The Pre-ROSCon Singapore Community Hackathon is designed to help new and aspiring contributors become active participants in the global ROS open-source ecosystem. Through a guided, hands-on programme, participants will learn how to build ROS 2, Gazebo, and Open-RMF from source, understand collaborative development workflows, and gain practical experience contributing to real-world robotics software.

Experienced mentors from the ROS community will provide hands-on support throughout the session, helping participants configure their development environment, identify suitable Good First Issues, and submit their first Pull Requests. Whether you are a student, researcher, software engineer, or robotics practitioner, the hackathon provides a welcoming environment to develop practical open-source skills while building lasting connections within the international ROS community.

ROS has become the foundation for robotics research and increasingly serves as the software platform behind commercial robotic systems. However, transforming a successful prototype into a reliable production deployment requires far more than functional software. Organisations must address interoperability, maintainability, operational ownership, cybersecurity, lifecycle management, testing, and long-term support while balancing business realities with engineering constraints.

This panel brings together leaders representing open source, translational research, enterprise deployment, and robotics product development to share their experiences building and operating production-grade ROS systems. Drawing on deployments across healthcare, security, logistics, and industrial automation, the discussion explores the technical decisions, organisational capabilities, and ecosystem collaboration needed to deliver robots that operate reliably in real-world environments.

As robots become increasingly autonomous, coordinating multiple fleets from different vendors becomes one of the biggest challenges in large-scale deployments. Independent fleet management systems often lack awareness of one another, creating conflicts, deadlocks, and operational inefficiencies.

Open-RMF addresses this challenge by providing an open software integration framework that enables heterogeneous robots to safely share infrastructure such as lifts, doors, and common spaces. Originally designed for healthcare and airport environments, Open-RMF is now expanding into manufacturing, warehousing, and industrial automation.

This session introduces the next generation of Open-RMF, featuring a fully modular architecture that allows users to choose traffic management, task orchestration, and execution components best suited to their deployments. The presentation also highlights the project's growing open-source community and explains how developers, integrators, and organisations can contribute to the future of robot interoperability.

Open-RMF was originally created to enable heterogeneous robots to operate seamlessly within healthcare environments, coordinating fleets from different manufacturers while integrating with hospital infrastructure such as lifts, automatic doors, and building systems.

Drawing on deployments across hospitals and nursing homes in Singapore, CHART shares practical lessons from operating Open-RMF in production healthcare environments. The talk examines technical and operational challenges encountered during large-scale deployments, including high-demand workflows and pandemic-response scenarios, and discusses how these experiences exposed new requirements for fleet coordination, resilience, and scalability.

These deployment insights directly shaped RMF2.0 (Healthcare), introducing capabilities such as Smart Sensor Nodes, Dynamic Goal Re-planning, Human–Robot Lift Sharing, and Fleet & Map Filtering to improve robustness and operational efficiency.

The session concludes by demonstrating how continuous deployment feedback continues to drive the next generation of Open-RMF and healthcare robotics interoperability.

Autonomous Mobile Robots (AMRs) have become an integral part of hospital logistics, automating routine delivery and transport tasks. As healthcare workflows become increasingly sophisticated, there is growing demand for mobile manipulators—robots that combine autonomous navigation with robotic arms to perform physical interactions with their environment.

Integrating these systems into an existing Open-RMF deployment presents challenges that extend beyond simply adding another robot fleet. Manipulation introduces dynamic robot footprints, new task semantics, tighter safety requirements, and the need to coordinate navigation, fleet management, and manipulation controllers within a shared operational environment.

This presentation shares experiences from a collaborative project between the Centre for Healthcare Assistive & Robotics Technology (CHART) and the Singapore University of Technology and Design (SUTD) to integrate mobile manipulators into an Open-RMF-managed healthcare ecosystem. Based on clinical use cases developed with stakeholders at Changi General Hospital, the session examines how manipulation capabilities can be incorporated into existing hospital workflows while maintaining interoperability across heterogeneous robot fleets.

Attendees will gain practical insights into dynamic footprint management, manipulation-aware task workflows, vendor-agnostic integration strategies, and coordination between Open-RMF, autonomous navigation, and robot manipulation systems. The talk concludes with lessons learned and future directions for enabling mobile manipulators as first-class participants in large-scale, interoperable healthcare robot deployments.

Industrial facilities often deploy robots and automation systems from multiple vendors, making system integration increasingly complex, costly and difficult to scale.

Seeded by Singapore's National Robotics Programme and A*STAR ARTC, RMF Industrial (RMF2) is an open-source robot middleware platform designed for manufacturing, warehousing and laboratory automation. The platform provides modular capabilities for fleet management, task orchestration, traffic coordination, simulation, device integration and VDA 5050 interoperability.

This presentation introduces the architecture of RMF Industrial, its modular design philosophy, and its alignment with the Next-Generation RMF Interfaces. Through real industrial use cases, attendees will learn how open robot middleware enables robot vendors, system integrators and facility owners to deploy interoperable robotic systems while reducing development effort and avoiding vendor lock-in.

Long-horizon bimanual manipulation requires a humanoid robot to understand task instructions, perceive relevant objects, determine the correct action sequence, allocate actions between both arms, and execute coordinated movements in dynamic environments.

This presentation introduces an embodied AI framework that combines vision-language scene understanding with task planning and humanoid robot execution. The system converts visual observations and natural-language instructions into coordinated manipulation behaviours for tasks including object sorting, stacking, handling, and cross-arm transfer.

The session discusses the overall system architecture, embodiment-aware planning strategies, the role of vision-language models, and practical deployment challenges including perception uncertainty, robot reachability, inter-arm interference, execution failures, and error accumulation during long-horizon manipulation.

Attendees will gain insights into how embodied AI is enabling more adaptive, intelligent, and executable humanoid robotic systems capable of operating in complex real-world environments.