Using Emotion Recognition and Temporary Mobile Social Network in On-Board Services for Car Passengers

In next-generation cars, passengers will have more time for fun and relaxation, as well as the number of unknown passengers traveling together will increase. Thanks to the progress in Artificial Intelligence and Machine Learning techniques, new interaction models could be exploited to develop specialized applications that will be informed of the passengers’ experience. The mood and the emotional state of driver and passengers can be detected, and utilized to improve safety and comfort by taking actions that improve driver and passengers’ emotional state. Temporary Mobile Social Networking (TMSN) is a key functionality that can enhance passengers’ user experience by allowing passengers to form a mobile social group with shared interests and activities for a time-limited period by utilizing their already existing social networking accounts. By minimizing isolation and promoting sociability, TMSN aims to redesign user profiles and interfaces automatically into a group-wise passengers’ profile and a common interface. This work proposes and develops the generation of TMSN-inspired music selection through the Spotify music streaming service. The results obtained are promising and encourage further development toward the concept of in-car entertainment. Finally, we evaluate the performance of lightweight and heavy intelligent models that recognize the emotion of a person from its face, using Raspberry Pi 4 B devices. The results show that it is possible to realize a system with face detector and facial emotion recognition models on edge devices with sufficient performance (Frame per Second) to detect at least emotions expressed through macro-expressions.