Improvement of the motion behaviour of robots through path planning
To reduce cycle times, high-speed webs are to be realized in many industrial applications. The definition of both the geometric path and the time law is usually independent of the robot-specific properties. Rectangles are often defined within the working area where a path of straight lines, circles or spline functions can be programmed. Additional criteria such as collisions, avoidance of areas in the workspace with poor properties, etc. are rarely taken into account.
Due to the increasing performance capability of today's drive technology, however, the aim is to adapt motion sequences as optimally as possible to specific requirements. In order to meet this requirement, an optimum definition of the geometric path and the motion profile must be determined, taking into account sensible optimization criteria (time-optimal, optimum performance, etc.) as well as robot-specific limits (maximum drive torques and forces, maximum drive speeds, etc.).
While the optimization of motion profiles  and the generation of time-optimal trajectories, see Figure 1, along predetermined paths have been successfully treated in several works, the planning of the geometric path for serial as well as for parallel and hybrid robots remains an elaborate and time-consuming search. Due to the non-linear relationships between drive and output quantities, inhomogeneous distributions of the maximum achievable forces and speeds over the working area result for this type of robot. Accordingly, a method must be specified that delivers a web within the working area that avoids areas with poor properties and thus enables optimal use of the robot.
The methods used in the literature are based on a prior calculation of these properties in different points, cells or intervals of the workspace and the storage of this information in the form of graphs. Subsequently, classical search algorithms, such as the A* algorithm or the Dijkstra algorithm, are used to generate the optimal path with regard to the robot-specific limits (size of the workspace, shape of the workspace, maximum performance, avoidance of collisions, etc.). This type of search is very time-consuming and requires a lot of storage space, especially for robots with multiple degrees of freedom. This research project implements and compares several approaches to railway planning. These approaches will be tested experimentally using the robots (KUKA KR125, plane and spatial parallel robots) available at the Institute for Mechanism Theory, Machine Dynamics and Robotics as an example. The aim is to develop a process tailored to these robots.
 Corves. B., Bonsch, F., Nefzi, M.: Optimization of motion profiles. In: VDI (Ed.): Motion Technology. Solution of motion tasks with couplings, cam gears and controlled drives (Getriebetagung Fulda, September 19-20,2006 / VDI-Gesellschaft Entwicklung, Konstruktion, Vertrieb). VDI Reports No. 1966, Dusseldorf: VDI-Verlag, 2006, S. 93-110.