Telepresence systems allow a human operator to act in a remote, inaccessible, dangerous or scaled environment and receive feedback from it without physically being there. Applications range from underwater and space teleoperation to minimally invasive surgery and nano-manipulation. The haptic (motion and force) feedback system - additional to the auditory and visual modality - enables the realistic and efficient manipulation in the remote environment. Ideally, a telepresence system is transparent, i.e. the human cannot distinguish between direct and tele-interaction with the remote environment.
Stability and transparency of haptic telepresence systems over non-ideal communication networks are challenging control problems.
The human operator and the remote environment are both unknown dynamical systems in a global closed control loop. Additionally, the human operator plays a special role in the closed control loop, in that he/she a) is coupled with the telepresence system through energy exchange, affecting it with his/her dynamics, and b) perceives the remote environment through the telepresence system with the degree of transparency strongly influencing his/her evaluation thereof. Consequently, the human operator should be an inextricable component of the control design process. The global control loop is closed through a communication network with time delay and other network-induced unreliabilities, such as packet loss. Without further control measures this may degrade the transparency of the telepresence system and even destabilize it. An additional challenge are the limited network resources, e.g. in underwater or space teleoperation. The ultimate goal would be to mitigate this, without impairing the system stability or distorting the human user experience.
This dissertation provides a comprehensive control-design concept for telepresence systems over non-ideal communication networks by bringing the perspectives of stability, transparency, and network utilization together using a common point of reference: the human operator.
A novel control framework is proposed, in which the exact modeling of dynamics is circumvented through the use of the theory of dissipative systems.
We manage to guarantee stability with communication unreliabilities and to improve the transparency of the system despite having only approximate knowledge on the human operator's (or other subsystems') energetic behavior.
Unique compared to existing literature is the utilization of the grip force of a human operator. Using this haptic measure, the human arm dynamics can be estimated online and used for the parametrization of the control mechanisms.
Finally, with regard to the communication-related challenges, an efficient method of transmitting haptic data is developed considering human factors, such as the haptic discrimination abilities of the human and the task performance. Haptic data can be reduced significantly, without impairing the human's perception of the remote environment.
The proposed methods are, throughout the thesis, extensively evaluated on telepresence systems with multiple degrees-of-freedom.
«
Telepresence systems allow a human operator to act in a remote, inaccessible, dangerous or scaled environment and receive feedback from it without physically being there. Applications range from underwater and space teleoperation to minimally invasive surgery and nano-manipulation. The haptic (motion and force) feedback system - additional to the auditory and visual modality - enables the realistic and efficient manipulation in the remote environment. Ideally, a telepresence system is transparen...
»