SALAMANDER

SYSTEM OVERVIEW

Firefighters typically spend 3–4 minutes setting up hoses, while the Awareness of Situation (AOS) process can take up to 5 minutes. In smoke-filled environments, this requires entering blind to locate the fire source, increasing the risk of delays and error.

SALAMANDER introduces a parallel workflow — autonomously assessing the environment during setup and providing clear directional guidance before entry.

01. ARRIVAL ON SITE

SALAMANDER is housed in a Pelican case alongside a portable ground station. Upon arrival, firefighters set up the station, connect to the web interface, and deploy the robot to begin autonomous exploration.

02. DEPLOYMENT

After multiple iterations, the team arrived at this exterior chassis design for its aerodynamic form, smooth organic lines, and porous structure that allows ventilation for the battery packs.

MATERIALITY

3D printing can be used for rapid prototyping and iterative development, offering faster turnaround compared to outsourced metal fabrication. The concept prototype uses PETG for its moderate heat resistance and safe printing properties, while the final deployment version adopts an aluminium sheet metal frame with heat-resistant coating for improved strength and durability.

BATTERY POWER SYSTEM

Under maximum load, the system achieves approximately 60 minutes of operation, with an additional 30-minute buffer for extended use.

03. EXPLORATION

Designed for adaptability in unpredictable environments, the system supports both known and unknown conditions through a dual-mode navigation approach. In mapped environments, it uses structured navigation for efficient path planning and localisation, while in unknown or compromised spaces, it switches to reactive, sensor-driven exploration that requires no prior data. This ensures continuous operation without interruption, regardless of available information.

04. THERMAL DETECTION

A thermal override algorithm continuously monitors temperature data to detect high-heat anomalies. Once identified, it calculates the heat source’s position and overrides navigation controls, directing the robot to orient and move toward the fire source in real time.

05. DATA RELAY

The interface is designed for high-stress environments, presenting critical information through three primary zones: key mission metrics, a real-time tactical map, and live telemetry with control functions.

Operators can plan and execute missions directly on the map by setting the robot’s position, orientation, and target location, enabling precise and intuitive navigation.

During operation, the system provides continuous situational awareness through live position tracking, breadcrumb trails, environmental monitoring, and clear visual alerts for safety-critical conditions such as temperature and battery levels.

IMPACTS

SALAMANDER addresses a critical gap in fireground operations by reducing the need for blind entry in low-visibility environments. By autonomously assessing conditions ahead of firefighters, it provides real-time spatial and thermal information that enhances situational awareness and supports faster, more informed decision-making.

This shift from reactive to informed response improves operational efficiency while significantly reducing risk to frontline responders, enabling safer and more coordinated firefighting efforts.

FUTURE

While currently focused on fireground reconnaissance, SALAMANDER is designed as a scalable system that can extend to a wide range of emergency scenarios, including industrial fires, underground environments, and disaster response operations.

Future developments include multi-robot deployment, enhanced autonomy and mapping capabilities, and deeper integration with command systems through advanced interfaces and emerging technologies such as augmented reality. Together, these advancements position SALAMANDER as part of a next-generation ecosystem for intelligent emergency response.

This shift from reactive to informed response improves operational efficiency while significantly reducing risk to frontline responders, enabling safer and more coordinated firefighting efforts.

In partnership with :

Acknowledgements

We would like to express our sincere gratitude to all individuals who have mentored and supported us throughout this Capstone project. First and foremost, we extend our deepest appreciation to our Capstone mentor, Prof. Edwin Koh, our Capstone Language Advisor, Dr. Bernard, and our SCDF mentor, Mr. Hassan Kuddoos, for their consistent guidance, encouragement, and invaluable support.

The group would also like to record our heartfelt thanks to Prof. Malika Meghjani and LTA Kaushik Ramnath, whose expert advice and constructive suggestions significantly informed the direction of this project.

We further express our appreciation to Mr. Low Hsien Meng and Ms. Ker Chin Tian from HTX for their guidance, as well as to the Marina Bay Fire Station firefighters, who generously shared operational insights that shaped our understanding of real-world constraints.

Lastly, we acknowledge the contributions and dedication of our group members and the Capstone administrative team, whose efforts were instrumental in supporting the successful progress of this project.