Advanced prediction and effective regulation on driver’s thermal comfort under actual driving conditions

The thermal comfort control in automotive intelligent cabins necessitates the integration of the weighted Predicted Mean Vote (PMV) evaluation index, which is based on air parameters near the human body surface. However, these air parameters near the human body surface cannot be directly obtained during real driving, rendering it impossible to accurately acquire the PMV value. To address this issue, this paper innovatively proposes a two-stage acquisition method of environment-PMV (e-PMV): environmental parameters from 13 seat monitoring points are extracted, and a mapping relationship between these parameters and the actual PMV is established by using BP neural network. The high accuracy of e-PMV is demonstrated through a combination of experimental and simulation approaches. Based on e-PMV, the eP-DP control strategy that balances energy consumption and thermal comfort is proposed and coupled with the air conditioning thermal system. The results indicate that, compared to on–off and PID control strategies, eP-DP strategy reduces compressor energy consumption by 44.4% and 18.5%, respectively. The e-PMV evaluation index enables the acquisition of the driver’s current accurate thermal sensation state without relying on human body conditions, making a groundbreaking contribution to the realization of human–machine interaction in future intelligent cabin air supply strategies.