Mountain-Wading Off-Road Robot Dog: 1.2m Submersion

I. Project Background

    A mountainous region located within a tributary basin of the Yangtze River is highly susceptible to heavy summer rainfall, frequently leading to flash floods and landslides. The area is characterized by dispersed villages, many of which are accessible only via fragile roads built along mountainsides and rivers. These roads are easily destroyed or submerged during disasters. Previous flood rescue operations relied on a "manual search and inflatable boat" model, which suffered from major limitations:

1. Floodwaters would destroy roads and bridges, blocking rapid access to affected villages.

2. Strong currents and water depths exceeding 1 meter in some areas hindered boat operations and posed significant capsizing risks.

3. The challenging post-disaster terrain—filled with mud, trenches, and fallen trees—severely hampered the movement of rescue personnel, leading to low efficiency. During a 2023 flood, the initial assessment of a single isolated village took 5 hours.

To improve emergency flood response capabilities in such environments, the local Emergency Management Bureau deployed five all-terrain amphibious robotic dogs in May 2024.

II. Implementation Process

1. Adaptation and Deployment (May 8 – May 23)

    The robotic dogs were specially modified for operations in flooded, mountainous terrain:

• Waterproofing: Designed with full-seal, IP68-rated waterproofing, allowing them to traverse waters up to 1.2 meters deep for a minimum of 30 minutes.

• Mobility: Equipped with high-traction tracks and retractable limbs. The tracks feature anti-slip rubber, and the limbs can adjust for variable ground clearance. This enables climbing 45° slopes, crossing 1.5-meter trenches, and stable movement on muddy surfaces.

• Sensors & Communication: Fitted with waterproof high-definition cameras, infrared thermal imagers capable of detecting human heat signatures through water, emergency lighting, and a voice broadcast system. Integrated satellite communication modules ensure connectivity in areas with no cellular network.

• Mapping & Planning: Using regional topographic maps and drone survey data, a detailed 3D rescue map was created, pinpointing villages, roads, and rivers. Ten pre-defined emergency response routes were programmed, and a standard operating procedure—"Aquatic Reconnaissance, Overland Traversal, Victim Localization"—was established. Full integration with the emergency command center was completed.

2. Simulation Drills and Optimization (May 24 – June 12)

    Eight simulated flood rescue drills were conducted in mountain rivers and abandoned villages, replicating conditions with 1.2-meter water depth, 2 m/s currents, destroyed roads, and muddy ground. Two core capabilities were enhanced:

• Stability in Currents: System refinements improved stability in flowing water up to 2 m/s by optimizing the center-of-gravity adjustment and track drive power.

• Underwater Detection: Algorithm improvements for the thermal imagers filtered out water interference, increasing the accuracy of human detection from 85% to 98%.
  Testing verified that in these complex environments, the robotic dogs reduced the search and assessment time for a single village from 2 hours (manual) to 30 minutes.

3. Operational Deployment (June 13 – Present)

    The five amphibious robotic dogs are now stationed at emergency posts throughout the region, with readiness heightened during the flood season (June-September). Upon a flood alert, they are immediately deployed along pre-set or command-center-directed routes.
    They reach cut-off villages by fording through water and navigating overland obstacles to locate trapped individuals. Once victims are found, their precise GPS coordinates are marked, and live video is streamed to the command center. The units use voice broadcasts to guide survivors and simultaneously direct human rescue teams to the exact location.

III. Application Results

• Expanded Rescue Coverage: The system addressed previous blind spots in areas with water up to 1.2 meters deep and where roads were destroyed, eliminating such coverage gaps entirely. During the 2024 flood season, it successfully conducted operations in three isolated villages, locating 23 trapped individuals.

• Significantly Improved Rescue Efficiency: The initial assessment time for a single village was slashed from 2 hours to 30 minutes, a 75% efficiency gain. The average time to rescue trapped individuals dropped from 5 hours to 1.5 hours, greatly improving survival odds.

• Enhanced Safety and Cost Savings: By using robotic dogs for initial reconnaissance in high-risk zones like fast currents and unstable slopes, the risk to human rescue teams was eliminated. The need for three manual rescue teams was removed, saving approximately 960,000 RMB annually in labor costs.

IV. Typical Scenario

    On August 10, 2024, torrential rains triggered flash floods, destroying road access to three mountain villages and submerging parts of them under 1 meter of water, cutting off all communication.
The amphibious robotic dog team arrived at 10:30 AM. The command center dispatched the five units to the three villages.
    By 11:10 AM, one unit reached a target village, fording through a 1.2-meter-deep submerged roadway. Using its thermal imager, it detected three elderly individuals trapped in the village clinic, unable to evacuate.
    The robotic dog immediately marked their location and uploaded live footage showing them stranded on the second floor with water rising. It used its broadcast system to calm the victims and instruct them to await rescue.
    Guided by this precise location data, the command center dispatched inflatable boats. The robotic dog remained on site to monitor water levels continuously. By 12:20 PM, the rescue team arrived and successfully evacuated all three individuals.
    In this operation, the robotic dog's ability to penetrate the flood barrier and rapidly pinpoint the victims provided a critical time advantage, preventing potential casualties from the rising waters.