Custom Search
 Hi
Featured Services
Tutorials on Embedded
System Design .
Tutorials on
Antenna Design.
Tutorials on
Image Processing Using
Open-CV.
read more

LATEST ARTICLES

HELP SHARE IDEAS.COM
Mail us at : info@helpshareideas.com

Are Self-Driving Cars Are Safe?

We are moving closer to the once unimaginable self-driving car becoming a reality, the question that strikes our mind is “Are Self-Driving cars are Safe?” Have you started to wonder on this?. Not only automobile companies like ford but also Software Giant Google is also putting its lot of investment on self-driving cars. The big question that arises in mind will be are they safe? Let us analyze the same in this article.

HelpShareideas- Tech news image

Click on the image to goto Article Source -It will open in Newtab

The Car manufacturers will have difficulty demonstrating just how safe self-driving vehicles are because of what’s at the core of their smarts: machine learning. The decision taking capability of the machine.
Some of the market research firms project that the self-driving car market will be worth $87 billion by 2030. As we all know, several companies, including Google, Tesla, and Uber, are experimenting with computer-assisted or fully autonomous self-driving car projects—with varying success because of the technical obstacles that must be overcome in their implementation.
The fact accepted by several researchers who believe that the nature of machine learning makes verifying that these autonomous vehicles will operate safely is very challenging.

Traditionally, engineers write computer codes to design and meet their requirements and then they perform tests to check that whether they met the desired results.
But the case with machine learning is that it lets a computer grasp complexity —for example, processing images taken at different hours of the day, yet still identifying important objects in a scene like crosswalks and stop signs—the process is not so straightforward. The difficult thing about machine learning is that how to write the requirements.
Engineers provide many human-annotated examples—say, what a stop sign is, and what isn’t a stop sign. An algorithm strips down the images, picking unique features and building a model. When a computer is subsequently presented with new images, it can run them through the trained model and get its predictions regarding which images contain a stop sign and which ones don’t.This is an inherent risk and failure mode of inductive learning as when we look inside the model to see what it does; all we get are statistical numbers. It’s a black box. Hence we don’t know exactly what its learning.

HelpShareideas- Tech news image

Click on the image to goto Article Source -It will open in Newtab

To make things more concrete, imagine if you test drive your self-driving car and want it to learn how to avoid pedestrians. So you have people in orange safety shirts stand around and you let the car loose. It might be training to recognize hands, arms, and legs—or maybe it’s training to recognize an orange shirt.

Or, more subtly, imagine that you’ve conducted the training during the summer, and nobody wore a hat. And the first hat the self-driving car sees on the streets freaks it out.There are an infinite number of things,” that the algorithm might be training which we don’t know until fully tested.
Google researchers once tried identifying dumbbells with an artificial neural network, a common machine learning model that mimics the neurons in the brain and their connections. Surprisingly, the trained model could identify dumbbells in images only when an arm was attached.

There’s also the challenge of ensuring that small changes in what the system perceives—perhaps because of fog, dust, or mist—don’t affect what algorithms identify. Research conducted have found that changing individual pixels in an image, invisible to the unaided eye, can trick a machine learning algorithm into thinking a school bus is not a school bus!

The Research on study to determine the best tests for autonomous vehicles are going on but they are going to be very costly. As of now we can’t exactly say whether Self-driving are safe or not until we have them on roads and we ourselves experience the riding.

Author: Satish Kumar V

IEEE Graduate Student Member, Bangalore Section.

Sathish Kumar is a Ph.D Research Scholar from Visvesvaraya Technological University Belgaum Karnataka, India. His area of interest is Communications, Embedded System Design and Image Processing.He has more than 6 papers presented in National and International Conferences.


<: Click the link LATEST ARTICLES to access the Latest Articles Page:>