Hello World!!!! It’s JITHU. Today we are talking about Self Driving Cars. What Self Driving Cars?????? That means no driver!!!. How could I bet my life on a computer? Yes, you can bet because it,s already proven that computers drive the car better than human but not as better as racers actually it might take some time.
Actually, Self Driving Cars already entered the market. Self Driving trucks already shipping the beer long distances. It reduces the cost for the customers because there are no humans. It reduces the accidents better than using airbags.
We are using Deep Learning for the computers to detect who’s life is important? we are collecting data from the human drivers and feed into the algorithm.
OK. now we will know discuss what factors does this depends on? Mainly it depends on 5 things
The name itself refers that visualizing the world by the camera. Here as there is no human the cameras are fitted on the top of the car so that it could cover in a 360-degree view while the human can cover only one direction at a time. And also it covers superhuman eyesight. Here the camera detects an object and that called classification in deep learning and so that it could detect what the thing is? For example, we want the computer to detect the stop sign we feed the algorithm we the thousands of images on a stop sign. So, that whenever later it sees the stop sign it can be identified.
Senor Fusion is nothing but combing the sensors. Here we are not using the cameras, But also Radar and Lidar. The only camera can detect up to some extent, camera cant detect the weather, distance tracking, velocity tracing, and Illumination. There the Radar and Lidar comes to the field. Radar is good at telling you that there is something in front but can’t able to say what it is, then lidar comes. The combination of both Radar and Lidar data we can easily detect things.
Localization means knowing that where the car is in the real world, after knowing about the real world. Know we can only use GPS and it gives for 2 meters. that leads to accidents. So we are using sophisticated algorithms so that we can up to the 10 cm.
This is the next step after knowing where the car is in the real world. The car creates the trajectory and knows where the car wants to go. First, it predicts how the vehicle is moving and decides in what direction should it go. According to the trajectory it says where to go.
Once the trajectory is created the car should move the steering and brakes and acceleration, according to the trajectory its slight a tricky one, but the cars are getting better nowadays. Control theory is the study to know how much the force should we apply for the objects to move.