Human-robot interaction with assistive social robots: applications for autonomous navigation and mask recognition in COVID-19 scenarios

Objective: The proposal of an assistive social robot with the ability to autonomously navigate in an environment and with good perceptive ability, which both are important for optimal human-robot interaction. The proposal is made in the context of the Covid-19 global health crisis. Thus the assistive social robot is able to navigate autonomously in an environment and recognize if someone it meets is wearing a mask or not, assisting humans in this capacity to combat the contagion. Background: Technology and artificial intelligence have seen great advancements in the last years and engendered the concept of human-robot interaction. The interaction between humans and robots promotes the support that robots can provide to humans to the best possible level and in the best possible way. The beneficial interaction between humans and assistive social robots requires that the robot meets certain requirements such as independence, good perception, navigation and communicative ability, and adaptation to changing ambiance. Addressing these challenges is important as it allows us to make the most of the potentials of human-robot interaction. The Covid-19 global health situation highlights the importance referred to above. For health reasons, to reduce the chances of infections and the spread of an influenza, assistive social robots would offer great assistance. The robot should be able to autonomously navigate in an environment, checking that people are following health guidelines of wearing a face mask and keeping a safe distance in an indoor setting. Methods: This study proposed a Pepper humanoid robot, an assistive social robot, that can navigate autonomously and perform face mask recognition. It uses a neural network architecture and heuristic path planning algorithm approach to recognise masked and non-masked people and navigate autonomously. Results: The tests performed with the proposed solution produce good results. Pepper is able to navigate autonomously in an environment with or without obstacles in its path and adapts to a dynamic environment. The robot is also able to perform mask recognition with high accuracy. Conclusions: In this work, human-robot interaction with assistive social robot has been studied using the platform of the humanoid robot Pepper. The two key areas of autonomous navigation and face mask recognition have been developed with the Pepper robot. The study has been presented in the context of the global health crisis, namely Covid-19. The importance of assistive social robots in helping to combat the current health challenge has been well identified. The ability of the assistive social robots to play such a role highlights two key areas which are great perceptive ability and autonomy. These robots would be deployed in a real life setting and interact with humans. With this scope, the study intended to contribute to the knowledge and solutions for autonomous navigation and object recognition by assistive social robots, ensuring that they interact optimally with humans, and their performance in terms of assistance jobs is optimal as well.