Flexible assembly work cooperation based on work state identifications by a self-organizing map

This study presents a method of realizing flexible assembly work cooperation in cases where neither the assembly process, nor the final form of the completed task is pre-defined in advance. To realize such systems, there exists an issue of identifying work states during the assembly, and determine when and what kind of support is necessary. As an approach of solving such issues, identifying the work states from a work model built by self-organizing assembly motions of human is taken. Examples of the work state identifications and a support system, which judges the situational necessity of support and selects whether to hand out or holds assembly parts are shown. The support is carried out based on the work states identified by the self-organized map. Experiments indicate that work state identification by a self-organizing map is effective in flexibly cooperating with human during assembly work.