Blind spot warning with hedging (BSW-H): driving simulator-based evaluation of a connected-vehicle warning for large truck–passenger vehicle interactions

Objectives Conventional blind spot warning systems for large trucks generally alert only the equipped vehicle and do not directly communicate blind-spot risk to surrounding passenger-vehicle drivers. This study introduces Blind Spot Warning with Hedging (BSW-H), a warning concept intended for future connected-vehicle deployment and evaluates its passenger-vehicle-facing warning interface in a driving simulator. The study assesses driver behavioral response to the interface rather than an operational Vehicle-to-Everything (V2X) communication architecture.Methods A within-subjects experiment involved 43 licensed drivers. Each participant completed three scenarios: no warning (S0), combined visual-and-auditory warning (S1), and visual-only warning (S2). Blind-spot entry and exit were logged in all scenarios. Outcomes were ordinal blind-spot duration and 5-s speed response, defined as mean speed during the 5 s after entry minus mean speed during the 5 s before entry. Repeated events were analyzed using participant-random-intercept cumulative link and linear mixed-effects models adjusted for blind-spot area; speed-model inference used participant-clustered bias-reduced linearization (CR2) standard errors. Sensitivity analyses used Gamma and log-normal mixed models for continuous duration and alternative 2-, 3-, and 10-s speed windows.Results In the primary duration model, neither S1 (Odds Ratio (OR) = 1.094, p = .410) nor S2 (OR = 0.962, p = .729) was significantly associated with ordinal blind-spot duration relative to S0. In the primary speed model, S2 was associated with a more positive 5-s speed response (β = 0.569 km/h, 95% confidence interval (CI): 0.081 to 1.058, p = .023), whereas S1 was not statistically significant (β = 0.428 km/h, 95% CI: −0.214 to 1.070, p = .185). A more positive response indicates greater acceleration or less deceleration. Continuous-duration sensitivity analyses supported the null Scenario finding. The S2 association was also significant at 2 and 3 s but not at 10 s; S1 remained nonsignificant across all windows.Conclusions The visual-only BSW-H interface was associated with a modest immediate speed response but was not significantly associated with ordinal blind-spot duration. These findings demonstrate outcome-specific behavioral adaptation in a simulator, not reduced crash risk or operational connected-vehicle effectiveness. Field, test-track, and naturalistic research is needed to determine whether such responses reduce blind-spot exposure or conflict risk.