The Real Future of Autonomous Vehicles?

On our visit to UMTRI last week, many in our class were introduced for the first time to the reality of the autonomous car becoming commonplace in coming years.  UMTRI has been playing around for a few years now with dedicated short-range communications to enable V2V communication.  However, the UMTRI is the leader in autonomous car development.

In Silicon Valley, CA, Google already has technology capable of safely driving a car through city streets.  Google’s autonomous cars use not only DSRC radio for V2V communication, but lasers, cameras, machine learning computers with sharp reaction times.  Google prototypes have driven over 100,000 miles so far on public roads, with two accidents total (both fender benders).  The author of the cited Business Insider article notes the relevance of automated vehicles in a society where the majority of cars on roads today have one solo driver and average 30 mph.

google-self-driving-car-11

 

Although still in the testing phase, project director Chris Urmson intends the automated car technology to be available to the consumer in the next ten years.  As opposed to UMTRI, the members of which seemed confident that DSRC alone could be first mechanism used to enable automated driving, Google incorporates a number of technologies to make autonomous cars a reality.

http://www.businessinsider.com/google-self-driving-cars-2014-5

3 comments to The Real Future of Autonomous Vehicles?

  • Louisa Ortiz

    That was my main question at UMTRI. The V2V and V2I(?) DSRC radio seemed like a great and cheaper alternative to sensors but what about things that do not have these radio waves? UMTRI said that they could be added to cellphones so the vehicles could sense people but that is not a full answer. People cannot be relied on to have a phone on them at all times and what about children or animals? I feel that a combination of both sensors and radio waves is the only answer.

  • Louis Ike

    I found it very difficult to believe that sensor technology wouldn’t be capable of mass production and working effectively without V2V technologies as UMTRI made it seem. When we visited with other people involved in the auto industry in Detroit the consensus seemed to be in the other direction, and I must say that I think sensors may be the way. If the right economies of scale can be reached, then it may be possible to equip even lower end vehicles with the necessary technology to make them autonomous.

  • Sensors – well, you’re talking about multiple radar – lidar systems in each vehicle. Add in testing and installation, wiring, control units, vehicle-specific tuning of software signal interpretation. Mass production can only lower prices so far, and hardware isn’t the only cost component. Plus the rollout will take decades. V2V is quick and cheap, yes with its own limitations but on a cost-benefit basis it seems attractive.