Adapting DIN EN ISO 14505-2 for thermal comfort assessment in highly automated vehicle cabins: Effects of occupant posture and in-vehicle activity on equivalent-temperature comfort zones

Highly automated vehicle cabins allow occupants to work or rest in seated postures that differ from conventional driving and can therefore alter the relationship between local equivalent temperature T eq and perceived thermal state defined in DIN EN ISO 14505-2. This study adapts the Nilsson equivalent-temperature comfort-zone framework for two sedentary use cases in such cabins: an upright working condition and a more reclined relaxing condition. Two climatic-chamber studies covered a range from 12 to 34 ° C . The first study used a paired cool-range design from 12 to 24 ° C and showed clear scenario effects on thermal sensation at lower setpoints, with smaller differences near thermal neutrality. The second study extended the range from 22 to 34 ° C and confirmed that these scenario-dependent differences remained relevant toward the warm side. A cross-phase comparison in the overlap range supported an interpretable combined use of both datasets, while differences in phase-wise sensitivity indicated that the relationship between thermal sensation and equivalent temperature was not constant across the full range. Scenario-specific regressions enabled an ISO-compatible threshold inversion and the derivation of separate comfort-zone diagrams. Compared with the working condition, the relaxing condition yielded higher neutral equivalent temperatures and the largest deviations occurred at the extremities. Facial infrared temperatures and an in-ear temperature proxy were evaluated only as complementary auxiliary observables and showed positive associations with mean thermal sensation on aggregated setpoint level. Overall, the main contribution is a scenario-specific re-parameterization of the DIN EN ISO 14505-2 interpretation framework for highly automated vehicle cabins, while physiological observables are treated as supportive rather than primary inputs.