The foreground software used in 2024 has been made available at https://robotics.cse.unsw.edu.au/gitlab/toyota-hsr/runsweep2024.

The software is described in the README.md file of the git repository. An
excerpt of the readme is provided below:

The ROS packages of the foreground software are described below.– manipulation
• hsrb-interfaces-unsw: A fork of the toyota HSRB interface package
• object-grasper: Package contains services for manipulation actions– msgs
• unsw action msg: Custom msgs for action dispatch
• unsw vision msgs: Custom msgs for unsw vision pipeline– navigation
• gmapping: Fork of hsrb gmapping package used for mapping
• nav utils: Contains navigation scripts such as person-following. Intention
to refactor to action servers
• rosnav: Edited fork of HSRB rosnav repository– planning
• action clients: State machines to dispatch planner outputted actions to
action action servers
• action servers: Servers for all base actions. Containing manipulation
actions at the moment
• state machine: Package used as proof of concept for continuous feedback
using smach library
• tr: In development proof of concept for teleo-reactive planner– speech
• sound stream: Package to publish microphone stream to ROS topic
• speech to text: Vosk based model for speech to speech to text

• NOT INCLUDED:
∗ nlp-rasa: Package containing scripts for nlp using rasa model– tasks
• carry my luggage: State machine/behaviour tree based control for carry
my luggage task using services
• clean the table: State machine based control for clean the table task,
using action clients and action servers
• inspection task: State machine based control for inspection task
• pick up apple: State machine based control for simple integrated vision
and manipulation task using services
• serve breakfast: State machine based control for serve breakfast written
at RoboCup 2024, unchanged and untested since competition. Will be
refactored
• store groceries: State machine based control for store groceries task
written at RoboCup 2024, unchanged untested since competition. Will
be refactored

– tools
• common utils: Scripts that don’t really belong anywhere such as “door
checker” to check if a door has opened– vision
• person-recognition: Person recognition and reidentification inserted in
vision pipeline
• utils: Specific vision nodes used for small tasks such as identifying a bag
handle or a person pointing
• yolov8: Wrapper of yolov8 for ROS– world model
• world-model: World model and scripts used to maintain belief store

We are using an external device for additional processing, compliant with
the rules of the Domestic Standard Platform League: Dell Mobile Precision
Workstation (Intel Core Ultra 9, NVIDIA GeForce 4090, 64Gb RAM)
connected via Ethernet to the HSR, mounted on the standard backpack mount.
UNSW uses the following third party software and libraries for the
competition:– Vision Processing: YOLO, Face Recognition– Grasping: MoveIt, Grasp Pose Detection (GPD)– SLAM and Navigation: GMapping, ROS navigation stack– Speech: Vosk– Database: Postgr