advantages align with AR benefits in manufacturing (Funk et al., 2016), suggesting its potential for complex assemblies.
Conclusion
Laser projection offers superior efficiency and reduced cognitive strain for VR-based assembly tasks, ideal for complex manufacturing processes. Pictorial guidance may suffice for simpler tasks due to cost-effectiveness, but laser projection’s benefits in speed and workload reduction support its use in smart factories. Future research should validate these findings in real-world settings.
Footnotes
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.