Navigating without GPS

In places we are not familiar with, we rely on GPS to determine our location. But what about situations where GPS is unavailable? Within the SENAV project, we are exploring innovative methods of localization and map creation in areas where GPS systems are not accessible.

GPS provides navigation

The Global Positioning System, known as GPS, has revolutionised the world of transport. Today, relying on a paper map laid out on the steering wheel or trusting our own intuition for navigation is nearly unimaginable. We have technology that can accurately determine our current location based on the positions of satellites orbiting the Earth.

GPS can measure our distance from these satellites very accurately and then use triangulation to calculate our exact position. This technology made navigation as we know it in modern cars and phones possible. Navigation refers to determination, at a given time, of the vehicle's location and velocity as well as its attitude.

💡 GPS is just one of the existing global navigation satellite systems (GNSS). This American system was the first, and therefore people do not remember other similar systems very well. The first GPS satellite was sent into orbit on 22 February 1978. Next, in time succession, came the Russian GLONAS system (first satellite launch on 12 October 1982), the Chinese BEIDOU system (first satellite launch on 31 October 2000) and, finally, our European Galileo system (first satellite launch on 21 October 2011). Modern navigation chips can receive signals from all navigation systems, which increases the accuracy of positioning.

GPS works as long as you are in direct line of sight with at least four satellites, which is why navigation can sometimes drop out in the forest. Buildings are a significant challenge for GPS. Roofs and walls can obstruct GPS usage because they block our view of the sky, where the satellites are located.

Navigation application in your phone or car does not provide only navigation, but as well guidance. Guidance refers to the determination of the desired path of travel (the "trajectory") from the vehicle's current location to a designated target, as well as desired changes in velocity, rotation and acceleration for following that path. Frankly we should call navigation systems as Navigation and guidance systems.

Autopilot - autonomous driving

Modern autonomous driving concepts control the movement of a vehicle based on the knowledge of the exact location and where the vehicle is supposed to go. Positioning for the autopilot is in the vast majority of cases provided by the GPS system. For autopilot to function, it requires one more crucial element - an accurate map of the environment it needs to navigate. This map is integrated into the autopilot system itself and can be updated as needed from a remote server via the internet.

Operation in unknown terrain without GPS support

What about situations where we do not have access to GPS or any maps? Can the vehicle still operate autonomously in such cases? The answer to this question is provided by modern methods of Simultaneous Localisation and Mapping known as SLAM, which manage localization and map creation using various sensors. In this case, sensors such as radar, laser, or cameras can be used. SLAM evaluates real time inputs from available sensors and creates its own three-dimensional maps "on the fly". The sensors allow SLAM to map the terrain very accurately and to record any larger objects that may present potential obstacles to the moving vehicle and need to be avoided.

The SENAV project

The SENAV project - Smart Space Exploration Navigation, on which we collaborate on, aims to enable breakthroughs in technologies and scientific instrumentation for space science and exploration missions. SENAV focuses on optical navigation for orbiters, landers, drones and robots. M2M Solutions' main task will be to implement multiple SLAM methods and to integrate a versatile autopilot capable of controlling orbiters, landers, drones, rovers and other robots. The work of our consortium consists of improving intelligent algorithms, optimizing software solutions and miniaturizing hardware modules, validated through test in a laboratory environment.

The payload processor used to run SLAM is the core unit of the navigation unit that we develop in the project. To evaluate our core technologies, we will use hardware available in common stores, steering clear of exorbitant costs and lengthy delivery times. In addition, this will allow us to reuse developed technologies and products also in terrestrial applications, e.g. storage management, in-door localisation.

To put things in perspective - the processor employed in NASA's latest Mars-bound Perseverance rover carries a staggering price tag of $200,000. This processor, the PowerPC 750, dates to the 1998 G3 iMac and operates as a single-core processor at 233 MHz—a speed that pales in comparison to today's smartphones, which are at least ten times faster.

However, this price tag has some reasons - mainly its adaptability to harsh space environmental conditions, boasting resilience to temperatures ranging from -55 to 125 degrees Celsius and robust resistance to cosmic radiation, in which the modern smartphone would survive just for a very short time.

Currently in the SENAV project we are testing the behaviour of SLAM algorithms in a simulator. The simulator contains the surface of Mars, which was created based on existing and available photographs of the Martian terrain. A drone connected to the simulator flies in a virtual world over this simulated Martian surface and produces a map as it flies over it. In this way, we are testing the accuracy of the position estimation. The accuracy depends on the accuracy of the connected sensors and on their individual measurement types, on the SLAM algorithm itself as well as on the environment in which it flies. Our sensor suite comprises an IMU for change measurement, a laser for range measurement and a camera. Iterative development and testing is important, as in the beginning, after the GPS signal was disconnected, we couldn't even maintain a stable position (Hovering) with the drone in the simulator. Now, we have developed a stable operating system.

The photos are maps made by drone.

It's great to simulate things first, as they allow you to run a few dozen flights in a day. A crashed drone is no tragedy either, because after restarting the simulation it works like new again. The most challenging part is simulating the image that we feed into the drone's "mind" in the virtual matrix, making it genuinely believe it's flying on Mars. In doing so, we mentally transport ourselves there alongside it; some of our colleagues even spend several hours there daily, in thought. Let us tell you, it is a great feeling to be exploring where no one has been before.

We can then utilize the localization technology on Earth, in warehouses where, similar to space, GPS signals are unavailable.

Finally, watch the real simulator video showing the creation of a map as a drone explores Mars' terrain. This captivating footage, showcased also at Drontex 2023 and featured on Spectrum 24 program of TV JOJ 24, offers a glimpse into the exciting future of space exploration that we are a part of.

A method of generating a map when a drone flies over simulated Martian terrain.