Exjobbspresentation: Communication-Aware Safety Analysis for Longitudinal Autonomous Driving Using Realistic V2V Channel Models

Cooperative vehicle-to-vehicle warning messages can support earlier responses to emergency braking, but their safety benefit depends on wireless propagation, message-delivery reliability, retransmission delay, and vehicle dynamics. This thesis investigates how these factors jointly affect longitudinal safety in two- and three-vehicle driving scenarios.

A communication-aware safety-analysis framework is developed to establish a continuous modelling chain from initial scenario parameters to the required safe distance. Urban and highway environments are considered under both line-of-sight and vehicle-blocked propagation conditions. Waveform-level simulations characterize decoding performance for different modulation and coding configurations, while distance-dependent path loss, shadow fading, vehicle blockage, multipath propagation, and mobility are incorporated into the communication model. The resulting decoding-failure probabilities are evaluated over blind retransmissions and repeated warning messages, with retransmission dependence approximated using a coherence-time-based criterion, to determine the event-level communication delay. This delay is then combined with longitudinal vehicle kinematics to compute the minimum safe inter-vehicle distance, which is compared with the initial inter-vehicle distance to obtain the corresponding safety margin.

The results show that modulation and coding scheme (MCS) selection has its greatest safety impact under degraded propagation. Under line-of-sight conditions, higher-rate configurations generally maintain reliable delivery and short delays. Under vehicle blockage, particularly in highway scenarios, less robust MCS configurations lead to more failed transmissions, additional message repetitions, and longer communication delays, thereby increasing the required safe distance. The analysis identifies that when communication performance is adequate, the safety outcome is governed mainly by vehicle kinematics; when communication becomes unreliable, message-delivery delay can become a major contributor to the required separation. These findings demonstrate the importance of evaluating wireless communication and vehicle dynamics within a unified safety framework.

Handledare: Galina Sidorenko & Aleksei Fedorov

Examinator: Michael Lentmaier