Emotional Interaction Between Human and Robot

With a goal to encourage task compliance and improve well-being, this research proposes a Markov Decision Process (MDP) model for socially assistive robots (SARs) to enable emotionally meaningful interactions with human users. The model considers the emotional states of the robot as well as the emotional and attentional states of the human user together with the effect of other people in the surroundings, therefore affecting the efficacy of the interaction. An adaptive decisionmaking method is provided by using human attention dynamics and emotional transitions to let the robot dynamically change its emotional state and tactics in real-time. This study focuses primarily on the emotional behavior of the human, the future study will entail the detailed set of actions to be executed by the human. The study incorporates an online learning system to dynamically update transition probabilities, therefore adjusting to fluctuations in human emotional reactions. The model was verified using a simulation-based case study to show its capacity to produce an optimum strategy for persuasive encounters under different environmental settings.