Self Driving Cars

What are we getting ourselves into?

© 2013, 2016, R. W.C. Stevens


The following is extrapolated from, observations made in a Letter to the Editor:
Letter by Jim Surjaatmdja, P.E.; to Mechanical Engineering magazine (a publication of the American Society of Mechanical Engineers).

  It has been proposed that vehicles be driven by computer-control. Taking a (potentially distracted) human driver away from the controls, and transferring the driving ‘intelligence’ to a computer – especially one that will be in continual communication with other vehicles sharing the roadway – will get more people where they want to be, faster and safer. Or so it is supposed.

  What will happen if pedestrians learn that it is safe to step in front of a computer-guided vehicle? What will happen when the youth fail learn that it is unsafe to be in front of a computer-guided vehicle? Either we will need additional infrastructure to deal with policing of ‘jaywalking’ laws, or traffic congestion due to ignorant pedestrians will become rampant!

  Perhaps part of a child’s upbringing and growing experience will be a multitude of rude awakenings, experienced as passengers in vehicles avoiding collisions. Will it make them more aware of the social impact of selfishness, and therefore more responsible citizens? Or will it tend them towards the thinking, ‘Others do it; and so I may too!’?

  If it makes better citizens of them, then it will be a self-limiting feed-back control. More bad experiences in one generation will lead to better people, and fewer teaching experiences for the next generation – leading to less-better people (or fewer better people) – more bad experiences ….

  If it makes more selfish citizens, then the system will strain and occasionally a jaywalking incident and a vehicle failure will coincide. When such happens, ‘Darwin's law’ may eradicate the insolent, selfish, challenger of the system.

  Interesting law cases are also foreseen. Vehicle occupants will sue pedestrians for (by way of a few examples): Damage due to spilled coffee (with damages ranging from dry-cleaning expenses to pain and suffering due to scalding); Damage to unsecured parcels, tipped off the seat, and smashed on the floor; Collateral damage from flying, unsecured parcels.

  Even if, instead of trying to totally avoid pedestrians, the computer intelligence is used to calculate and balance the damage to pedestrians, versus the damage to the vehicle and its occupant(s), there will certainly be legal challenges over the validity of the assumptions made. Should vehicle occupancy be based on the number of seats with weight on them before the seat-belt was buckled; or vehicle maximum capacity? Should the presence of unsecured cargo be assumed; or should the ‘driver’ have made effective efforts to prevent its shifting in emergency maneuverings? Indeed, should a seat, with weight but with the belt unbuckled, be taken to indicate the presence of a [presumed] non-fragile parcel; or the presence of a critical passenger who can not wear a belt, and so needs and deserves transport more gentle than the norm?

  Case law may make the determinations of blame in the situations laid out in the paragraph above. Worse, the case-law determination in one jurisdiction may not agree with the determination in another!

  Is this really the direction in which we, as society, want to be advancing?


The following, simpler observation, is by another, to whom I would willingly give credit, if only I could find the source.
Perhaps, I first heard it on a podcast from the BBC Radio Service.

  Machine vision may find it easier to identify a shopping cart, than one of the many designs of baby stroller, baby carriage, three-wheeled ‘racing’ stroller, baby-trailer for a bicycle, etc. Shopping carts are known to, at times, carry babies (and at other times shoppers use the ‘baby’ area to pile items which may, to a machine vision system, be indecipherable from a baby). Baby strollers are often enclosed, and even people can not see if they are occupied, or empty of precious cargo. Thus:-

  The first time the wrong choice is made between these two objects, the futurist went on to point out, the headlines may read
Robot-Car Sacrifices Baby, to Save Groceries
and that will be the end of autonomous cars.


Driverless-Cars should probably be able to recognize and cope with situations, some peripheral to regular driving, such as the following:


Other vocalism comments are indexed here.
Permission to redistribute this copyrighted article may be easy to obtain.


 Robert’s Home Page   The latest version of this page may be accessed at
http://www.wendygamble.com/RwcS/vocalism/Robot-Cars.html
 Pleased To Be Of Service, RwcS.