Tutorials

Cooperative cyber-physical systems with human in the loop

Prof. Ruzena Bajcsy,  University of California, Berkeley, and Center for Information Technology Research in the Interest of Science (Director Emeritus of CITRIS).

In this presentation we will explore development of predictive models of interaction between Cyber-physical systems and people during a given task. We will model the Human physical activities using robotic technologies embodied in kinematics and dynamics of the human body structure. Then we will explore the predictive models of the human and a robotic systems during a collaborative task. The control of such a coordinated activities must entail communication of the intent of the human to the robot and vice versa. The parameters which are communicated have to be transformed into the task space since the degrees of freedom and the forces of the human and the robotic system are very different. We will demonstrate these concepts, algorithms on two different applications, one in the Driver driven car that is negotiating the road with an autonomous car, the other in health care domain, where elderly subjects are coached to perform exercises either by an avatar coach or by a human coach who is available remotely.

Human-centered haptic shared control: one design philosophy across different applications

Prof. David Abbink, TU Delft.

 Manual control tasks­ like driving or flying ­ are prone to human errors, which may cause accidents. A popular solution is to increasingly automate the task, pushing humans into a more supervisory role. While this may be fine in a fully predictable environment, literature has shown persistent issues with human-automation interaction when unexpected situations occur: lack of situation awareness, overreliance, difficulty in handing back control etc. In other words, the conventional solution helps to prevent certain accidents, but causes others. In this talk David Abbink aims to show that the main cause of these problems lies in a lack of intuitive communication: the human doesn’t fully understand the system, and vice versa. He will argue that the solution for the near future lies in shared control, where human and system will flexibly and dynamically share control over the task. He proposes haptics (the sense of touch) as a powerful channel for communication. The talk will present case studies of haptic shared control developed and evaluated over a wide variety of areas, with a focus on real-world applications of the fundamental research. Topics will range from a haptic driver support system for car-following (marketed by Nissan as Deceleration Control Assist in 2008) and steering (patented, and in development), to ongoing research concerning shared control of intelligent remote robot arms operating in complex hazardous environments (maintenance of nuclear reactors, deep-sea mining, space).