This invention discloses a method for predicting and evaluating urban
street light and heat human factors based on multimodal behavioral
coupling, belonging to the field of urban
microclimate environment and
human comfort assessment technology. The method includes the following specific steps: S1
Multimodal data acquisition; S2 Data preprocessing and multimodal input fusion; S3 Light-heat-human factor
coupling modeling; S4 Response prediction and
deep learning modeling; S5
Exposure integral index calculation; S6 Interpretable output of results; S7 Digital twin feedback and optimization. This invention is the first to deeply couple light environment factors with thermal environment factors and human behavioral responses, overcoming the limitations of traditional
thermal comfort assessments that only focus on a single
thermal index. By automatically extracting complex relationships through a
deep learning model, it improves the accuracy and robustness of comfort prediction. Compared with methods that simply use indices such as PMV / UTCI, this invention provides more accurate predictions of subjective
thermal sensation in outdoor dynamic environments, and its prediction of behavioral responses fills a gap in existing technology.