Optimal robot positioning for sustainable process execution

Kurzfassung

In this paper, the methodology to implement an inference algorithm based on the robot parameters for process execution on different Key Performance Indicators (KPI’s) is presented. This work is part of holistic objective to implement a system to perform virtual validation of automated process with robots, where the impact of relative position of robot to required task and vice-versa could be inferred virtually. For the above implementation, it is proposed to develop the digital model of the necessary infrastructure in the simulation environment. To start with, computational data from simulation is stored to a database, which is then analysed and appropriate regression based inference model is formulated. To ensure that the digital model imitates the real-time system, the feedback data from hardware execution is used to improve the parameters of regression model. With this implementation, the digital model would represent the digital twin (DT) of the hardware under consideration. The whole execution is performed on the pre-assembly cell at the Institute for System Architectures in Aeronautics, Hamburg. Use case for the digital twin implementation is the pre-assembly of overhead structural truss that assists in realizing the modularized cabin assembly process. Implementation of simulation model for cabin assembly is a two-fold approach, where the robot localization for reachability is computed followed by computation of joint trajectories. From the obtained trajectories, the energy and time consumed by the robot for a given task is calculated. The computed information is then stored in a database, which is then fed to an inference algorithm. Implementation of the algorithm is desired based on the time taken for path computation, and using this algorithm the optimal robot location and joint angles for any new unknown task could be com- puted.