Autonomous Shopping Cart: A New Concept of Service Robot for Assisting Customers

Several lines of research into Robotics currently aim at the development of robots for assisting humans in their everyday lives. These robots, also called service robots, should perform tasks that are dirty, dull, distant, dangerous, or repetitive to humans. This work proposes an autonomous shopping cart to assist customers in supermarkets. Customers, especially those with some disabilities, can have their groceries carried by these robotic carts while they are shopping. Moreover, a system was created to compute the shortest tour for a robot cart that must go through all aisles required by customers based on their shopping list. The system models this problem as an instance of the Traveling Salesman Problem (TSP) and uses a Genetic Algorithm (GA) to find approximate solutions. Other features that provide autonomy to our robot carts, such as mapping, localization and navigation, are also addressed in this paper. Experiments were carried out with simulations in order to evaluate this new shopping cart concept. The final design and architecture necessary to deploy our autonomous shopping carts into real-world environments are also presented.