Containerization enhanced systems integration for robotics code development and deployment

Over the decades, robotics deployments have leveraging synergies with rapid in-parallel research and developmental advances in sensing, actuation, simulation, algorithmic control, communication, and high-performance computing among others. Collectively, their integration within a cyber-physical-systems framework has supercharged the increasingly complex realization of the real-time ‘sense-think-act’ robotics paradigm. Successful functioning of modern-day robots relies on seamless integration of increasingly complex systems (coming together at the component-, subsystem-, system- and system-of-system levels) as well as their systematic treatment throughout the life-cycle (from cradle to grave). As a consequence, ‘dependency management’ between the physical/algorithmic inter-dependencies of the multiple system elements is crucial for enabling synergistic (or managing adversarial) outcomes. Furthermore, the steep learning curve for customizing the technology for platform specific deployment discourages domain experts from rapid prototyping and validation of the technological piece. This creates a need for frameworks that can provide adequate compartmentalization for domain experts (to carry out platform agnostic research) and yet permit flexible encapsulation of multiple robotic code deployment architectures (legacy or otherwise). In this work, we explore various facets of these challenges for autonomous operations with a simulated/physical Clearpath Husky robot by developing Robot Operating System (ROS) based Docker containers, that isolate different functions of the robot operations and yet interact with each other in real-time for a synergistic deployment.