One of the greatest advantages that robotics and automation offer the mining industry, is the ability to remove people from potentially very dangerous environments.
Robots can enhance our capabilities in many areas, and data collection is the perfect example of a task that machines can perform faster, more safely and more thoroughly than humans.
CL: Tell us about Emesent’s Hovermap technology: how was it developed and how can it be applied at mine sites?
SH: Hovermap’s underlying technology builds on more than ten years of R&D at CSIRO’s robotics group, bringing together LiDAR-based mapping and drone autonomy into a single system.
Having seen the pull from industry for early Hovermap prototypes, we formed Emesent to commercialise and continue developing the IP, turning Hovermap into a commercial product that is solving challenging, real-world problems.
Hovermap is a two-in-one system which provides both LiDAR mapping and drone autonomy. When mounted to a drone it allows the drone to self-navigate underground, flying into challenging inaccessible areas. At the same time, Hovermap collects rich LiDAR mapping data of these areas.
Hovermap can also be removed from the drone and used as a mobile mapping device. It can be hand carried, mounted on a vehicle, attached to a tether to be lowered down vertical shafts etc.
Underground mines contain large inaccessible areas, such as stopes. These unsupported underground voids are a source of ground falls that can endanger personnel, underground infrastructure and equipment.
Mining engineers and surveyors are required to regularly inspect these areas. Traditionally, they use a cavity monitoring system (CMS) scanner deployed from a boom, which is time-consuming, produces poor quality, incomplete scans, and endangers personnel.
Hovermap takes away all of those issues. Rather than take three hours to scan a stope with a CMS, it can be achieved in 15 minutes with Hovermap. This is a signiﬁcant time saving and allows for the stope to resume production quickly. Using Hovermap also ensures personnel keep a safe distance from the edge of the stope and means they are in unsupported ground for far less time.
Hovermap produces ‘shadowless’ point clouds, which means scan data covers the entire stope and has a consistent spatial resolution. Mining engineers can identify geological features or conduct volumetric analyses, which is not possible with low point density CMS data. Our customers are using Hovermap to map any underground area, draw points, ore passes, and old underground workings.
I understand you’re about to launch a level 2 automation feature for drones, what does this involve?
Autonomy Level 2 (AL2) allows the drone to fly autonomously beyond line-of-sight (BVLOS) and beyond communication range. A map is generated and streamed back to the operator in real time, displayed on a tablet. The drone is operated through the tablet interface by placing goal waypoints in this map. The drone then self-navigates to these waypoints while avoiding collisions.
This allows new areas to be accessed and mapped that were previously beyond reach. Old workings, ore passes, vent rises, larger, more complex stopes are now all within reach.
AL2 make the operation extremely simple, enabling almost anyone to operate the drone safely in these challenging underground environments.
Does the system depend upon artificial intelligence to navigate the environment and unexpected obstacles?
Hovermap employs the well-established robotics concept of plan-sense-act. It perceives the world around it using the LiDAR data, decides what the best action is to take given the goal it is trying to achieve and then executes that plan. This happens over and over many times per second.
This may sound simple but there are many thousands of complex calculations being performed every second onboard the device.
One of the key challenges to overcome is to estimate how it is moving, without using an external signal such as GPS. This problem is solved using a technique called simultaneous localisation and mapping (SLAM). It uses the LiDAR data in real-time to estimate its own motion. This same technique also produces the detailed point clouds from the LiDAR data.
The strict definition of AI requires a system to learn while it is operating and adapt its behaviour based on the learning. Hovermap does not employ this type of on-the-fly learning, but we are using machine learning to analyse the data collected and detect features of interest in the data.
What are the next steps for Emesent’s technology?
There’s ongoing development of the technology, especially in-flight autonomy, and multiple product enhancements will be released over the next year. We’re improving our data analytics capability, not just collecting the data but actually automating the analysis because we can add a lot of value for our customers.
How do you see drones impacting data collection at mines of the future?
The digital mine and analytics are increasingly important areas for improving mine productivity and eﬃciency, and drones will play an increasingly significant role in the creation of digital twins. The use of drones above ground is already well established in the mining space and they have proved their value there.
We’re pioneering a new frontier of underground drones and Hovermap is a key enabler in this space. It enables collecting rich new data that was previously not possible.
I’m excited to see how this will start to change and improve some of the decade-old mining methods. If a mining engineer knows that he or she can confidently capture data deep inside stopes or draw points that was not previously possible, will this open up new design options and mining techniques?
I think we’re just seeing the tip of the iceberg in terms of the potential value that underground drone-based scanning can provide. There are certainly exciting times ahead.