Evaluation of assembly routines with multitasking execution in a physical robotic cell

Abstract

This research covered the development of multitasking execution programs and evaluation of twelve assembly sequences, in terms of efficiency and effectiveness, applied to a robotic cell under two software control methods. The assembly sequences were defined via analytical methods and verified later with a physical simu I at ion. The analytical methods used to define the sequences were the SPT rule, the LPT rule and the Branch and Bound algorithm. The software control methods were a single task execution program, and a multitask execution program. The single task execution programs performed all the activities in a sequencial mode. The multitask execution program allowed two activities to run simultaneously. The physical simulation was performed in a robotic cell containing two TeachMover robot arms, a central assembly area, and two bin cells built with Fischertechnik components. The part assembled was a representation of a circuit board with four microchips made of machinable wax. The control software was coded using ARMBASIC for the robot arms and TurboBASIC for the main program. The analytical results showed that using a single robot, the sequence with the best completion time was generated with the Branch and Bound algorithm. Also, this result was verified with the physical simulation that generated the best completion time using the Branch and Bound algorithm. Using two robot arms, the analytical results showed that the Branch and Bound algorithm under a multitask execution mode generated the best assembly sequence. In this case, the physical simulation showed different results. The best sequence found with the physical simulation was an adjusted sequence, running under a multitask execution mode, that removed the physical conflicts to avoid collisions in the assembly area. The physical interference could not be observed by the analytical methods. Therefore, the use of physical simulation to evaluate robotic motions is recommended.