The predicted massive update of mobile robots in the coming years [IFR World Robotics, oktober 2018] depends on sufficient handling of safety aspects - not just on the physical side but equally on the mental side. The latter is often greatly overlooked in healthcare, industry and the service sector alike, but is often key to successful robot integration and adoption. A way of addressing 'mental safety' is by enabling the robots to be 'context aware' - i.e. have an understanding of its surroundings so it can act safely and according to the required social setting. This project will attempt to elucidate the potential of some of such socially context-aware mobile robots, by addressing concrete end-user cases and testing technological components that will support it.
This project will utilize object detection to implement a social aware perception module that enables compliant navigation for mobile robots. The project will be exploring the use case of navigating in areas where humans are working and interacting with the goal of minimizing distraction. The use of a social context-aware robot can be applied in sectors were human-robot interaction is unavoidable, including but not limited to healthcare, construction, manufacturing, industry, etc.
Most of the current mobile robotic solutions uses path planning with the goal of calculating the shortest path to a specific location. Without any additional knowledge regarding human poses and social context, the robots will unknowingly interrupt human communication and at the same time breach their personal space by getting too close.
With novel artificial intelligence (Al) and computer vision algorithms, it is possible to detect poses (position and orientation) of humans in the surrounding environment. By extrapolating pose information, the navigation system will be able to avoid interruption of social contexts and increase the level of safety. Thus, allowing for smoother and compliant navigation of mobile robots in environments containing people.
Furthermore, ensuring that mobile robots respect the social interaction space of humans will advance the implementation in areas, where acceptance of mobile robots is harder to achieve. Examples of these areas are elderly care centres and hospitals, where both patients and workers can easily be disturbed by mobile robots, or a warehouse where a robot must execute an emergency stop in the presence of humans.
