HMI Design and Customisation

Between trials, participants had the opportunity to customise their HMI using an alert settings screen (Figure 13.1) that appeared on the central console between trials. This provided the ability to hide or display aspects on the HUD, cluster, and central console, including wording regarding driving mode, a driving mode icon, a time to takeover readout, takeover question wording, and a coloured edge frame that indicated driving mode. A ‘road sense’ view could additionally be displayed on the cluster or central console; this graphically indicated the presence of vehicles around the drivers’ car.

Analog settings included the ability to adjust the brightness of ambient light, which indicated automation mode based on the colour. Audio volume could be adjusted that would affect both audio tones and vocalisations generated by the automation system. The vibrate setting allowed the intensity of the haptics to be adjusted. The takeover questions setting allowed the driver to specify the number of questions they wanted the system to ask as part of the takeover protocol.

When a participant completed making their adjustments, the settings were saved, and the customised HMI was employed in the HMI of the next trial.

Procedure

After arriving at the JLR facility, participants were welcomed and presented with a participant information sheet informing them of the details of the study. Their right to halt the study at any time was explained; they were then provided with an informed consent form which they had to read, initial, and sign in order for the study to continue. On completion of the consent form, participants were given a bibliographical form to complete to capture demographics data. They were then introduced to the simulator. The main controls were explained, and they were informed of the functionality of the customisation screen and the effects it had on the HMI. The customisation screen provided dynamic updates of the HMI as changes were made, allowing the participant to see the effects, and to test audio levels, light, and haptic intensities. Participants then took part in a test run, where they experienced three takeovers after 1-min OOTL intervals. After completion of the test run, the customisation settings were reset to defaults and the study started. Trials started with the participant being asked to accelerate onto the motorway, join the middle lane, keep up with traffic, and follow the HMI’s instructions. After a period of approximately 1 min, the HMI indicated to the participant that automation was available and informed them via text, icon, vocalisation, and two flashing green steering wheel buttons. Participants then activated automation by pressing the two steering wheel buttons; this was followed by the HMI indicating that automation was now in control of the vehicle. Participants then picked up the secondary task tablet and started to play the Tetris secondary task. After a period of either 1 or 10 min, dependent upon counterbalancing, the automation indicated via the HMI that the driver was required to get ready to take control. Participants were expected to put aside their secondary task, and follow the instructions presented by the HMI. The takeover protocol consisted of a set of questions designed to raise situation awareness. These questions could be presented in vocal, word, and icon form, dependent upon the HMI customisations. Participants responded vocally to each question, the answers to which were judged by an experimenter taking the part of the automation system using the Wizard of Oz approach. Incorrect or missed questions were repeated twice before moving to the next. Once the questions were answered by the participant, the HMI indicated for them to take control, which they did by pressing the two green buttons on the steering wheel. This constituted one takeover, the process was repeated twice more, and after completion, the participant was asked to pull safely to the hard shoulder and stop the vehicle. The customisation screen was then displayed on the central console, and the participant was asked to make adjustments to the HMI for the next trial. Once the participant was happy with their customisations, these were saved, and they were presented with three questionnaires in electronic form. Completion of the questionnaires marked the end of a trial. Four trials took place in total, two featuring 1-min OOTL durations and two featuring 10-min OOTL durations. Once all four trials were complete, the participant filled out a post-trial interview questionnaire.

Analysis

Standard questionnaires were used to assess drivers’ subjective attitudes and perceptions about takeover experience in HAD environment. The forms were administered immediately after they finished each trial in the simulator. Data was collected from 68 participants who took part in this study. Forms from three participants were excluded for analysis because they were not completed as the trials could not be finished. Missing responses were replaced by the median values.

Workload

Participants’ perceived workload was assessed by the NASA-TLX. As workload is a multidimensional construct, the scale was comprised of six questions regarding mental, physical, and temporal demand, as well as performance, effort, and frustration. Each item ranged from very low to very high, divided into 20 intervals. Averaged raw scores (raw TLX) were used to gain each participant’s overall workload score (Hart 2006). This scale has been used to evaluate drivers’ subjective workload in autonomous driving situations in a wide range of studies (de Winter et al. 2014).

Usability

The system usability scale was selected because it is a useful tool to examine subjective usability considering appropriateness of systems, or tools to a purpose viewed in relation to the context that they are used. The scale is composed of items demonstrating a wide range of aspects of usability such as complexity, the need for training, and support. It consists of ten questions including five positively and five negatively worded questions (Brooke 1996).

Acceptance

The acceptance of advanced transport telematics scale was adopted to measure the participants’ acceptance about takeover experience. It consists of two subscales assessing system usefulness, reflecting practical aspects, and satisfaction, representing pleasantness. Subscales are comprised of five and four items, respectively. Scores from the two subscales were calculated after verifying reliability was above the suggested value (Cronbach’s «>0.65) (Van Der Laan, Heino and De Waard 1997).

Trust

The scale of trust between people and automation was used to assess trust in the interfaces during the takeover. It consists of five questions regarding the level of distrust, and seven questions regarding the level of trust. Items were scored on a Likert scale ranging from 1 (not at all) to 7 (extremely) (Jian, Bisantz and Drury 1998). Mean values were used as representative scores because this method has been employed in relevant studies (Gold et al. 2015; Blomacher, Nocker and Huff 2018; Feldhiitter et al. 2016).

Data Analysis

Perceived workload, usability, acceptance, and trust of takeover experience through customisable interfaces were investigated based on comparisons between trials. The purpose was to identify the participants’ perception and attitude towards the takeover experience before and after customising HMI settings. Data sets measured after the first trial with the default settings, and last (fourth) trial, carried out with the last setting after three customisations between trials, were used because the settings for the last trial were more optimised than those for the second and third trials. Wilcoxon signed-rank test was used to test the differences in subjective experience: perceived workload, usability, acceptance, and trust of customisable interfaces in autonomous vehicles between the first and the last trials. It was chosen because it allows a comparison of two sets of data collected from the same participants (Field 2017) when the data is measured on an ordinal scale (McCrum-Gardner 2008). IBM SPSS Statistics 24 was used for the analysis.

 
Source
< Prev   CONTENTS   Source   Next >