Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

The Director: A Composable Behaviour System with Soft Transitions

Ysobel Sims, Trent Houliston, Thomas O'Brien, Alexandre Mendes, Stephan Chalup

TL;DR

The paper tackles the limitations of existing robotic behaviour systems, notably the need for smooth transitions, environment-aware multi-implementation choices, and strict resource control. It introduces The Director, a pub/sub-based framework built around Providers and Tasks, with Provider Groups, and semantics such as When and Causing to enable soft transitions and conditional behavior changes. Key contributions include soft-transition support, the ability to have multiple implementations for a single task via provider groups, and a strict resource-control mechanism validated through RoboCup deployments and automated tests. The framework supports rapid development, safe integration with higher-level reasoning like LLM-powered strategies, and demonstrates practical effectiveness in humanoid soccer contexts, indicating broad applicability to autonomous robotics.

Abstract

Software frameworks for behaviour are critical in robotics as they enable the correct and efficient execution of functions. While modern behaviour systems have improved their composability, they do not focus on smooth transitions and often lack functionality. In this work, we present the Director, a novel behaviour framework that addresses these problems. It has functionality for soft transitions, multiple implementations of the same action chosen based on conditionals, and strict resource control. The system was successfully used in the 2022/2023 Virtual Season and RoboCup 2023 Bordeaux, in the Humanoid Kid Size League. It is implemented at https://github.com/NUbots/DirectorSoccer, which also contains over thirty automated tests and technical documentation on its implementation in NUClear.

The Director: A Composable Behaviour System with Soft Transitions

TL;DR

The paper tackles the limitations of existing robotic behaviour systems, notably the need for smooth transitions, environment-aware multi-implementation choices, and strict resource control. It introduces The Director, a pub/sub-based framework built around Providers and Tasks, with Provider Groups, and semantics such as When and Causing to enable soft transitions and conditional behavior changes. Key contributions include soft-transition support, the ability to have multiple implementations for a single task via provider groups, and a strict resource-control mechanism validated through RoboCup deployments and automated tests. The framework supports rapid development, safe integration with higher-level reasoning like LLM-powered strategies, and demonstrates practical effectiveness in humanoid soccer contexts, indicating broad applicability to autonomous robotics.

Abstract

Software frameworks for behaviour are critical in robotics as they enable the correct and efficient execution of functions. While modern behaviour systems have improved their composability, they do not focus on smooth transitions and often lack functionality. In this work, we present the Director, a novel behaviour framework that addresses these problems. It has functionality for soft transitions, multiple implementations of the same action chosen based on conditionals, and strict resource control. The system was successfully used in the 2022/2023 Virtual Season and RoboCup 2023 Bordeaux, in the Humanoid Kid Size League. It is implemented at https://github.com/NUbots/DirectorSoccer, which also contains over thirty automated tests and technical documentation on its implementation in NUClear.
Paper Structure (23 sections, 2 figures, 2 tables)

This paper contains 23 sections, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Framework
  3. Providers
  4. Provider Groups:
  5. Provider Types:
  6. Needs:
  7. When:
  8. Causing:
  9. Tasks
  10. Root Tasks:
  11. Priority:
  12. Done Tasks:
  13. Idle Tasks:
  14. Data
  15. Uses:
  16. ...and 8 more sections

Figures (2)

  • Figure 1: This figure shows a Director graph for a humanoid robot playing soccer. It shows active (solid blue), blocked (dashed red), and inactive (dotted black) Providers and Tasks. The values of Tasks indicate their priority.
  • Figure 2: An example of priority, where the kick challenges the walk for control of the legs and succeeds in taking control due to its higher priority.