It is common knowledge, that Autonomous vehicles, (or self-driving cars as they are commonly called) currently lack the common sense needed to navigate using a traditional human map as they are unable to decipher or interpret content. They consequently have to rely on complex map signals that are unaffected by the many variables affecting it such as signalling issues, updates and precision.
Maps for CAV: It is hence quite clear that a new class of machine maps have to be designed and created with self driving vehicles in mind as they will allow safe and predictable vehicle autonomy. Nonetheless, such a task can be quite daunting and complex as collecting data and making use of it for humans is substantially easier than what it is for machines.
For example, your usual conventional maps for humans similar to the one on your iPhone or Android, rely on GPS and simple receivers that are accurate to a few meters. This does not however mean they are pin point accurate as they rely on the fact that as a human we are able to make the “decisions” on the ground.
To compensate for their inability to operate safely and reliably from context alone, AVs need more of the scenes they encounter to be pre-mapped. No amount of sensor data can substitute: Maps tell an AV there’s a traffic light coming, for instance, even when a large truck blocks the view. Up-to-date maps also let AVs reroute to avoid tough situations like unprotected left turns or intersections under construction.
It is crucial that these machine maps hence meet a variety of complex abut highly important criteria and demands such as:
- Incredible precision, to ensure the car can compensate for its lack of understanding context and know where it is within small distances.
- Granular instructions, for example, like which lane the car is in, or the traffic rules that apply to that lane to ensure a safe trip.
- Constant connection, which continues to provide information even when GPS signals are weak or missing.
The catch: Collecting accurate 3D data of cities and keeping the information on them up-to-date have both historically been incredibly expensive and time-consuming. Even a thorough one-time map is almost useless for autonomy because cities are constantly evolving.
The challenge is making the hardware footprint scalable and lighter, whilst still delivering sufficient enough accuracy and precision. The fact that autonomous vehicles and all their components are being released in a very slow and safety-aware fashion, allows for companies to improve their technology and be able to minimise hardware size whilst increasing its effectiveness.
Article by Nicholas Kalavas for Y-Mobility
What are your thoughts on the above? Any comments or suggestions of something we should add to our Blog?
Think we can help you? Send us an email at info@y-mobility.co.uk, or follow us on Twitter and LinkedIn where we release educational and informative content every day.
