user study用户体验实验流程

1  [1] Makris S , Karagiannis P , Koukas S ,et al.Augmented reality system for operator support in human–robot collaborative assembly[J].CIRP Annals - Manufacturing Technology, 2016:S0007850616300385.DOI:10.1016/j.cirp.2016.04.038.

The experiment followed the two-stage procedure shown in Fig. 3. In
the first stage, a subject reported to the experiment by signing a standardized ethical agreement. Next, the experimenter explained the
experimental instructions, helped the subject put on the devices related
to AR user interfaces, and demonstrated the manual assembly tasks. This
stage lasted about 10 min in total. The subject then practiced the tasks
for three more minutes. After a 3-minute break, the subject proceeded to
perform the tasks for three times in the following sequence. The first trial
was conducted without any user interface as a baseline. The second and
third ones were assisted by the user interfaces based on visual and tactile
cues, respectively, which order was randomly determined. Each trial
lasted 5 min and followed by a 7-minute break in which the subject filled
out questionnaires and rested.
 

关于实验虽然在实验室环境中进行，但仍然有价值的说法

The experimental settings in this work, including the assembly task,
the robot, test environment, and participants, are far from real conditions in industry. However, we argue that the experimental findings still
have generalized applicability of AR user interface design in humanrobot collaboration. First, conducting evaluation experiments in laboratory settings or recruiting non-professionals as experiment participants is a common practice in human factor or usability research.
According to the systematic review of usability studies in AR [30], most
of the experimental studies were run with young non-expert participants. The most recent review on evaluating AR/MR applications [31]
reported the similar conclusion from analyzing 458 papers. It also
confirmed that the majority of evaluation experiments were conducted
in a controlled laboratory environment. The number of participants
generally ranged from 11 to 20. To recruit that many industrial operators usually would not be feasible in practice. Nevertheless, it is advantageous to have experts, even only a small group, participate in
usability assessment. Their feedback can complement or strengthen that
of non-expert participants
 

同一个人做情况可以第一天做一部分然后其余的第二天再做

5.1. Learning effect
A subject would complete the experiment in two consecutive days as
shown in Fig. 3. They performed the manual assembly three times on the
first day. The first time did not involve any user interface and the
remaining two were conducted with visual and haptic cues in a random
order. They experienced the remaining two interfaces on the next day,
also in a random order. This arrangement aimed to minimize potential
learning effects generated from earlier trials, which could be examined
from change of the work performance from the second to the fifth trials.
The ANOVA result does not show a significant learning effect during the
two-day period (F=1,54, p = 0.21). The participants’ performance did
not vary with the sequence of the user interface to be tested, either.
 

2 Peng Wang1 · Yue Wang1 · Mark Billinghurst2 · Huizhen Yang3 · Peng Xu4 · Yanhong Li，BeHere: a VR/SAR remote collaboration system based on virtual replicas sharing gesture and avatar in a procedural task[J], Virtual Reality. https://doi.org/10.1007/s10055-023-00748-5