"DRIVERLESS CARS" is the Topic for a Panel Discussion at MIT


Error message

Deprecated function: The each() function is deprecated. This message will be suppressed on further calls in _menu_load_objects() (line 579 of /home/neighbormedia/public_html/includes/menu.inc).
Stephen Kaiser's picture

Questions About the Future

    On Monday morning June 13 MIT hosted a panel discussion on "Driverless Vehicles."  The panel was sponsored by the Metropolitan Area Planning Council (MAPC) and a Boston advocacy group, Transportation for Massachusetts.

    Driverless vehicles are a concept that has been largely stimulated by the independent research by Google.  The idea is to have cars that are controlled in their movements by computers and radar sensors so that they can be moved automatically along public roads.  Designers are hoping that the idea starts small and transitions into a much larger idea.  The ultimate system would be for all roads to be composed entirely of driverless cars.

    Planners have recognized a whole range of intermediate technologies.  Some of these variations go beyond cruise control and could steer a car and keep it in lane ....  assist in doing parallel parking .... providing emergency stopping to avoid accidents .... and numerous other concepts.  These transitional technologies all involve sharing of control of the car between car and computer.  The computer would be capable of overriding the driver in certain circumstances, and the driver could override the computer in others situations.    Currently most of the focus is on changing the cars, and not making any changes to roads until later.

     Speakers included MIT Urban Planning Professor Chris Zegras ... Tony Dutsik, Policy Analyst with the Frontier Group ... Mark Draisen, executive director of the MAPC ... Lauren Isaac, Manager of Sustainable Transportation at Parsons Brinckerhoff ... Jonathan Koopman, Senior Engineer with the Volpe Center ... Mayor Joseph Curtatone of Somerville .. and Kent Larson, from the MIT Media Laboratory.  

    The sub-topic for the meeting was "What will change and when?" Many panelists discussed possible changes could occur in the way we get around the city and how we plan the activities throughout the day.  Some speakers listed advantages and disadvantages, but no speaker hazarded a prediction about when driverless cars would become a practical option for the Boston area.  

    The reason for such caution may lie in the many novel issues and unresolved design concepts which have yet to be studied in depth. The field is progressing and changing so rapidly that new technical developments occur almost every week.  The Internet and enthusiasm for driverless systems has generated much energy from MIT, Google, Apple, the global automobile industry, politicians and government bureaucrats.  Federal transportation officials are proposing spending $4 billion in the next year on research into the practicality and safety of driverless cars.   
    Prof. Zegras began with references to the history of major transportation initiatives (including the ill-fated Inner Belt). Some of the transit implications have been realized just this year as evidence came out of Uber in California of many passengers using Uber vehicles to access mass transit service.  Single vehicles can be used to carry multiple passengers, and reduce the tendency of single-occupant vehicles to contribute to severe congestion on roads.
    Tony Drazik emphasized the safety aspects of our current highway system in Massachusetts.  In 2015 there were 349 high deaths, with over 4,000 injured and estimated costs for highway accidents of $5.8 billion. The clear implication of this safety data is that driverless vehicles have the potential to reduce the frequency and severity of road accidents.  By this month, one reported accident has been caused by an automated vehicle, and that between a very slow-moving bus and a Googlemobile.  However, I have not heard of any injuries in any accident.   The industry has not yet had its Titanic incident.
    Lauren Isaac is manager of Sustainable Transportation at Parsons Brinkerhoff, a venerable transportation firm that did much of the early construction for the New York subway system and later moved into highway construction. She had previously been working in ride-sharing programs, and has been helping provide a better understanding of the pros and cons of driverless vehicles and how to deal with local and regional governments and the regulatory needs.  

    While Google's original concept was for small electric minicars, modern concepts have expanded to include larger vans and transit vehicles, and Uber has been considering the possibilities of developing an automated fleet of vehicles, thus avoiding the uncertainties of using private cars and drivers. One concern of the planners is the possibility of driverless cars actually producing more vehicle trips and thus contributing to roadway congestion.  Here the option to allow shared-vehicle  rides would allow for an expansion of the transit function of driverless vehicles and would reduce vehicle miles traveled. Another possibility is to use driverless cars as a way to feed riders into an existing urban transit system, so that downtowns would not get clogged with more cars.

    Jonathan Koopman from the Volpe Center represented the engineer on the panel and was in a position to report on international developments in the field, as well as research into the safety aspects of driverless vehicles. His past experience was with Volvo truck division, and hence he recognizes the diversity of vehicles using the public roads and which would interaction with regular and driverless cars.

