Exploring Deep Learning for Cyclist Emotion Recognition

This study explores the application of Recurrent Neural Networks (RNN) to recognise emotions using multimodal physiological signals from cyclists, specifically heart-rate variability (HRV) and electrodermal activity (EDA). These signals are used to develop prediction models(such as LSTM and GRU) that can be deployed on bike-mounted edge devices. This approach aims to enhance urban cycling safety by enabling timely adaptations. The study evaluates the feasibility of these models on edge devices and provides recommendations for their effective deployment to reduce accidents and improve cycling safety.