Google I/O'18 Keynotes: Updates on Self Driving Car technology

Each year Google organize an annual developer festival name Google I/O, where they present some of the amazing stuff they have been working on. Surely since the rudimentary period of Artificial Intelligence and Machine learning, Google has invested a lot and have made a meteoric changes with the help of Artificial Intelligence and Machine learning.

Waymo Self Driving Car

One of the biggest project that google have been focusing on is Self Driving car. Google has started to test on self driving car since the year 2009. Google has transformed its Self Driving car project to an autonomous firm named WAYMO. At the Google I/O'18 Keynotes, John Krafcik, who is CEO of Waymo took over the stage to address the progress they have made in the field of Self Driving cars.

John Krafcik CEO of Waymo

Key points John addressed are:-

• In this year 2018 Self Driving car will be accessible and will be up and ready to help people get around the corners of the city. As of now it is only available for the people of Phoenix Arizona, United States of America.

•  With the help of Deep Leaning which is a part of Machine Learning, Self driving Engineers teamed up with AI Engineers from google and solved the pedestrian detection problem. Well it was pretty amazing that they were able to minimize the rate of error by 100x, which is 100 times better. 

• To avert Waymo car from road mishaps, they have been running the car over 6 million miles on public road, which have surely created a copious database which has every possible real time world interactions.

• To train the neural network for Waymo cars they have used Tensor Flow and Google Data Center TPU's, which has increased its efficiency by 15 times.

• They have also been running 25000 cars across the roads of the city and have traveled around 5 Billion miles for real time simulation.

•  They have also announced that Self driving car won't be limited to drop and pick up facilities. It will be soon integrated with Health and Medical Emergency services and make Self Driving technology available to other autonomous car manufacturing companies as well.

• In difficult weather conditions, the sensors tuned in with the car picks up ambiguous data or noise, which makes it difficult for the car to take decisions, map the track or detect other cars, pedestrians.

This what the car see in the snowy weather.

Distortion in sensor data due to Snow flakes

However this problem was taken care of with the help of Machine Learning (ML). When Machine Learning was applied to the data gathered by the sensors, it was able to filter the distortion caused by snow. After filtering the distortions, the car is able to identify its track mapping and also all the objects around it.

This is how it looks like when distortions are removed.

Filtering distortion with the help of Machine Learning.

Finally, Waymo along with Google cannot wait to make these cars accessible to other parts of city and countries all around the world. Making our road and travel experience convenient and a lot safer.

Check out this cool video of How people are getting along with Self driving car.

Also Check out Google I/O'18 in 10 mins

Thank you for reading out my blog. Please give your suggestions in the comment box below and click the Subscribe Button to stay connected.