Artificial Intelligence and Business Automation in Automobile Industry
Abstract
Artificial Intelligence has the automotive industry under the spotlight. Hybridization, electrification, connectivity, and stand-alone driving are the main directions pursued by all major automotive manufacturers in the world. It all involves developing the artificial intelligence of vehicles by forcing builders either to develop their own IT divisions or to engage with collaborations with companies in this field. Artificial Intelligence (AI) already exists in most of the cars on the road. Computers that manage engine operation, tire sensors, ABS and ESP are the most readily understandable uses of artificial intelligence in vehicles.
Reasearch
An autonomous car is a vehicle that uses sensors, cameras, radar and artificial intelligence (AI) to travel between destinations without the human involvement. A car is completely autonomous when it’s able to navigate without human control to a destination over roads that haven’t been adapted for it.
Self-driving also advances rapidly. Many cars can park alone or move while keeping the lane without driver intervention. Levels 1 and 2 of "autonomous driving" are already available and the third – where the driver and the car can transfer their responsibility for driving control – is a few steps away. However, it is only in 2025 that the transition to level 4 – long distances without the driver's intervention on the condition of separation from the opposite traffic – is foreseen – and strong controversies have emerged about level 5.
According to the initial studies, this last level, which means that the steering wheel and the pedals are no longer needed and that passengers can stay in the car without getting involved in driving, could have been in use around 2030.
In the automotive industry we have 6 levels of systems, see fig.1.
Fig.1
In January 2014, the Society of Automotive Engineers (SAE) classified the future of the car and defined six levels of automation, from Level 0 to Level 5, precisely for the public and even the automotive industry to speak a common language.
Level 0
The driver deals with everything that means driving the car. No system intervenes and, at most, the driver is assisted by a warning about certain aspects / dangers of driving. The driver must always be 100% involved in driving.
Level 1
The autonomous driving system can take control of the direction of acceleration (ie deceleration / brake) in limited driving scenarios. All the other functions of the car are handled by the driver. In addition, the driver must be ready at any time to immediately take over the car's full control.
To reduce the abstract level, we can call here two technologies that make a car to be Level 1 in terms of self driving.
A. Lane Departure Prevention (LDP), which maintains the lane if the driver is inattentive or falls asleep at the wheel.
B. Adaptive Cruise Control (also known as ACC – Adaptive Cruise Control or Cruise Control Adaptive), which can make a car equipped with such a thing track the vehicle just ahead (from motorway speeds to stopping) without the driver to intervene on the accelerator pedal or the brake.
Level 2
The autonomous driving system can take control of both direction and acceleration (including deceleration / brake) simultaneously in certain driving scenarios. For example, running on the highway in relatively light traffic on a properly marked road surface.
The driver must always monitor the (autonomous) driving action and must be ready to take control immediately if the system asks for it.
Level 3
The autonomous driving system can take the steering and throttle control (including deceleration / brake) in certain driving scenarios. Unlike Level 2, the system is able to identify the limits in time and notify the driver to take over the vehicle.
The driver must no longer monitors the (autonomous) driving action, but he must be prepared to take control of the machine if the system asks for it. Compared to Level 2, the driver demand will not be immediate, but gives him a 5-10 second respite. The system will say messages such as "please take your machine's order in 5, 4, 3, 2, 1 …". If this does not happen, the machine is able to pull on the right and stop the damage (the system thinks the driver is asleep or has a medical problem).
Level 4
The driver can fully delegate the driving task to certain driving scenarios, being able to use the steering, acceleration and braking alone, turn signal lights or lane lights, wipers, headlights etc.
At this stage, the driver is no longer needed to monitor the car or back-up, as long as it remains in the driving scenario to which the autonomous car can handle.
Level 5
The machine can take over the command completely in any driving scenario. The driver may be completely absent from the vehicle, the machine is completely autonomous. Even the steering wheel and pedals may be missing at this stage.
How AI works for autonomous cars?
A basic requirement for advanced sophisticated driver assistance and automated driving is a thorough understanding and accurate assessment of the entire traffic situation. To enable automated vehicles to take control of drivers, they need to develop an understanding of the actions ready for all road users to make, so the car can always take the right decision in different situations. This task is best accomplished through training algorithms using automated advanced learning methods.
Just as drivers perceive the environment through their senses, they process this information using their intelligence, make decisions, and implement them with their hands and feet to control the car, and an automated vehicle must be able to do all that. It can be done here if vehicles get to act at least the same way as people's senses.
Artificial Intelligence opens new possibilities for visual modeling assisted by artificial intelligence. For example, AI can detect people and interpret their intentions and gestures.
Just like humans, artificial intelligence systems have to learn new skills: people in a driving school, AI systems using "supervised learning". To do this, the software analyzes huge amounts of data that includes situations where it has been successful and unsuccessful and then applies the experience to the vehicle. This essential skill of learning algorithms is continually developing. For advanced driver assistance systems, the right dates for this type of learning are available, for example, in the form of radar signals and cameras recorded from real driving situations.
Ford claims that in the future autonomous cars will be able to run on public roads without stopping at intersections. The manufacturer tests a technology that allows vehicles to "communicate" with each other, and the speed is adjusted so that collisions are avoided. Thus, the road would be traveled without stopping to the destination.
Ford is exploring new dimensions in developing autonomous cars in the future. The constructor imagines a world where intersections will no longer have traffic lights because autonomous vehicles will be connected so that their speeds adjust and collisions will be avoided.
Intersection Priority Management (IPM) is tested by Ford in the United Kingdom as part of a program to reduce the number of road accidents and to respond as quickly as possible to rescue crews.
IPM uses Vehicle-to-Vehicle (V2V) communication to coordinate autonomous vehicles approaching the intersection so that the speed is adjusted for each and the machines cross the junction safely without stopping at stop.
Bibliography
1. https://timesofindia.indiatimes.com/business/india-business/artificial-intelligence-driving-future-of-car-industry/articleshow/64009927.cms
2. https://www.inf.ed.ac.uk/teaching/courses/nlu/assets/reading/Gurney_et_al.pdf
3. https://media.ford.com/content/fordmedia/fna/us/en/news/2018/07/24/ford-creates-ford-autonomous-vehicles-llc.html
4.