How Autonomous Vehicle Map Building Benefits ADAS Vehicles.

October 31, 2017 Tasking

Toy car on a world map




It’s always been interesting to me that the explorers of the past were cartographers. It makes sense, there was no point in exploring something new if you weren’t able to map it out and show it to the people back home. Even today map makers are still on the edge of the unknown, although this time the frontier is autonomous navigation. Vehicles with advanced driver assistance systems (ADAS) are becoming more numerous each day, and most of them need high definition maps in order to operate. Interestingly, they’re not only used by ADAS vehicles but can also be created and updated by connected cars. These maps won’t just show roads and bridges but will model streets down to curbs and potholes with centimeter accuracy. High definition outlines of roads will give future vehicles something to rely on when their onboard sensors aren’t working. They may even eventually reduce the need for complex multi-sensor arrays on ADAS enabled cars.


Map with pinpointers on it
Current maps are not high enough resolution to bed used in ADAS cars.




What is an HD map?

The maps needed by modern-day ADAS enabled vehicles are like the ones your old TomTom or Garmin used, only much more detailed. Where the maps that we use daily on our mobile phones are accurate within a few meters, these will need to outline the streets in terms of centimeters. They’ll also have to be updated daily or weekly since cars will be using these charts as part of the backbone of their navigation systems.


I have to admit that I occasionally take a turn too sharp and hit a curb now and then. When I’m driving I may find that acceptable, but if I’m in a self-driving or semi-autonomous vehicle bouncing off things will make me a bit nervous. That’s why new HD maps are moving from meter resolutions, down to centimeter accuracy. Users will want their cars to navigate effortlessly, and that will mean roads modeled down to every curb, pothole, and sign. In order to do this, most companies use things like LIDAR and other sensors found in ADAS vehicles to accurately map roadways.


As Bob Dylan might say, the roads they are a-changin', which is why they need to be re-assessed on a regular basis. Many companies have already taken to the streets to do just that. Some develop and run their own systems to chart out highways, while others are planning to contract out that work to people with ADAS cars or the requisite sensors. These updates will give us constantly changing maps that evolve or devolve in the case of potholes, as our roads do. They could also replace user reported events, like traffic delays, that are already integrated into navigation apps.


car wheel avoiding a pothole
New maps will help us avoid potholes like this one.




How do these maps fit into vehicles?

Autonomous vehicles have been driving the streets for years already, so why do we suddenly need these high-quality maps? Well, some of those cars have already been using those maps to find their way. Even if a car uses a large and diverse sensor array to navigate, it will still need detailed charts to use as a backup system. Beyond physical driving, these maps let companies simulate scenarios during development. This allows them to investigate potentially dangerous circumstances on a computer rather than out in the real world.


  • Primary Navigation - We all know that Google has had self-driving cars on the roads for years. What you may not know is that they did this by using highly detailed maps. The problem with using these outlines as the main means of travel is the car won’t work when there aren’t maps. That’s why those companies are working on mapping every pathway they can.

  • Backup System - One of the main challenges for ADAS vehicles is ensuring safety. A car may use a suite of sensors and their data to explore its environment, however, those sensors may not always work. LIDAR can be inhibited by rain or snow, and then your vehicle can have an accident. Highly detailed maps coupled with advanced GPS can provide a secondary system in case the main sensor array fails.

  • Simulation - It’s always better to test a system in simulations before risking a costly prototype. That’s what Waymo is doing with their HD maps. They use data gathered in the real world to simulate roads and then drive virtual cars on them for testing. This allows them to come up with a viable solution to counteract difficult conditions before experimenting with a real car.


All of these are reasons that we need detailed virtual maps, and together they become an imperative. Google started the detailed mapping, but now other companies have joined the fray in a race to see who will control the virtual landscape of the world’s roads. Once streets have been mapped ADAS vehicles can drive on them safely, either using the map as part of a primary or secondary navigation system.


While others are intent on developing HD maps of the world, you’re focused on developing programs for the cars that will use them. Just like those vehicles need a map to guide them, you could use software to help you during development. TASKING has developed a variety of products, like a standalone debugger and static analysis tool, that are made specifically for the ADAS industry.

Have more questions about HD maps? Call an expert at TASKING.

Previous Article
How an Interrupt Driven Embedded System Can Save You Power
How an Interrupt Driven Embedded System Can Save You Power

Using interrupts to handle peripherals will save you power, but take care to implement ISRs correctly.

Next Article
Advantages of Sensor Fusion With GPS in ADAS Vehicles
Advantages of Sensor Fusion With GPS in ADAS Vehicles

The best way to use GPS in a connected car could be as a supervisory sensor.

Get Your TASKING Free Trial Today.

Free Download