Knowledge-Enabled Robotic Agents for Shelf Replenishment in Cluttered Retail Environments

摘要: Autonomous robots in unstructured and dynamically changing retail environments have to master complex perception, knowledgeprocessing, manipulation tasks. To enable them act competently, we propose a framework based on three core components: (o) knowledge-enabled perception system, capable of combining diverse information sources cope with occlusions stacked objects variety textures shapes, knowledge processing methods produce strategies for tidying up supermarket racks, the necessary skills confined spaces arrange semi-accessible rack shelves. We demonstrate our an simulated environment as well real shopping using PR2 robot. Typical products are detected rearranged rack, what was found be misplaced items.