Robot Base Placement Optimization for Pick-and-Place Sequences in Industrial Environments

This work proposes a highly parallelized and efficient optimization procedure for finding the globally optimal robot base placement in industrial assembly tasks. This procedure fully embeds a point-to-point (P2P) trajectory planner to evaluate the placements, considering kinodynamic robot limits, collisions, robot configuration changes, and a cyclic constraint. The investigated optimization goals are cycle-time optimality, time-energy optimality, and adaptivity to dynamically changing throughput. The latter uses an optimization-based method to scale the cycle-time-optimal trajectory in low-throughput situations to reduce energy consumption. The approach is used to optimize the base placement in two robotic work cells, showing a reduction in the cycle time of up to 44% and a reduction of energy consumption of up to 53% in a best-to-worst comparison.