Safety vs Utility : A Tale of Autonomous Vehicles

Imagine a fully autonomous level-5 vehicle , stuck on the road due to red light. Standing behind a loaded truck, it has traffics in both side as it stands in the middle of a 3 lane highway. Just when the truck begins to move , the loads started falling behind and those heavy , gigantic & round shaped bulks were about to roll over the car. What should that do then ? Hit the next-lane cars ? Let the car itself get crushed ? Even if it hits the next-lane car , which one is to be hit ? This thought experiment by Ted-ed, is a model of famous “ Trolley Problem” . On refuting the famous statement “ Automation is the perfect solution for car crashes ” , some of the tech-geeks chose the ethical response citing Trolley Problem , Downs-Thompsons Paradox or even the idea of “Freezing Robot Problem” . On top of that 2016 TESLA Car Crash in autopilot mode fuelled the debate as the car mistook the side of a truck for a washed out sky and ended up crashing into that . In brief the key problems the autonomous vehicle producing firms looking for solutions are

Predicting the upcoming move of a human driver and communicating with him.
Detecting obstructions or objects distinctively when they are situated very close to one another.
In case of an accident who is to take the accountability of that incident.
How the software ethics would act while the vehicle cannot but damage any of the stakeholder during intense situations like Trolley Problem.
Ensuring the system of the car not falling prey to Hackers !
How would the computer process data from the cornucopia of sensors like LIDAR , Camera , Ultrasonic Sensors , Radar and so on and still take the best possible action while a human a could have taken that decision in split second.

Well… it’s not just the hopeless pessimistic analogy. “Waymo” has already started upgrading their AV’s while Ford is testing more than 30 of their cars. Similar kind of progress we can also see in Tesla as they had launched “Autopilot Mode” in their cars.

Figure : Overview of the range and placement of the sensors of an autonomous vehicle ( Collected from www.intechopen.com)

Let’s find out how the automated vehicles have come this far :

AV’s are designed with lots of sensors placed back and forth. Radars which are commonly used are being assisted by LIDAR that use light wave instead of radio wave for helping the cause of quick identification. It helps understanding the objects placed around the vehicle.

While LIDAR adds an unprecedented dimension to the AV , camera has been used mostly to read traffic signals and signs . Experts came up with an idea to solve the problem of communication of the self-driving vehicles. In order to setup communications with other vehicles , there has been added LED Display that would show “ You can overtake me ” or “ I would get slower in a minute ” . It can also be understood by the AV’s and thus can address a pressing problem. For safety concerns , the mathematics of the movements of the vehicles are determined with this formula .

Figure : Safe Distance Formula ( Collected from EETimes)

While all these questions have been raised and tech-giant corporations are looking for the solutions, utility of autonomous vehicles cannot be undermined and they can be laid out as it saves a huge part of our time providing enough hours to complete the chores or doing things we love . Alcoholic driving can be completely alleviated , speed limits and traffic rules could be followed with perfect accuracy . This list can be endless but still the debate over safety vs utility outweighs everything else .

Let’s see if Waymo , Tesla or Ford can answer this question !

References :

[1] Is There Math for ‘Drive Cautiously’? by Junko Yoshida on EETimes

[2] Autonomous vehicles and the truth about LiDAR sensors by Majeed Ahmad

[3] Sci-show

[4] Ted-Ed

[5] www.intechopen.com

*This article got published in IUT AMBER 2019

Google Sites

Report abuse