Thorium 5G The New Oil Internet knowledge Driverless cars Astronomy Photography Contact

[ back ]

Driverless vehicles

In 2019 most of us with modern vehicles have some sort of inbuilt intelligence/driver-assist technology in our cars, such as cruise control, parking sensors, collision alert, etc.  We are aware of ongoing research and development into driverless cars but we don't have them ourselves yet.  From my experience if you talk to someone who has not researched or thought deeply about the subject, but only been exposed to media coverage, they tend to be apprehensive or at least have some reservations on the subject.  When the media splash headlines when one woman was killed in an accident with a driverless vehicle (March 2018), but they ignore the thousands of deaths on the roads every day (one person killed every 25 seconds in 2010 - wikipedia), this is not surprising.

This BBC news item (1 February 2018) announced a plan by the government to have driverless cars by 2021.  As part of that, a driverless car will take a 300 mile journey in December 2019 using twisty lanes and fast roundabouts.  The initial reaction to this on first learning of it, by many people, is one of shock or concern.  I had the same reaction.  However, to put things in perspective consider the hypothetical case of meeting a car going the other way on a narrow twisty lane.  Would you rather that car was driverless or driven by a human arguing with someone on a mobile phone?

There should be public debate on this, as there are ethical problems involved and our normal method of dealing with new technologies is to wait and see what happens, and then try and fix any problems that arise.  It is better than doing nothing, and accident investigation has had and always will have an important role in dealing with unforeseen problems.  However we should not just rely on after the event methods but try and be proactive.  These are three videos that provide food for thought and a news item on the subject.

  1. ethical considerations
  2. the Trolley problem
  3. transport networks
  4. In a crash, should self-driving cars save passengers or pedestrians? 2 million people weigh in

An anecdote; Eoin Tracy of Southbank Investment Research Ltd, had to attend a meeting in Phoenix, Arizona.  This is where Google are currently trialing their driverless cars.  So while he was there he conducted a mini-survey.  He asked everyone he met what they thought of the driverless cars, did they like them, were they concerned, etc.  He found that all the people he asked were all oblivious of the driverless cars.  So the media (US press) stories of people throwing rocks at the driverless cars may well be biased reporting.

In my view there are two problem areas, both transitional.  The first is the point brought out in the Guardian article on the Uber fatality.  The current phase where cars are not fully autonomous but rely on a "safety driver" who in turn is not paying attention and relying on the automatic handling of the car leaves us with no one in control in an emergency when it really matters.  The second problem I see is the transition between mostly human to mostly autonomous when the mix is half and half.  Then, it seems to me, that autonomous software needs to know if a subject vehicle is autonomous or human driven.  Cars need to communicate with each other and broadcast their intentions, so 5G needs to be fully rolled out.  Is this part of the ongoing research?  Who knows?  Autonomous shipping would seem to be a simpler problem to solve but seems to be behind, or do we just not hear about it?

Copyright © 2004 - JG Weston, all rights reserved.

This document is