Take-Over Strategies and their Effect on Control and Recording of Automated Camera Tracking Shots

There is a multitude of tools aiming to support camera operators in their work (such as dollys or cranes). Recent developments extended these tools with motors and control systems that allow computer controlled motion of the camera [2,8,3]. The tools available on the market mainly follow the approach of either manual remote control [4] or automatic control [6]. Research in the field of robotics found a trade-off between unpredictability and workload that comes along with the introduction of automation [5]. In Consequence less workload (so to speak simpler systems) come at price of less predictable results.
In the context of professional cinematography with the goal of crafting a certain aesthetic the approach favoring automation might introduce easier operation, but also very likely not good enough or controllable results. Smart systems and human operators perform well at different tasks [7], so that a dynamic change of control seems the favorable approach in consequence. However only few examples of systems allowing a fluid take-over of the motion control exist [1]. In the context of cinematography, the very act of the take-over from computer to human control should at best be invisible in the recorded material. A clever mechanism therefore should ideally address human control behaviours (e.g. oversteering).
This thesis aims at identifying thoses issues and developing and evaluating concepts that address the found issues. The followed approach will be rapid and user-centered working in near life circumstances making use of a motorized and remote wirelessly remote controlled camera slider.

• Literature review on related work
• Identification of human control issues in taking-over of motion control
• Concept development / strategies addressing the identified issues
• Implementation and evaluation of the developed concepts
• Written thesis and presentation of your work