Self-driving cars are about to come to a street near you. The integration of AI in the automotive industry is making remarkable changes to how we manufacture and use vehicles. Manufacturers are continuously integrating artificial intelligence into every feature of the automotive industry. The industry has been evolving hugely, providing some of the most advanced auto tech features until now.
Evolution of automotive artificial intelligence market
The rise of industry standards such as blind spot alert, ADAS(advanced driver assistance systems) and ACC (adaptive cruise control) would give rise to the evolution of automotive artificial intelligence market. According to MarketsandMarkets report, the automotive AI market is predicted to reach 10,573.3 million USD by 2025 from 782.9 million USD in 2017, at a CAGR of 38.46% during the forecast period.
Driving Assist has been used from the year 2015-2017. Driving Assist is a technology developed to prevent accidents by helping a hands-free vehicle continue to be on a track. This is not a good example of a completely autonomous driving and the possibilities of mistakes are quite high. It is designed not to see everything around us i.e in a 360-degree view, which is why it doesn’t alert or escape from crash situations. So, the driver needs to be alert while driving.
But from the year 2018, autonomous driving cars are expected to gain widespread popularity across the globe. These are completely hands-free vehicles with limited scenarios, where it can only work on national highways. Autonomous driving means there is no need to keep an eye on the road for a driver. The system will provide 10-30 sec downtime for the driver to gain control and if the driver doesn’t, the system will move the car out of the way and stop it.
Fully autonomous driving cars are expected to dominate the automotive industry in 2021, where hands-free vehicles can drive successfully in cities with high mobility, just like Uber and Ola without a human driver. By 2035, more than 18 million partially autonomous vehicles and 12 million fully autonomous vehicles are predicted to be sold per annum globally.
Why autonomous driving cars or self-driving cars are required?
Some of the reasons for the existence of self-driving cars are:
Improves the utilization of Cars
Mostly, cars are utilized only 4 percent of the time, and are placed in a parking place for the rest 96 percent of the time. People invest a lot on a car, still it is not used well. The reason may vary from person to person. For example, few people may think that they are very tired and can’t drive. Self-driving cars can be utilized well in such situations.
Reduces cost per kilometer
Some services like Lyft, Uber, Ola etc., offer mobility-on-demand services by evaluating the cost per kilometer and the driver consumes 50% of the cost. Self-driving cars can help you save a lot in this regard. You can completely discard the driver by adopting hands-free vehicles, and shifting to electric fuel reduces the cost per kilometer undoubtedly.
Manages Add-On Services
As autonomous cars have passengers those are not engaged in driving, they can think and share their queries. Sensors in the car track what you are speaking about and respond to all your queries. They will give you some ideas like where you can get some good food and coffee on your route, show discount offers in shops on that route etc. Tech companies offer similar services and improve their productivity by collecting data from the passengers.
Artificial Intelligence in self-driving cars
Autonomous driving cars need to understand the data stream moving from a wide range of sensors such as lidar, cameras, radar, and ultrasonic to react precisely and immediately in a fraction of a second. The first step is to analyse what the sensors gather and what the cameras see. This can be achieved successfully through artificial intelligence technique, which can direct cars to identify different objects. Three features play a crucial role for a car to drive on its own, which is why artificial intelligence is used in cars.
Credit: U.S. Department of Transportation
Sensing- Autonomous cars require sensors to detect the surrounding objects
We humans will have senses such as hearing, touching, and seeing. These collect data and send to the brain that guides us to take secured actions. Similarly, sensors need to be provided for a self-driving car that covers complete 360 degrees using cameras and radar scanners. The laser scanners gather the data from all the sensors and send to a high-performing computing device to develop an environment model that can:
notify where all the objects are present around the car such as cyclists, pedestrians, signals, road turns, barriers, symbols etc.
notify the status of these objectives
notify about the way i.e whether a car is approaching a highway or not
Mapping- Detailed and updated maps are required for a self-driving car to drive successfully
Mapping that you provide should not be confusing because it helps the vehicle to know everything about the drivable paths, the delimiter, the roadway etc. So, how you design and update these maps in an effective way is important. These maps should show the things that are actually seen and must be updated according to the changes in the environment immediately. The maps that are available now are not updated as such. This brings up the need for crowdsourcing to design detailed maps for the self-driving cars that will be popular in the next few years. Without these updated maps, it is not possible to drive a hands-free vehicle.
Driving Policy- Needs to be implemented carefully to understand the traffic negotiations
A broad intelligence needs to be enclosed in autonomous driving. A car must be able to inform what vehicles and pedestrians are doing and needs to understand driving policies. This knowledge doesn’t come automatically to us, it comes with both driving lessons and experiences. Since it is already confusing for humans, systems are going to face more difficulties, because it is very difficult to understand some situations like which vehicles to give way to etc. Moreover, driving policy varies according to the places, as traffic negotiation is not same in all the cities. So, effective measures must be taken to implement driving policies into the system.
Few companies that are already in this field
Audi deals with legal, social, and ethical issues of autonomous and automated driving. The Audi designed Aicon that drives completely autonomously without any steering wheels and pedals is expected to launch in 2018.
Waymo Intel Collaboration
Intel collaborated with Waymo to launch autonomous driving platform for a better and safer world. Waymo cars with Intel technology has already operated some self driving cars on U.S. roads.
It is one of the largest AI and internet technology company announced that they are planning to deliver Level 3 vehicles by 2019 and Level 4 autonomous cars by 2021.
Although many tech companies and government executives are concentrating on self-driving cars globally, they are still facing many problems with respect to technology. Six accidents have been registered with these self-driving cars so far, and among these, five vehicles are from Google. Autonomous cars technology is emerging and is still subjected to continuous research. Maybe not today or tomorrow, but they are coming soon into the real world. WinterGreen survey reported that 90.87 million self-driving cars are predicted to be on the road globally by 2023.