    The Volpe Center also does considerable transportation work with MIT, which may be the local hotbed of driverless vehicle research. (Google has been doing most of its research in California).  California and the Federal Government have been stimulants for investigation and regulatory initiatives, with the Feds proposing to spend $4 billion on research in the coming year.
    Mayor Curtitone spoke of the numerous efforts the city of Somerville was making in both new development and transportation technology.  He spoke of more traditional technologies which would seek optimal traffic signal timing to improve vehicle flow at now-congested Union Square. This area is now targeted for over 2 million sq. ft. of new commercial development.    
    Kurt Larson from the MIT media lab used visual simulations to illustrate the potential for new vehicle technologies to be used in the city, including a "dual mode" bicycle which could arrive under automatic control and also be used in manual mode. His group has been taking the widest look at the full width of options for automation of all transportation modes.

    Some features and virtues of driverless vehicles are unclear or in dispute.  It does appear that parking needs and actual supply can be reduced, but there is a wide range of expectations in terms of vehicle travel.  Some analysts see vehicle miles traveled (VMT) dropping by more than 30%, but that conclusion appears unlikely because of the increase in empty vehicles in circulation.  Other analysts are forecasting up to 270 percent increases in VMT.  Such an increase would be untenable for any urban or suburban community suffering from traffic congestion today.
    Beyond bemoaning the severity of accidents (aka "crashes") across the state, the presenters could provide no estimates of how effective driverless cars would be if applied to all roadways.  No one wanted to touch the safety issues associated with transitional situations when driverless cars in large numbers are mixed with conventional vehicles.

    The safety situation remains buried in safety and liability uncertainties.  For example, in Massachusetts, driverless cars would need to be able to drive through rotaries.  They would need to make left turns at unsignalized intersections --  where in Boston drivers charge half way across the intersection and then bull their way through the other half in a separate movement.

    The panel was composed almost entirely of advocates for driverless vehicles, and the panel session was intended to help spread enthusiasm for the concept.  There were no arguments or points of contention among panel members.  No one set out a listing of the great unanswered questions that need to be asked and resolved.  No one offered an action agenda to move into the next step of investigations and detailed testing.

    With the dynamic and incomplete condition of ongoing research, some of those unanswered questions can be assembled :

* How much will the price of cars be increased?  Will the gadgetry wear out faster or could it last 15 to 20 years (longer than most car components)?

* What happens with the annual auto inspections?  Will they become more complicated and expensive, or could the inspection be done by simply plugging a cord into the car and everything is tested automatically?

* What happens with cars in a fender bender or more serious accidents?  Who certifies the repairs?  Will such vehicles require mandatory safety inspections?

* What would be done with the used automobile market?  How is quality control maintained?   
* How does an automated vehicle turn into a residential driveway?  If a contractor realigns the driveway, how is the vehicle told of such a change?

* Suppose two cars approach each other on a narrow road, and one needs to back up to let the other one get by.  How does the computer-controlled car do that?

* What does one do about motorcycles and bicycles?  Will they always be in the vehicle stream, or does someone have a plan for automating them too?
* What does a driverless car do when it reaches a dirt road?

* Black ice is a winter condition where ice forms but drivers often cannot see it.  Could driverless vehicles be more sensitive to black ice and be able to detect it more readily?
* How do we anticipate weird roadway conditions and events.  Here is one example : Over two decades ago, on I-290 near Worcester a man was driving his car along at 60 mph.  He had prepared to load his car by placing his young baby's car-seat on the roof while he got ready to leave home -- and he left the car-seat on the roof by accident.  On I-290 the wind forced the car-seat to slide off the roof and onto the road  with the baby still strapped inside.  The car-seat simply slid along the pavement, slowly losing speed.
    Traveling just behind this car was an elderly man who saw an object slide off the car's roof.  Puzzled, he gradually slowed down and came to a stop in the lane, with his flashers going.  The father realizing he had made a mistake rushed back to the scene and was reunited with his son.  How would an automated vehicle have responded to such a situation?

*  What happens to people who own antique cars?  Can a driverless car co-exist with a 1937 Hudson?

* Finally, what do governments and industry do with the driving enthusiasts, who enjoy driving their cars and do not want to give it over to a computer?  They could see owning and driving a car as a vital freedom, and would resist anyone taking away this right.  Would there be a furor equivalent to gun rights, where gun owners resist any effort to ban guns in any way?  This resistance by traditional drivers could become a key Constitutional battleground.