Automated vehicle's human-machine interfaces design with haptic feedback: A meta-analysis

Automated vehicles (AVs) enable drivers to engage in non-driving-related tasks (NDRTs), creating challenges for safe transitions of control. Haptic feedback has been proposed as an effective takeover request (TOR) modality, yet evidence regarding feedback location and task context remains fragmented. This meta-analysis synthesizes results from 19 empirical studies to examine the effects of haptic feedback on drivers' response times during TORs. Comparisons were made between haptic, visual, auditory, and multimodal feedback, with moderation analyses considering feedback location and sensory-conflict-based NDRT categories. Results showed that haptic feedback generally outperforms visual feedback, particularly when drivers are engaged in visual-demanding NDRTs, and that multimodal feedback can further improve performance compared to haptic feedback alone. Feedback location emerged as a relevant design factor, with hand-based haptic feedback often associated with faster responses than seat-based feedback, although effects were context-dependent. These findings provide integrative evidence to inform AV design and highlight priorities for future research.