Out of nowhere: track underground miners in bad connection areas

Many mines around the world use hand tracking to monitor the position of miners underground, the task often performed by a mine manager who, at the start of each shift, must provide the dispatcher with a list of people’s names and their locations in the mine.

Once inside, if a miner needs to go to another work area, they communicate this to the dispatcher via a rotary phone in the mine. The dispatcher then updates the list of current miner locations.

But as with many manual processes, this type of tracking has several drawbacks, the US Department of Health and Human Services warns. For example, a miner’s location may be shown as being in too large a work section to determine their exact position. Or a worker might forget to notify the dispatcher when moving to another work location.

For this reason, electronic tracking technologies such as player-based tracking, node-based trackingand inertial tracking are being developed to overcome the limitations of currently available manual tracking.

Even with the existing methods, such as GPS and Wi-Fi tracking, in places such as open pit underground mines or disaster areas with destroyed existing infrastructure, the connection often faces interruptions and outages.

In several cases, the position of the mines did not allow a secure Wi-Fi connection. For cases like these, researchers from the Institute of Intelligent Systems and Artificial Intelligence at Nazarbayev University have rolled up their sleeves to come up with an algorithm that could improve the safety of underground workers, by determining the location minors only based on existing Wi-Fi hotspots. .

Algorithm development

The researchers started with indoor location, using available Wi-Fi signals to determine a person’s location, before moving on to deep learning stages, where they form a network to predict someone’s location. one in an underground mine. By simulating tests in a building at Nazarbayev University, they succeeded in developing a neural network that predicts the exact positioning of a person inside the university.

“We collected a three-story dataset and trained a model with an error distance of about 2.5m,” said Mukhamet Nurpeissov, data analyst at the Institute of Smart Systems and Intelligence. artificial from Nazarbayev University.

After reaching this error distance, the team decided to improve the model by adding information from inertial measurement unit (IMU) sensors, which, after further data collection, increased the range of the movement when locating a person or a phone.

Right now, the team is collecting the data needed to start building the machine learning algorithm that will receive the information in the form of signals from available Wi-Fi spots, as well as IMU sensors.

“One of the reasons we added the IMU sensors is that when a device is not receiving any Wi-Fi signal, it can locate itself by calculating the displacement via the IMU signals with an overall position estimate,” explains Nurpeissov.

These sensors can also be found in pedometers and smartphones and are used to determine a person’s acceleration and from the Earth they can also successfully determine a person’s location. Using the IMU means that in areas without a stable Wi-Fi connection, a signal can still be found.

“We pitched this idea in front of a company, who said they had a small existing infrastructure with underground Wi-Fi hotspots, so we took our job from there,” he says.

Adapt and evolve

The development of the technology is not without obstacles as the team says the main challenge for them is collecting data in order to train the model. And one of the trickiest topics is recording the exact position when receiving signals.

“In the university building, we solve it using landmark markers, where a camera sensor can calculate its relative position to the marker. So we put these markers everywhere and we know the exact position of these markers, relative to the base point to calculate our position,” says Nourpeissov.

He also notes that the different conditions in the underground mines could create unpredictable situations, which the team tries to anticipate. For example, there are many hotspots inside buildings to help them locate each other better, while mines have limited Wi-Fi hotspots whose full performance is yet to be achieved. tested.

As part of the trials, for now, the workers move in predefined pairs, after which the positions of this walking path are extracted from the map. Once they have obtained a complete dataset, the team will proceed to train the models.

In order to meet future demand, the researchers ensure that 5G networks can also be used, instead of Wi-Fi hotspots, as the base stations of 5G networks have a much higher density.

Instrumental to improve safety where standard GPS systems fail due to occlusions and signal blockages in buildings, the development of this algorithm gives another hope for improved surveillance, which could replace manual tracking and help miners feel safer underground.