What is SLAM?
At GeoSLAM, we pride ourselves in being the experts in 3D SLAM technology
(it's in our name, after all).
But what exactly is SLAM? How does it work? Where did it come from? And why is it so important?
First things first, SLAM stands for Simultaneous Localization and Mapping. SLAM can trace its early development back to the robotics industry in the 1980s and 1990s. Robotics engineers had a problem – how to get their autonomous vehicles to move around the factory floor without bumping into walls, people, or indeed each other. Being indoors, they couldn’t rely on GPS for navigation. The engineers realised that if they could design an algorithm (like a set of digital instructions) that could concurrently map and navigate through that space, they could solve the problem. SLAM was born.
SLAM devices take data from sensors to build a picture of the environment around them and where they are positioned within that environment. These sensors may use visual data (such as camera imagery), or non-visible data sources (such as Sonar, Radar, or Lidar) and basic positional data, using an inertial measurement unit (IMU for short). The device utilises this information to compute a ‘best estimate’ of where it is within the environment. By moving its position within the environment, all environmental features (i.e. walls, floors, pillars) will move in relation to the device and the SLAM algorithm can improve its estimate with the new positional information. SLAM is an iterative process – the more iterations the device takes, the more accurately it can position itself within that space.
There are many different types of SLAM algorithms. There is visual SLAM, which uses visual data as its main source of information. There is 2D SLAM, which is only interested in movement in 2 axes.
Then there is 3D SLAM. In 2008, the Commonwealth Scientific and Industrial Research Organisation (CSIRO) developed a powerful and robust SLAM algorithm primarily focused on accurate 3D measurement and mapping of the environment, rather than autonomous navigation. You may not be familiar with CSIRO, but you’ll certainly be familiar with their work – they invented Wi-Fi, 30-day contact lenses, plastic banknotes and the list goes on. They’re also Australia’s national science agency, pure experts in their field.
In 2012, Data61, the digital innovation arm of CSIRO teamed up with UK geospatial market-leaders 3D Laser Mapping to commercialise their new SLAM. The result was GeoSLAM.
What makes our SLAM better?
We’ve already introduced different types of SLAM, but SLAM algorithms and approaches to SLAM, greatly differ. The SLAM technology used inside GeoSLAM’s product portfolio was developed by some of the smartest people on our planet. Our original SLAM (seen in our earliest solutions) was most comfortable in indoor, feature-rich environments. But we weren’t content with limiting ourselves to the indoor world. Over the course of the last six years, and over thousands of projects, our SLAM algorithm has been refined, honed and tweaked. We’ve pushed our algorithm to breaking point – and sometimes beyond – to see what makes it tick. We’ve taken our technology into notoriously difficult scanning environments, such as repetitive corridors, smooth-sided tunnels and spaces with moving people, and really put it to the test. The result is a constantly improving SLAM algorithm, one that is so robust that it now works equally well in outdoor open environments as it does indoors. It turns out, not all SLAM is created equally.
So how does ‘GeoSLAM’s SLAM’ work?
All of GeoSLAM’s mapping technology, from 2013’s ZEB-1 (the world’s first SLAM-based mobile mapping system) through to the new ZEB-HORIZON, feature our improved SLAM algorithm at their core. Our products use Lidar (Light Detection and Ranging) sensors, which emit multiple pulses of laser light per second. Depending upon the product solution, these range from 43,000 to an astonishing 300,000 pulses per second. These pulses (invisible to the human eye) reflect off surrounding surfaces and back to the sensor. By calculating the time taken for each pulse to return, the distance from object to sensor is known.
We also utilise an industry-grade inertial measurement unit. The IMU is used to estimate an initial position within the space to be mapped. Our solutions feed information from the IMU and the Lidar sensor into the SLAM algorithm, to generate a trajectory (or path) through the space. By knowing the distance from each point on the path to the surrounding features, the device builds a three-dimensional point cloud of the space.
The device then moves forward, and the entire process is repeated. In fact, this process is being repeated every single second, throughout the duration of the scan. With each new iteration, the device improves its understanding of its surroundings and its position within those surroundings. So as the user moves through the environment, the point cloud is built, in real time around them.
To further optimise the scan quality, the user is always encouraged to return to the starting point of the scan, referred to as ‘closing the loop’. By returning to this initial position, the user is giving the SLAM algorithm the best chance to tie all the environment’s data together, thus reducing the risk of data ‘drifting’ or ‘slipping’, which can occur whilst moving. Help the SLAM and the SLAM will help you.
ZEB-REVO RT system being used to map a construction project. The RT in the name stands for "real-time" and the live data can be viewed while capturing to make sure there are no gaps or occlusions.
What are the benefits of using SLAM?
SLAM was designed for devices that are moving through a space. SLAM based systems are inherently mobile, they are at their best when used on the move. For anyone involved in the acquisition of geospatial data, this is a huge game changer. Gone are the days of multiple, static set-ups of bulky, tripod-based systems. With a SLAM mobile mapping system, it’s possible to rapidly and simply walk through an environment, building a digital map as you go. By removing laborious set-ups from the equation, the time and cost savings are enormous. SLAM based mobile mapping systems can be over 10 times faster at acquiring data.
GPS doesn’t work indoors, it requires a line of sight to at least three satellites to function. But it isn’t just indoors that is out of bounds to GPS based systems. Forests prove difficult, as tree canopies block the line of sight to the sky. Similarly, urban canyons and tall buildings block signals in built up environments too.
SLAM based technology overcomes all these obstacles, by cutting out GPS altogether. With no reliance on remote data, SLAM systems are truly go-anywhere technology. This opens a whole new realm of environments to be surveyed, that were previously highly problematic, time consuming, or both. For the built environment, this opens large opportunities;
- Fast, weekly progress monitoring of construction sites
- Real-time surveys of residential, commercial and industrial facilities
- Comprehensive site surveys of existing structures to be refurbished, remodelled or extended.
It’s easy to see how SLAM mapping devices are considered a ‘disruptive technology’ in the survey industry.
But it’s not just the built environment that benefits. The pre-cursor to the ZEB-1 found its birth in a complex cave system, in south-east Australia. From these humble beginnings, GeoSLAM products have been utilised in caves, mines, forests and open fields, globally. To cement our market position, we are proud to have built an international dealer network of almost 80 organisations, in over 50 countries, across all six continents.
That’s why we are called the experts in go-anywhere mapping 3D mobile mapping technology.
What’s Next for Us?
You might think that after six years of rapid growth and development we’d be taking a break. But in this fast-paced, innovation hungry industry, slowing down is not an option. After all, our SLAM isn’t standing still, it’s constantly evolving and so are we. This is evidenced by the recent launch of our new, in-house innovation hub, GeoSLAM Labs, headed by GeoSLAM & 3D Laser Mapping Founder Graham Hunter.
The initiative behind the foundation of Labs is to cement innovation at the heart of all that GeoSLAM does, allowing us to expand our R&D capabilities, to keep delivering industry leading products and to fortify our place at the cutting edge of global SLAM development. The last six years have been quite the journey for GeoSLAM and with these exciting new developments, the next six years promise not to disappoint.