Russia Develops Archaeology Robot to Automatically Search for Artifacts Underground
Scientists at the Moscow Institute of Physics and Technology (MIPT) have developed and tested the Terra Cognita Bot robotic system to automate archaeological surveys. The robot moves independently and scans the ground using ground-penetrating radar, a metal detector, GPS, inertial sensors and computer vision.
Data collected by the robot are synchronized and processed by software algorithms, allowing the system to build a digital model of the site. The developers plan to train a neural network to automatically classify detected objects and distinguish potential archaeological finds from natural structures.
Works Reliably in Any Weather
Searching for underground artifacts with ground-penetrating radar is difficult, painstaking work. Normally, an archaeologist has to walk every meter of a site at a steady pace while recording coordinates along the way. Keeping the survey error-free becomes particularly challenging on difficult terrain, in bad weather and amid radio interference. But now, some of the more arduous work can be handed over to a machine.
The robot surveys the area autonomously using ground-penetrating radar, a metal detector, GPS navigation and computer vision systems. The collected data are processed by an onboard neural network, which immediately builds a detailed 3D map of the underground layer and marks detected anomalies. The Terra Cognita Bot system uses a 400 MHz radar antenna, allowing it to scan soil to depths of up to 10 meters.
The system has already undergone field testing at archaeological sites in Old Ryazan. “We worked at an archaeological site with traces of burned wood. These objects produce a strong contrast in ground-penetrating radar data. We believe there may also be remains of wooden structures, ovens and hearths underground,” Oleg Bulichev, a senior researcher at MIPT’s Laboratory of Cognitive Dynamic Systems, said of the system’s first deployment.
More Detailed Mapping and Classification of Finds
The archaeology robot can already be used to survey terrain, but its developers plan to further improve the system. In the future, it could detect and immediately classify objects, distinguishing, for example, human remains, household items and structural elements. The bigger challenge, however, is teaching the robot to recognize debris such as rocks, tree roots and voids, so archaeologists do not waste time investigating them. The team could also work on improving the robot’s autonomy and navigation stability over rough terrain, combining ground-penetrating radar, magnetometer and visual data, and automatically generating 3D maps of archaeological cultural layers.
Archaeologists Are Changing Their Toolkit
Ground-penetrating radar, lidar, drones and AI are already widely used by archaeologists around the world. Russia’s archaeological community is among the most successful, including in the development and use of new technologies. What makes Terra Cognita Bot different is its autonomous mobile platform, which combines several methods of detecting and processing information in a single system.
Terra Cognita Bot could let archaeologists survey large sites faster, pinpoint excavation locations more accurately and reduce unnecessary work. Experts expect the next step to be autonomous survey robots capable of independently covering large areas and delivering archaeologists a ready-to-use digital map of potentially significant objects.
The technology’s commercial potential extends well beyond archaeology. The same capabilities could be used for engineering surveys, locating underground structures and utilities, geophysical surveys and other applications that require autonomous terrain scanning.
