What it does
Modular landslide cleanup machine group, set flight reconnaissance, plasma cutting, handling and grasping modules, 15-minute delivery, AI scheduling and parallel operation, zero-contact risk avoidance, adapting to multiple scenarios
Your inspiration
When focusing on emergency rescue, it is found that traditional landslide equipment is “single-function and poorly synergized”: reconnaissance equipment is difficult to break barriers, breaking tools lack positioning, and handling machinery is difficult to adapt to narrow seams. Analyze the Wenchuan earthquake, mining disaster and other cases, combined with UAV and industrial robotics technology, put forward a modular program. Drawing on the docking structure of firefighting vehicle and aerospace, we design the reconnaissance - barrier breaking - handling module, and use AI scheduling to solve the bottleneck of “single thread”
How it works
The design of the modular construction of the rescue system: flight reconnaissance module with infrared thermal imaging (640 × 512) and LiDAR, 15 minutes to complete the three-dimensional modeling (error ± 3cm); plasma cutting module 8000 ℃ arc with the vision servo, 20mm concrete barrier breaking accuracy ± 5mm; handling module 7 degrees of freedom robotic arm (500kg load), narrow slit deformation mechanism can be over 30cm width. AI scheduling is based on reinforcement learning, dynamically assigning tasks according to modeling, and 5G + Beidou realizes 10km zero-contact manipulation, increasing efficiency by 2.8 times.
Design process
From the germ of the concept to the finished product, I first used Rhino to build a modular framework: I split the rescue function into “eyes” (reconnaissance module), ‘scissors’ (cutting module), "hand " (handling module), and parametric modeling to adjust the interface dimensions to ensure that the quick-release structure fits snugly. When rendering, we added brushed metal to the aluminum alloy skeleton in C4D, used the plasma cutting head to make the 8000℃ burning glow effect, and then used the physical camera to simulate the diffuse reflection of the soot at the rescue site. In order to make module collaboration more intuitive, the AI scheduling logic is also marked with dynamic dotted lines in the rendering diagrams. For example, after the reconnaissance module is modeled, the path of the cutting module to break the barriers will automatically generate an orange guide line, and the reconstructed rescue scene profile is superimposed on the final figure to highlight the trajectory of the equipment in the debris.
How it is different
Rescue equipment on the market is mostly a “one-trick pony”: those that can fly can't be dismantled, those that can be dismantled can't be moved, and those that meet complicated landslides are stuck in their shells. My design splits the rescue function into three modules: “scouting eye”, “cutting knife” and “carrying hand”, which can be freely combined like building blocks - the flight module first shoots a map to find survivors, and the cutting module immediately burns through the survivors. - The flight module takes a picture of the map to find survivors, the cutting module immediately burns through the obstacles, and the transportation module synchronizes to clear the road. 15 minutes to the scene, the AI system, like a “commander”, allows the modules to work at the same time, and it can handle narrow seams and heavy loads, which are difficult jobs, and it is more adaptable to 50% of rescue scenarios than the traditional equipment.
Future plans
In the future, I want to add a “brain” to the device and control it with my mind; and then install “shock-absorbing legs” to make it more stable when walking on bad roads. Next find the rescue team to try, according to the feedback to change the algorithm, make the device lighter, but also let it and the community first aid point linkage, to help more people.
Share this page on