Advancing The Robotics Software Development Experience: Bridging Julia's Performance and Python's Ecosystem

Gustavo Nunes Goretkin,Joseph Carpinelli,Andy Park

2024-06-06

Abstract:Robotics programming typically involves a trade-off between the ease of use offered by Python and the run-time performance of C++. While multi-language architectures address this trade-off by coupling Python's ergonomics with C++'s speed, they introduce complexity at the language interface. This paper proposes using Julia for performance-critical tasks within Python ROS 2 applications, providing an elegant solution that streamlines the development process without disrupting the existing Python workflow.

Robotics

What problem does this paper attempt to address?

The problem discussed in this paper is how to balance usability and runtime performance in robot software development. The traditional approach is to use two languages, Python (usability) and C++ (high performance), but this leads to the complexity of a multi-language architecture. The paper proposes a new approach, which is to use Julia in Python's ROS 2 applications to handle performance-critical tasks, in order to simplify the development process and maintain the continuity of the existing Python workflow. The paper points out that although Julia has the characteristics of high performance and usability, its ecosystem is not yet mature and lacks resources such as ROS 2 client libraries, which limits its application in robotics. By interacting with Julia and Python, the dependence on C++ can be reduced, leveraging the performance advantages of Julia while reducing system complexity and performance overhead. The paper also demonstrates Julia's significant advantage in runtime efficiency through a case study comparing the performance of FlexIK.jl written in Julia and the inverse kinematics solver flex ikpy written in Python. Finally, the paper suggests that future work should focus on expanding Julia's ecosystem in the ROS 2 environment to fully exploit its potential in robot programming.