CommRob

The project will be guided by a defined set of scenarios that exemplify the issues raised when a robot trolley is tasked to operate in a cluttered supermarket type environment.

Scenario S1: the trolley passive mode, using a haptic interface, for a person used to the shopping centre who may need a walking aid

Scenario S2: the trolley active mode, using a haptic interface, for a person who wants more services or who may need a walking aid

Scenario S3: the trolley passive mode, using a multimodal interface, for an “agile” person used to the shopping centre

Scenario S4: the trolley active mode, using a multimodal interface, for an “agile” person who requires more services

The communication model to be used in the proposed project will be based on this discourse model. For this purpose, it needs to support both communication between humans and robots, and communication among robots. In addition, it needs to cover multimodal communication and challenges related to asynchronicity, mixed task-initiative, open-endedness and simultaneous input and output.

An underlying assumption concerning the robot’s interaction design is that it should be based on principles of human-human communication in order to provide an interface that is intuitive and easy to use. The development of the communication platform envisioned in this project also provides another research challenge related to the dialogue design, namely that dialogue models should be generic and reusable. This project will also need to consider the general challenges that we are facing when developing and designing multimodal interfaces for human- robot interaction in situations where the human and the robot are sharing the same environment. We will address these challenges by designing dialogue models based on established principles for human-machine interaction and ensuring that these models are thoroughly evaluated in realistic usage situations.

The control for the mobile assistant robots will be realized with a generalized behaviour-based approach. The base architecture will be capable of managing safety, mobility in a dynamic human-centred environment as well as social behaving. Robot functions as planning, navigation and interaction/communication with humans and other robots will be integrated in the system as special behaviours. Three main goals will be addressed by this architecture:

• Open modularity: due to new application domains or new sensor motor components, new behaviours can easily be integrated and enlarge the abilities of the system.
• Safety and dependability: the control will be capable to cope with robot actions in dynamic environments, disturbances and maybe partially uncertain components. None of these will affect the safety for humans and environment.
• Autonomy and flexibility: autonomous planning, targeted navigation, behaving abilities based on the reliable situation interpretation and intuitive usability in direct interaction will be integrated.