An Integrated Model of Visual Attention for Homecare Robot with Self-awareness

This paper proposes an integrated model of top-down and bottom-up visual attention with self-awareness for a homecare robot. For mimicking the human attention processes, a robot self-awareness model with fuzzy decision making system is developed and utilized, which is an important improvement on the existing robot attention models. Besides the task-driven object-based biasing, a robot self-awareness model can generate other parts of top-down biases in a robot visual attention model. Some results from the self-awareness model are obtained here. In order to update the weights in robot memory, the learning process is carried out through the Bayes' rule. Three images are tested to evaluate the developed methods. Four types of saliency maps are compared to see how self-awareness can affect robot attention selection.