Institute of Mechatronic Systems Research Research Group: Robotics & Autonomous Systems Research Projects
Generation of Task Specific Robot Manipulators via Combined Structural and Dimensional Synthesis

Ongoing Research Projects

Generation of Task Specific Robot Manipulators via Combined Structural and Dimensional Synthesis

Team:  M. Sc Moritz Schappler
Year:  2018
Date:  01-01-18
Funding:  German Research Foundation (DFG)
Duration:  2018 - 2022

The increasing utilization of robotic systems and automation technologies in new areas requires the development of task-specific manipulators in order to perform a given task with best possible performance (e. g. small robot size, low weight, high accuracy, high acceleration/velocity, low energy consumption). Besides industrial robotics (especially in small and medium-sized companies and in the special purpose machinery manufacturing), some examples of these new areas are medical, nursing, service, and space robotics. Traditional robot manipulators are usually designed for a wide range of applications and, therefore, are often oversized. In some cases, the mechanisms cannot fulfill the aforementioned requirements or even perform the desired movement.

The goal of the proposed project is the research and development of new methods for the dimensioning and configuration of robot manipulators. This will allow for the generation of robots with task-specific characteristics by a synthesis of all suitable architectures and a subsequent optimization of the kinematic parameters. During the optimization process, the dynamics of the robot will be considered. Additionally, the sensitivity of the generated solutions will be evaluated. The present project focusses on non-redundant, serial and parallel kinematics.

The first project phase applies the result of the structural synthesis of serial chain robots from a previous project on the synthesis of parallel mechanisms. The method for the generation of parallel kinematics mechanisms includes the reduction of the set of solutions by kinematically identical structures that only differ in their description ("isomorphisms").

Subsequently the dimensions of all task-fulfilling serial and parallel robot structures from the structural synthesis will be optimized ("dimensional synthesis").
This is based on the automatic modeling of the kinematics and dynamics of the robot structure according to useful optimization parameters. Then it is possible to calculate kinematical and dynamical performance indices as well as cost functions for the optimization process.

A result of this combined structural and dimensional synthesis are optimally dimensioned robots for the demanded task with different structures, from which the one that is suited best can be chosen.
With this approach also the synthesis of new, previously unknown, robot types is possible. Further, known robot structures can be dimensioned optimally or the most suitable of available robots can be chosen.

As result of the optimization, a mechanism, as well as the equations of its respective model suitable to apply in the current control architectures is obtained. Wide knowledge in multibody systems is not needed.

In a second term of the project especially robots in tasks with five out of six possible Cartesian degrees of freedom are investigated. The redundancy of the sixths degree of freedom is particularly mathematically challenging and corresponds to a rotation around the axis of the tool mounted at the robot end effector. It appears in tasks like milling, drilling or welding. If the redundancy is already included in the synthesis this allows the selection of robot structures that exploit the redundancy to full extent and can by this means e.g. be more energy efficient.

Past Research Projects

Generation of Task Specific Robot Manipulators via Combined Structural and Dimensional Synthesis

Team:  M. Sc Moritz Schappler
Year:  2018
Date:  01-01-18
Funding:  German Research Foundation (DFG)
Duration:  2018 - 2022

The increasing utilization of robotic systems and automation technologies in new areas requires the development of task-specific manipulators in order to perform a given task with best possible performance (e. g. small robot size, low weight, high accuracy, high acceleration/velocity, low energy consumption). Besides industrial robotics (especially in small and medium-sized companies and in the special purpose machinery manufacturing), some examples of these new areas are medical, nursing, service, and space robotics. Traditional robot manipulators are usually designed for a wide range of applications and, therefore, are often oversized. In some cases, the mechanisms cannot fulfill the aforementioned requirements or even perform the desired movement.

The goal of the proposed project is the research and development of new methods for the dimensioning and configuration of robot manipulators. This will allow for the generation of robots with task-specific characteristics by a synthesis of all suitable architectures and a subsequent optimization of the kinematic parameters. During the optimization process, the dynamics of the robot will be considered. Additionally, the sensitivity of the generated solutions will be evaluated. The present project focusses on non-redundant, serial and parallel kinematics.

The first project phase applies the result of the structural synthesis of serial chain robots from a previous project on the synthesis of parallel mechanisms. The method for the generation of parallel kinematics mechanisms includes the reduction of the set of solutions by kinematically identical structures that only differ in their description ("isomorphisms").

Subsequently the dimensions of all task-fulfilling serial and parallel robot structures from the structural synthesis will be optimized ("dimensional synthesis").
This is based on the automatic modeling of the kinematics and dynamics of the robot structure according to useful optimization parameters. Then it is possible to calculate kinematical and dynamical performance indices as well as cost functions for the optimization process.

A result of this combined structural and dimensional synthesis are optimally dimensioned robots for the demanded task with different structures, from which the one that is suited best can be chosen.
With this approach also the synthesis of new, previously unknown, robot types is possible. Further, known robot structures can be dimensioned optimally or the most suitable of available robots can be chosen.

As result of the optimization, a mechanism, as well as the equations of its respective model suitable to apply in the current control architectures is obtained. Wide knowledge in multibody systems is not needed.

In a second term of the project especially robots in tasks with five out of six possible Cartesian degrees of freedom are investigated. The redundancy of the sixths degree of freedom is particularly mathematically challenging and corresponds to a rotation around the axis of the tool mounted at the robot end effector. It appears in tasks like milling, drilling or welding. If the redundancy is already included in the synthesis this allows the selection of robot structures that exploit the redundancy to full extent and can by this means e.g. be more energy efficient.