Project
Enhancing the effectiveness of automated vehicle sensory-based alert systems
With advanced driving assistance systems (ADASs), drivers are alerted using visual, auditory, and haptic methods. Unfortunately, the conspicuity of these alert systems does not always lead to desired driver behavior. For example, drivers may ignore ADAS alerts if they are too frequent, or worse, drivers disarm the corresponding ADAS technology if they find the alerts to be distracting or irritating. The purpose of this project is to: (1) review and synthesize literature concerning the implementation of human-machine interfaces (HMIs) in automated vehicles; (2) conduct a simulator study that examines the efficacy, trust and acceptance of drivers using promising HMIs identified in step 1; and (3) generate a set of recommendations for automated vehicle alerting systems. The literature review will be conducted by systematically searching using pre-determined terms through search databases and cross-referencing to find other relevant articles. We will generate a literature map to connect the relevant documents and compile a list of candidate HMI designs for the simulator study. For the simulator study, the main factor that will be varied is the HMI. Additional factors to be studied include demographic variables (e.g., age, gender, and socioeconomic status). Participants will be recruited from the University of Massachusetts Amherst campus and the surrounding community. There will be multiple drives, including a baseline drive where the vehicle is driven manually and multiple experimental drives where the automation’s HMI issues warnings to the driver. During all drives, drivers’ behavior will be recorded through the simulator (e.g., speed and acceleration), in-vehicle cameras, and eye tracking equipment. After participants complete the driving simulator portion of the experiment, they will complete a series of questionnaires to gauge the in-vehicle interface’s usability. The results will be analyzed using analysis of variance and multiple regression, where appropriate, to determine the effectiveness of the HMIs.
