PiELo: A Reactive Infrastructure for Swarm Programming

We present PiELo, a novel programming language for swarm robotics. A robot swarm is a decentralized network of robots working to achieve a common goal. Robot swarms have a multitude of useful applications, including exploration, construction, and mining. When programming swarms, the robots need to respond to changes in the environment. If a sensor reading updates, the robots should respond to the new readings and take action if necessary. This process, known as reactivity, is the essential building block for robots to take appropriate actions and contribute to the swarm's overall goals. Additionally, in a problem unique to swarms, robots need to be able to form a consensus, i.e., they must reach an agreement on important aspects of their environment. Programming these behaviors currently requires a significant effort and expertise from the programmer. PiELo explores addressing these concepts as first-class features of a programming language. When programming in PiELo, reactive primitives enable the programmer to easily write functions that will stay up-to-date as the robot performs its duties. Additionally, PiELo provides multiple means through which individual robots can come to a decision and compute problems together. This enables many foundational swarm algorithms, such as flocking and gradient formation. In our paper, we show the effectiveness of this new language through a series of demonstrations of common swarm algorithms and problems, comparing the required code in PiELo to the same algorithms written in other contemporary robotics languages.