Emergency Search Robot Dog: Panoramic-Thermal Imaging

I. Project Background

    A national forest park, covering an area of 200 square kilometers, has a forest coverage rate of 95%, primarily consisting of highly flammable species like pine and fir. During the hot, dry summer months, the park is at high risk for forest fires. Its terrain features high mountains, deep valleys, and complex topography with narrow trails, making search and rescue operations for trapped tourists or rangers exceptionally challenging once a fire breaks out. The primary difficulties are:

1. Dense smoke severely limits visibility, hindering the manual location of trapped individuals.

2. Thick vegetation and rugged terrain slow down manual search teams, with operations often taking over 10 hours per square kilometer.

3. High temperatures at the fire site, along with risks of re-ignition and unstable ground, jeopardize the safety of rescue personnel. During a 2023 fire, the rescue of two trapped tourists was delayed by 6 hours due to smoke; they were eventually saved but sustained minor burns.

    To enhance the efficiency and safety of search and rescue operations during forest fires, the park administration, in partnership with the local Emergency Management Bureau, deployed four emergency search and rescue robotic dogs equipped with panoramic and thermal imaging systems in May 2024.

II. Implementation Process

1. Adaptation and Deployment (May 10 – May 25)

    The robotic dogs were specifically adapted for forest operations:

• Sensors: Equipped with a 360-degree panoramic camera and a high-definition thermal imaging camera capable of detecting human body heat through smoke at distances up to 50 meters for precise victim location.

• Communication & Guidance: Fitted with satellite communication modules, emergency lighting, and a voice broadcast system to maintain command center contact without cellular networks and to relay instructions to survivors.

• Mobility: Featured optimized limb structures and anti-slip, wear-resistant tracks, enabling them to climb 50-degree slopes, cross 1-meter obstacles, and move steadily on soft forest ground.
  A 3D search and rescue map was created by integrating park topographic data with drone aerial imagery and LiDAR scans. This map details key scenic spots, trails, and water sources. Twelve emergency search routes were pre-programmed, and a standardized operational protocol—Thermal Detection → Panoramic Verification → Location Reporting—was established. The system was fully integrated with the emergency command center and aerial drone networks.

2. Simulation Drills and Optimization (May 26 – June 15)

    Four comprehensive forest fire rescue drills were conducted in the park, simulating conditions of heavy smoke, high heat, and difficult terrain. Key system improvements during this phase included:

• Enhanced Thermal Imaging: Algorithm upgrades improved the thermal camera's ability to filter out flame interference in smoky environments, raising human detection accuracy from 88% to 99%.

• Extended Endurance: High-capacity lithium batteries were installed, increasing operational duration from 4 hours to 8 hours to support prolonged missions.
  Drill results confirmed the robotic dogs reduced the average search time per square kilometer in a forest fire scenario from 10 hours (manual) to 2.5 hours, achieving a victim location accuracy rate of 99.5%.

3. Operational Deployment (June 16 – Present)

    The robotic dog units are now permanently stationed at key park security posts. Patrols are intensified during the high-fire-risk summer season (June-September). Upon receiving a fire alert or distress call, the units are dispatched immediately. They conduct searches along pre-defined or command-center-directed routes. Using thermal imaging to locate survivors and panoramic cameras to assess the environment, they transmit real-time data to the command center. Once victims are found, their exact coordinates are marked. The units then activate emergency lights and broadcast voice instructions to guide victims to safety while simultaneously directing human rescue teams to the precise location.

III. Application Results

• Dramatically Improved Search Efficiency: The average search time per square kilometer during forest fires was reduced from 10 hours to 2.5 hours, a 75% increase in efficiency. During the 2024 summer fire season, the system successfully assisted in the rescue of 3 trapped individuals, with the average operation time shortened to 1.5 hours.

• Enhanced Operational Safety: By deploying robotic dogs into high-temperature, smoky, and high re-ignition risk zones instead of personnel, the risk to human rescuers was eliminated. The units can operate continuously in darkness and adverse weather, overcoming limitations of manual search teams.

• Superior Rescue Coordination: Real-time transmission of panoramic and thermal data provides the command center with an accurate, comprehensive view of the fireground and victim status. This enables effective coordination with drone teams and firefighters, raising the overall mission success rate from 85% to 98%.

IV. Typical Incident

    On August 12, 2024, a lightning strike ignited a forest fire in the park, trapping two hikers. The robotic dog team arrived on scene at 14:20. Using live drone footage, the command center designated search sectors, dispatching the four units accordingly.
    At 14:50, one unit detected human thermal signatures through dense smoke. Its panoramic camera confirmed two individuals sheltering in a small clearing surrounded by burning debris. The dog instantly geotagged the location, sent the data to command, activated its lights, and used its broadcast system to instruct the victims to cover their faces with wet cloths and move toward a safer direction.
    Guided by this precise location data, the command center directed a fire crew to clear a path through the burning vegetation. The rescue team reached the victims at 15:30 and successfully evacuated them. The entire operation concluded in 1 hour and 10 minutes with no rescuer injuries, representing an 80% improvement in efficiency over a comparable 2023 rescue.