Abstract: In this talk I will discuss the development of autonomous mobile robots that navigate outdoors in unstructured environments under the remote supervision of human supervisors.  We are interested in two inter-related problems that focus on satisfaction on human-driven needs: assuring that the robot is trusted by a human supervisor, and collecting data according to learned behavioural cues.
In the case of trust, we focus on building models of human-robot trust when a human supervisor closely monitors an autonomous vehicle, such that the robot can quantitatively model, and then modulate, the amount of trust that a operator may have.  In this case we focus on human-robot interaction and classical learning methods.  In the context of data collection, we focus of problems where an autonomous vehicle is deployed to collect data in the air or underwater, and human guidance takes the form on example data used to teach the robot how to behave.

Abstract: Mobile manipulators with a large number of degrees of freedom have increasingly become of interest to industrial and service robotics, because of their unlimited workspace and their remarkable versatility. Robots have to work in unstructured environments together with humans, where safety aspects have to be taken into account and fast reactions to the dynamically changing environment is needed. This leads to the necessity of controlling the unified robotic system and exploiting their redundancy. Intuitive operation and programming has to be tackled to really use the flexibility and capability these complex robot systems are offering.