II-New: CyPhyHouse: A Laboratory for Evolving Distributed and Mobile Cyber-Physical Systems Research

University Of Illinois At Urbana-Champaign, Urbana IL

Investigators

Abstract

This project develops CyPhyHouse: an innovation playground for enabling rapid development and deployment of distributed and mobile robotic systems. Distributed robotic systems are at the frontier of research in manufacturing, transportation, logistics, and exploration. Yet, managing and experimenting with distributed robotic platforms can be a laborious and error-prone undertaking. Deploying a new algorithm on an existing hardware platform can take weeks; porting an application to a new platform can take months. This high overhead of deploying robotic applications takes a toll on the robustness of the scientific claims made in this area. CyPhyHouse provides application programming interfaces (APIs) and an easy to use software environment which reduces the above efforts by orders of magnitude. This new infrastructure provides easy to use, open source, software and hardware platform for rapidly programming, simulating, deploying, and visualizing a rich family of applications on heterogeneous fleet of vehicles. It thus will enable large-scale experimental research needed for developing computing systems that support and rely on coordinated and controlled mobility, for example, in the study of mobility of WiFi access points, debuggers for mobile systems, and augmented-reality with real and virtual Unmanned Autonomous Vehicles. The hardware infrastructure consists of a collection of about mobile vehicle platforms (a mix of ground vehicles and aerial vehicles). The software infrastructure consists of APIs for coordination and control of the heterogeneous fleet based on distributed shared-memory (DSM) and reach-avoid functions. These APIs are integrated in an extension of a familiar programming language (e.g., C++ or Java) and have implementations on all the hardware platforms. Supporting software tools are developed for compilation, simulation, logging, and monitoring that help with development and deployment of application programs. The infrastructure enables a researcher to quickly deploy experiments involving dozens of vehicles and visualize them. The same high-level application programs will be portable across different hardware platforms. The platform enables large-scale, repeatable experimental research needed for developing computing systems that support and rely on coordinated and controlled mobility. All the software and hardware specifications will be open-sourced and made available to the broader community. In several graduate courses, CyPhyHouse will provide students with a realistic platform for projects. It aims to help train graduate students and provide research opportunities for undergraduates in areas like cyber-physical systems, mobile and distributed computing, and multi-agent systems.