Self driving future


The future of self driving where no human intervention is needed and vehicles can safely go between two places have been nearer and nearer every year. Different companies have put into tremendous effort into this race for the future with use of various Innovative technologies. Self driving is one of the great example of machine learning or AI as all the data vehicles see has to be processed like a human and be better at it.

Various companies nad their technology

machinelearning

Autonomous level determines self driving capability on scale of (0-5).

🔸Level 0 Automated system have not much control except few warnings.

🔸Level 1-'hands on'-Driver have to keep their hand on steering wheel. Cruise control is one example of this Level.

🔸Level 2-'hands off'-Vehicle do all the tasks of steering and braking but driver should be present to intervene if anything happens.

🔸Level 3-'eyes off'-Vehicle runs on its own. Driver should be present to intervene but he can keep his eye off the road.

🔸Level 4-'mind off'-Driver can do his own things and completely forget about the drive. But this feature maybe limited to some areas only where extensive mapping have been done. If vehicle cannot proceed,it will park.

🔸Level 5-"no human intervention"- Vehicle will move between twobplaces with no human intervention. It works on all parts of world and at any conditions.

Lidar vs camera

This debate has long been going on as two forefront companies on self driving have taken very different approach.

Tesla heavily invests its time and money on camera technology while Waymo,one of google company,boasts the superiority of its Lidar technology.

Lidar uses lasers or light beams to sense the surrounding. It can map out its surrounding in 360° precise to centimetres. But lidar vehicles uses maps for self driving which heavily undermines its versatility. These maps are generated either by mapping cars or the consumer fleets. Mapping is one of the key for Lidar and it limits the uses of Lidar into least used roads or outside the mapped zone.

Camera technology uses machine learning to its highest. It analyses all the data from the camera like an eye of human in real time. The cameras in all side of vehicles watch for the surrounding. They cannot precisely measure the surrounding but they are less bulky and least expensive. The main thing is to make AI powerful enough to work better than humans in this technology. Tesla AI have been able to distinguish traffic lights,signs and cones till now. Tesla have been teaching its AI with over 3 billion miles of data recorded from tesla vehicles all over the world.

Tesla claims of Level 5 autonomy jumping direct from Level 2 in 2021 as its Full self Driving beta version shows huge advancement in field of self driving.

Waymo and other lidar technology users have also made huge jump in their self driving capabilities as most of them have already achieved Level 3 autonomy.

No matter who wins the race first,future is bright for self driving vehicles. Legislation issue have been huge problem till now but Level 5 vehicles will change the human civilization foever.

#machinelearning