SMiLE-AI

The core element of research activities in the field of assistive robotics for care is the SMiLE ecosystem for operating heterogeneous robot systems. This involves the development of different control modes that enable targeted command of the robots, adapted to the current operating environment. These control modes are supplemented by the option of controlling the robots remotely (teleoperation). This remote control is intended to enable autonomously operating robot systems to be put into use at an early stage. From a control centre, trained tele-care assistants can remotely control the robots if necessary, should the autonomous capabilities of the robot be insufficient in a given situation. With regard to user-related control options for the robots, different levels of autonomy are considered in the SMiLE ecosystem. In direct control, the robot is operated via a joystick, for example, and "directly" executes the movement commands initiated by the user. In the shared control autonomy level, user commands are adapted by AI-based software to facilitate the operation of the robot. Similar to a lane departure warning system in a car, partial autonomy helps to keep the robot on the right track for the task at hand. This can help, for example, to achieve the necessary precision when bringing the spout of a water bottle over a glass to pour a drink. However, shared control also helps to coordinate all degrees of freedom of the robot, for example, to drive a wheelchair through a door while the robot arm opens the door. At the autonomy level of supervised autonomy, the robot is operated via a graphical user interface, e.g., a tablet app, or via voice commands. Individual actions, such as fetching an object or opening a refrigerator or door, are performed independently by the robot at this level of autonomy. At the highest level of autonomy, a robot should be able to perform complex tasks independently and, if necessary, without direct human command. An adaptable assistance system for everyday care or personal support for people with physical impairments must also ensure flexible and smooth switching from one autonomy level to another. In addition, the autonomy functions must be further expanded for long-term and robust use in real-world environments.