Experimental setup for collaborative robots using RaspberryPi scale model cars

Nowadays, it has become clear that autonomous devices and vehicles play an important role in our society and will keep providing dramatic benefits to many industries in the future. Besides an increased human comfort or improved safety, autonomous systems now benefit from tremendous connectivity and computational capacities, allowing them to combine their individual strengths to effectively achieve specific tasks. For the reason that these large systems are difficult to control and often geographically distributed, they have the particularity of being operated in a decentralized manner. In order to make the best of these technologies, many approaches have been developed to model the behaviour of autonomous entities, from single-agent to large collaborative multi-agent systems. Demonstrating evidence of these systems practicability not only calls for accurate simulation models, but also for real experiments highlighting their effective performance and reliability, thus for the development of small-scale experimental setups able to partly reproduce large-scale use cases. In this context, the goal of this master thesis is to lay the foundations of such a small-scale experimental setup. In accordance with the lab research activities, the setup aims to provide the building blocks for developing mobile objects tracking and autonomous vehicles collaboration experiments. This task was performed by deploying three SunFounder PiCar-S, RaspberryPi RC cars programmable in Python. These RC cars, referred to as « picars » in this document, were eventually programmed using the BDI (Belief-Desire-Intention) framework, a widely used robots software model inspired by human practical reasoning. The objective of this document is to expose the concepts and motivations which led to this programming model, justify implementation choices, and put them in perspective with the requirements of the experimental setup. Proof-of-concept is achieved through simple practical use cases.