Formation control of autonomous vehicles with emotion assessment

Autonomous driving is a major state-of-the-art step that has the potential to transform the mobility of individuals and goods fundamentally. Most developed autonomous ground vehicles (AGVs) aim to sense the surroundings and control the vehicle autonomously with limited or no driver intervention. However, humans are a vital part of such vehicle operations. Therefore, an approach to understanding human emotions and creating trust between humans and machines is necessary. This thesis proposes a novel approach for multiple AGVs, consisting of a formation controller and human emotion assessment for autonomous driving and collaboration. As the interaction between multiple AGVs is essential, the performance of two multi-robot control algorithms is analysed, and a platoon formation controller is proposed. On the other hand, as the interaction between AGVs and humans is equally essential to create trust between humans and AGVs, the human emotion assessment method is proposed and used as feedback to make autonomous decisions for AGVs. A novel simulation platform is developed for navigating multiple AGVs and testing controllers to realise this concept. Further to this simulation tool, a method is proposed to assess human emotion using the affective dimension model and physiological signals such as an electrocardiogram (ECG) and photoplethysmography (PPG). The experiments are carried out to verify that humans' felt arousal and valence levels could be measured and translated to different emotions for autonomous driving operations. A per-subject-based classification accuracy is statistically significant and validates the proposed emotion assessment method. Also, a simulation is conducted to verify AGVs' velocity control effect of different emotions on driving tasks.