Multi-Sensor Cooperative Robots for Shallow Buried Explosive Threat Detection: Radar Sensors and Optical Sensors Integrated by System Software

This project has fully demonstrated a new paradigm using a team of cooperating robots carrying different sensors, with data fusion for detection and discrimination of surficial and buried landmines. The system is operated remotely with negligible risk to the operator. The integration of information collected by three robots has been achieved by software architecture. With this architecture, it is possible to plan the mission path, optimize the sensor settings and visualize the data from a remote terminal (e.g., a tablet or smartphone). The sensors used for the detection of plastic and metal landmines are installed in a configuration of three robots: microwave radar (Robot #1 with a UWB radar and Robot #3 with HSR radar), LiDAR (Robot #1 and Robot #3), high-resolution camera (Robot #1), and metal detector (Robot #2). The acquisition and processing of data from multiple radar sensors displayed improved detection and classification for small plastic-cased low-metal-content landmines (e.g., M-14 and Type 72) using high resolution microwave radar. These are generally difficult to detect with metal detectors alone. In real scenarios like the ongoing war in Ukraine, surface landmines like the PFM-1 (“butterfly”), are often scattered on the surface, and we demonstrated how artificial intelligence can be applied to detect this type of threat. Finally, because of booby traps rigged to protect minefields, demining also requires the detection of tripwires. For this purpose, the first robot in the team can detect trip wires of different types (metal, fiber or plastic) in real time. The robotic platforms are equipped with GNSS providing mapping of targets with accuracy better than $\text{1 0 ~ c m}$.