Andrew Chatham, Google’s principal software engineer for the Self-Driving Car Project, spoke to the Institute of Transportation Engineers 2014 Technical Conference this winter. I would have loved to have heard the talk, but wasn’t able to make it out to California.
Michelle Bridsall has an interesting article in the latest ITE Journal about ITE working with Google.
Chatham explained that the self-driving car works from a combination of pre-programmed maps, gps, and sensors. It’s using the maps to drive along the route. Then the system is using a stream of real time data from on board sensors (LIDAR, video, and radar) to look for anomalies the car has to take into account (i.e. slow down for the car in front of it).
The takeway from the article for me is how our industry can assist the automated car industry. Here are some things Chatham believes we can do for them:
- Keep building curbs, medians, and striping. This infrastructure makes it easier for the car to navigate (they make it easier for us humans too).
- Provide a data stream of up to date construction project information. Advanced knowledge of works zones (better yet the layout of the construction zone and the time-frame it will be implemented) makes it easier for the computers to adjust to the situation.
- Provide data on when construction projects will be finished and what the new infrastructure will be. This will allow them to have current maps, which are the foundation for their system.
- Provide signal timing plans, especially recently updated signal timing plans.
- Provide data on upcoming street closures (for construction or events – fairs, block parties, running races, etc.).
- Provide data in latitude/longitude instead of mile markers (aside, we don’t use slide rules anymore – you’d think we could do away with mile markers).
In a nutshell, any data we can provide them in advance is better for the computer systems than forcing the on board system to adjust to changed conditions in real time.