Acquisition of Product Purchasing Behavior by Robots Based on Problem Decomposition and Parallel Reinforcement Learning
摘要
Some elderly people living in depopulated areas struggle to secure essential supplies. In these regions, poor transportation infrastructure makes it difficult to maintain delivery services. Introducing shopping robots to replace human shoppers could help sustain delivery services. Recently, convenience stores with a wide variety of products have become common. Shopping can be divided into two main stages: leaving home and reaching the store, and purchasing items inside the store after arrival. This paper focuses on the latter—navigating inside the store efficiently to find and purchase desired products. To demonstrate effectiveness, the study simulates a robot starting near the store entrance, searching for the target item, taking it to the checkout, and completing the payment process.