Driving the future towards Driverless Cars
Did you ever wonder how many hours you spend driving in your entire life?
Don’t worry. We got you covered.
According to a survey, , if a person starts driving by age of 17 and drives until he is nearly 79 then the person spends almost 37,000 hours driving the car throughout his life covering 1,284,256 kilometers, approximately it is three times the distance taken to drive to the moon and back.[1]
Now Imagine… may be not completely, but at least a part of the 37k hours being saved. That is not by taking other transport or by appointing a driver but with the help of a self-driving car. YES!
These driverless cars are capable of driving without any type of human interference. For some, It sounds a little spooky but it’s not. Thanks to Artificial Intelligence. These driverless cars are built to use data that is processed using image recognition systems, neural networks, along with machine learning to drive.
These cars not only save time but they do save lives too.
According to a study, all around the globe, nearly 1.3 million people die of road accidents every year I., e.. an average of 3,287 deaths per day.[2]
94% of these accidents are because of human error, 2% because of environment, 2% by the vehicles and 2% because of other factors.[3]
On a positive note such human error can be avoided by these cars as these cars do not get distracted, they obey traffic rules, do not drive fast, do not drive drunk, they do not fall asleep.
These driverless cars are capable of adjusting vehicle’s speed in accordance with other vehicles and objects on the road, as it uses automatic brake systems in times of emergencies making it a safe ride.
Automatic parking systems, autopilot systems makes them user-friendly
· Car’s capacity to read traffic-signs makes it adaptable, and there are many other features which make it more safe to not just the person driving but for the pedestrians as well.
But how does the car manage to identify the road signs, other vehicles on the road and people too?
But how will identify the road signs, vehicles and people on the road ?
The LIDARs and sensors present in the car, use computer vision and translate the image into a relevant object, features, or patterns for the processing unit of the car which is responsible for decisions taken by the machine. This is where Deep learning takes place. Deep Learning is a machine learning technique, where machines learn features directly from data and such methods of deep learning are proven provide a better accuracy with regards to image classification than human. So, the more the machine works on the image the more features of the objects are learnt by the machine.
A Developer builts a machine learning model involving the training data to train the machine A deep learning framework is used to make this process easy and quicker. The components of such framework help in designing, training and validating the deep neural networks which process data using mathematical models. Some of the popular Deep Learning frameworks are Tensorflow and keras(google), mxnet (Amazon), PyTorch (Facebook), Caffe ( Berkeley).
In simple terms, Image Recognition helps the machine recognise objects and people from the images that are fed to it as data.
Before feeding the machine, these images go through a process called annotation.
Annotation, a process of labelling and annotating data for training the machine(car) to recognise the objects on the road. This process helps in giving an accurate reference to coordinate with the reality of the traffic situation. There are various tools used for annotating an image which include rectangle tool, cuboidal tool, polygon tool, circle, point, and bound boxing tool. Based on the shape of the object the tool is to be selected.
Vehicles like cars and trucks are annotated using cuboidal tool. This gives a 3D effect to a 2D image.
As the road signs are normally square or rectangle the rectangle tool can be used
Polygon tool is be used to annotate human as they do not have a definite shape.
In this process, labelling is the first step which means giving a name to the object and then one of the tools are used for annotating the object in the image.
To make annotation walk on cake we have OCLAVI. It provides
• Tools which annotate objects even low-quality images with precision.
• the developer is given a choice of approaching the company's freelancer, or even bring his team.
• the developer can connect to cloud storage for getting things done without any disturbance in model building.
These self-driving cars can be considered as a revolution in the industry of automobiles and AI is the head behind such revolutionary technology. Technology like this are a boon to the mankind and it is our part to use such technology to the fullest and make a right use of it.
Explore more OCLAVI
References:
https://offthethrottle.com/blog/2018/04/09/much-time-spend-cars/
http://asirt.org/Initiatives/Informing-Road-Users/Road-Safety-Facts/Road-Crash-Statistics
https://blog.lawinfo.com/2017/09/06/human-error-causes-94-percent-of-car-accidents
