A Behaviorally Informed Life-Cycle Assessment of Autonomous Electric Vehicles: Integrating Service Type and Size Preferences

Autonomous electric vehicles (AEVs) are widely expected to contribute to transport decarbonization; however, their environmental performance remains uncertain because post-adoption user behavior is rarely integrated into life-cycle assessment (LCA) frameworks. This study investigated how service-type-specific user preferences for vehicle size shape the system-level life-cycle greenhouse gas (GHG) emissions of AEV systems. Using stated-preference survey data from Busan, South Korea, a mixed ordered logit model was estimated to capture heterogeneity in vehicle-size choices across private, shared, and taxi-based AEV services. These empirically derived behavioral preferences were integrated into a behaviorally informed, cradle-to-grave LCA framework to simulate system-level emissions under a fully deployed AEV scenario. Results showed that under full AEV deployment, total GHG emissions decrease by 44.7%, with average emission intensity declining from 107.3 to 59.4 g CO 2 e/pkm. Although electrification and automation contribute to these reductions, behavioral factors, particularly occupancy, emerged as the dominant drivers of emission outcomes. Higher occupancy in shared-AEV services substantially reduced emissions, whereas persistent preferences for midsize and larger vehicles offset part of the potential downsizing benefits. These findings highlight that effective AEV-led decarbonization depends not only on technological efficiency, but also on service design and behavioral interventions, such as promoting ride pooling and managing vehicle size preferences.