Influence of color and icon encoding themed HMl on trust calibration in automated vehicles

Conditional driving automation, also known as SAE Level 3 automated driving, allows drivers to perform non-driving related tasks NDRT when certain conditions are met without the need for constant monitoring. However, these automated systems require human drivers to be prepared to take over control when faced with operational constraints, and in an emergency, the automated system will send a take-over request (TOR) to the driver via the human-machine interface (HMI). As a result, the in-vehicle HMI is becoming an increasingly complex and important information interaction system. This study systematically investigates the combined effects of color and icon coding in human-machine interfaces HMIs on trust calibration during SAE Level 3 automated driving scenarios, with a focus on emergencies. Twelve females and thirteen males made up the 25 valid data samples. The sample driving experience range was 0 to 5 years (Mean ​= ​1.56, SD ​= ​0.77), with a maximum age of 30 and a minimum age of 20 (Mean ​= ​22.68, SD ​= ​2.19). A one-way experimental design using a combination of subjective and objective data was used to study subjects' driving trust and NDRT performance under three different static driving interfaces. Distinct from previous works focusing on unimodal encoding effects, our research pioneers in examining the synergistic relationship between color semantics and icon semantics in emergency scenarios. Additionally, we propose a novel dynamic trust assessment framework integrating both subjective scales and ocular metrics (fixation count/dwell time) validated through psychophysiological literature. The study used three sections of the experimental road, and participants had to complete the non-driving related task in each section. It was found that (1) colour coding of information in the driving interface affects driving trust, and (2) the combined effect of color coding and icon coding led to higher subjective trust than either coding method alone. • To investigate the impact of color coding in the human-machine interface of level 3 autonomous driving on drivers' trust in NDRT tasks. • 25 participants were subjected to simulations of three emergency takeover scenarios. • Key findings: The color coding of the interface significantly affects trust levels; the combination of color and icon coding enhances subjective trust, outperforming single coding methods.