Research puts forward perception-based cognitive workload evaluation methods to help VR developers and users measuring their workload when playing with a VR application. Approaches to measure workload based on biosensors have progressed significantly, while evaluation based on subjective methods still rely on standard questionnaires such as the NASA-TLX table, the Subjective Workload Assessment Technique and the Modified Cooper Harper scale. The pre-defined questions enable operators to carry out experiments and analyse the data more easily than with biofeedback. However, the subjective evaluation process can bias the results because of unperceived internal changes and unknown factors among users. It is therefore necessary to have a method to handle and analyse this uncertainty. We propose to use the Technique for Order Performance by Similarity to Ideal Solution (TOPSIS) model to analyse the NASA-TLX table for measuring the overall user workload instead of using the classical weighted sum method. To show the advantage of the TOPSIS approach, we performed a user experiment to validate the approach and its application to VR, considering factors including the VR platform and the scenario density. Three different weighting methods, including the fuzzy Analytic Hierarchy Process (AHP) from fuzzy logic, the classical weighting based on pairwise comparison and the uniform weighting method, were tested to see the applicability of the TOPSIS model. The results from TOPSIS were consistent with those from other evaluation methods; a significant reduction in the coefficient of variation (CV) was observed when using the TOPSIS model to analyse the NASA-TLX scores, indicating an enhanced precision of the workload evaluation by the TOPSIS method. Our work has a potential application for VR designers and experimenters to compare cognitive workload among conditions and to optimize the settings.

研究提出了基于感知的认知工作负载评估方法，以帮助VR开发人员和用户在玩VR应用程序时衡量其工作负载。基于生物传感器的工作量测量方法取得了显着进步，而基于主观方法的评估仍然依赖于标准问卷，例如NASA-TLX表，主观工作量评估技术和改良的Cooper Harper量表。与生物反馈相比，预定义的问题使操作员可以更轻松地进行实验和分析数据。但是，由于用户内部未觉察到的内部变化和未知因素，主观评估过程可能会使结果产生偏差。因此，有必要提供一种方法来处理和分析这种不确定性。我们建议使用“类似于理想解决方案的订单性能技术”（TOPSIS）模型来分析NASA-TLX表，以衡量整体用户工作量，而不是使用经典的加权和方法。为了展示TOPSIS方法的优势，我们进行了一个用户实验，以验证该方法及其在VR中的应用，同时考虑了VR平台和场景密度等因素。测试了三种不同的加权方法，包括来自模糊逻辑的模糊分析层次过程（AHP），基于成对比较的经典加权和统一加权方法，以查看TOPSIS模型的适用性。TOPSIS的结果与其他评估方法的结果一致；使用TOPSIS模型分析NASA-TLX分数时，观察到变异系数（CV）的显着降低，这表明使用TOPSIS方法提高了工作负荷评估的精度。我们的工作对于VR设计人员和实验人员来说具有潜在的应用潜力，可以比较条件之间的认知工作量并优化设置。