Tiny Swarm Robots Boost Mining Efficiency

What comes to mind when you think about the types of robots that are involved in mining? For most people, it’s probably big autonomous trucks that are used for hauling, or robotic drilling and blasting systems that clear earth out of the way. However, these big, impressive machines are not the only ones that have a place in the field. There are also tiny robots that scurry about to make mining safer and more efficient.

A particularly interesting system of this sort was just described by a trio of researchers at the University of Adelaide. Instead of focusing on massive industrial equipment, the team explored how groups of inexpensive swarm robots could cooperate to transport ore more efficiently. Their work borrows ideas from nature, specifically the foraging behaviors of leafcutter ants and honeybees, to determine whether decentralized robots can outperform more conventional approaches in simplified mining scenarios.

The researchers implemented three different strategies using a commercial robotics platform. The first, called the Baseline model, is intentionally simple. A single robot travels down a haul route until it discovers an ore deposit, returns it to the base, then heads back out to continue searching. While easy to understand and implement, this approach wastes considerable time repeatedly traveling over the same ground.

The second strategy, inspired by leafcutter ants, divides the work between two robots. One serves as a loader that continuously explores and identifies ore, while the other acts as a dedicated hauler responsible for transporting material back to the base. This tandem approach allows exploration and transportation to occur simultaneously, reducing idle time and improving overall throughput.

The third strategy takes inspiration from honeybees. Rather than stopping to collect ore immediately, a single robot first surveys the entire route and records every ore location in memory. Once exploration is complete, it calculates an efficient retrieval sequence and collects the deposits afterward. By separating exploration from collection, the robot avoids many unnecessary return trips. In the team's longest 16-meter test with eight ore blocks, this memory-based strategy reduced total travel distance by as much as 80%, cut estimated energy consumption roughly in half, and completed retrieval about 60% faster than the Baseline method.

Each robot is built around the Pololu Zumo 2040, which uses the Raspberry Pi RP2040 microcontroller. The platform combines the dual-core RP2040 with quadrature wheel encoders, reflectance sensors for line following, proximity sensors for object detection, an OLED display, RGB LEDs for status indication, and dual tracked gear motors. Powered by four AA batteries, the robots followed marked routes while identifying simulated ore blocks and coordinating their behavior through finite-state control algorithms.

Although these experiments were performed on a tabletop testbed rather than in an actual mine, they provide something many swarm robotics papers lack: hardware validation instead of simulation alone. The researchers acknowledge that large-scale testing and field trials are still required, but the work demonstrates that even modest RP2040-powered robots could make future mining operations more efficient, scalable, and resilient.