A Comprehensive Guide to Autonomous Cars: Technology, Levels, and Current Models

Not so long ago, it seemed totally futuristic to us not having to physically drive for our car to get us from point A to point B without having to take control, but that day is closer than we think.

In today’s blog we tell you what an autonomous car is, what technologies make the car autonomous, what is the classification they have according to their level of autonomy and the current models that incorporate these technologies. Let’s start? Or rather “Start KITT”.

What is an autonomous car?

An autonomous car is one that, with a series of sensors and computer logic, is capable of interpreting the world around it and driving through it as a human would do without your intervention or with your partial assistance.

These sensors allow obtaining information on the location of other vehicles, their position, the distance and the lane in which they are traveling. This, together with GPS information and sufficiently accurate maps, indicate to the car its situation and the elements around it at all times. In some cases, it also allows adaptation to weather conditions.

Technologies that allow the car to be autonomous

Different technologies are being used for the development of autonomous cars. We explain the most used:

For example, navigation through cameras located around the vehicle used by Tesla. This autonomous driving system is based on a combination of 8 cameras, 12 support ultrasound sensors and a forward radar. Overall, the car is able to “see” the environment and interpret it as a person would, but with superior reflections.

Another system that companies such as Google, Ford, Toyota / Lexus, Nissan or Volkswagen are using is LIDAR. The LIDAR is a rotating laser beam emitter, which is usually located on the roof of the vehicles and which is responsible for reconstructing the environment 360 degrees around the vehicle. Through this emission of rays and their reflection, it is able to create a digital map of the obstacles within its reach, and together with support sensors such as the previous case (ultrasound, radars) it interpresses and acts in sequence.

Apart from the main systems that have been mentioned (camera, LIDAR), assistive technologies are also used that are used to improve the information captured, other types of signals that provide a complete assessment of the environment.

These technologies are the ultrasound sensors used for short distances, but due to their low cost, they allow to have a perimeter sensor not only for the navigation of the vehicles, but for the safety in the parking of these.

Another widespread system is radars, which, unlike LIDARs, emit radio waves that bounce off objects and tend to work better in adverse weather conditions.

Classification of the autonomous car based on the level of autonomy

Autonomous cars are classified according to their level of autonomy, that is, the degree of independence that the car has with respect to the driver.

The first classification in this sense was established by NHTSA in 2013, the US transportation safety agency, through a scale of 0 to 4.

As systems have evolved, new classifications and ways of evaluating them emerged, until the current method that is used in practically all cases, and the most widespread, has been reached. We are talking about the SAE J3016 standard, the last revision of which took place in 2016.

This classification tries to group all those that have emerged in the different countries, giving recommended evaluation guides, but that each brand can use or not according to their interests.

This guide establishes 6 levels of autonomous driving, ranging from 0 to 5. These levels only establish the minimum capacity that the vehicle must have to reach each of the levels, and the functions that it must automate.

• Level 0. Without any automation.
  • The driver is 100% responsible for all elements and provisions of driving.

• Level 1. Basic driving assistance.
  •  The vehicle has basic elements that help or free the driver from simple tasks.
  • It can be the automation of speed or trajectory maintenance, but only one of them.
  • All other tasks remain the responsibility of the driver.
  • There is no automatic detection system for obstacles, objects or events while driving.
  • The assistance can only be used in ideal conditions.

• Level 2. Partial automation.
  • Both speed and trajectory maintenance is automated.
  • The driver is released from functions related to the movement of the vehicle.
  • There is no automatic detection system for obstacles, objects or events while driving.
  • The assistance can only be used in ideal conditions.

• Level 3. Conditional automation of driving.
  •  As in the previous case, the speed is controlled both with the path of the autonomous lever.
  •  There is already a system for detecting objects and driving situations, as well as obstacles.
  •  The system can operate without driver interaction unless the system requires it or there is a potentially hazardous situation that requires their intervention. This notice must be provided well in advance so that the driver has time to react.
  • Only allowed to use under certain conditions.

• Level 4. Full case automation of driving.
  • As in the previous case, the speed is controlled both with the path of the autonomous lever.
  • There are enough sensors to give a complex response to situations and obstacles, reacting completely to them.
  • The vehicle has an alternative system in case the main fails or is not able to take control. This alternative system does not allow all the driving functions, but it does allow enough to drive the vehicle to a safe area to stop.
  • There may be situations where the system does not perform properly and you cannot drive.

• Level 5. Complete automation.
  •  As in the previous case, the speed is controlled both with the path of the autonomous lever.
  • It is fully automated the response to unforeseen driving or obstacles.
  • There is a backup system capable of taking control until the vehicle stops.
  • There is no limitation for the use of the system and it can be used in any condition, as a human would.
  • It is not necessary the presence of people inside the vehicle.

 

Current models of autonomous cars

We can find more and more autonomous car models on the market, with different degrees of autonomy. Within the classification provided above, we can find cases of vehicles of different levels.

At level 0, almost all the existing models on the market, both economic and old, which do not have any assistance and require 100% action from the driver. Also included in this group are those with assists such as emergency braking or lane or speed keeping.

If these vehicles add the adaptive cruise control function, we would be talking about level 1, since it is capable of maintaining the distance to the preceding vehicle automatically. Also included are those that incorporate the assisted parking capacity if it only acts on the turning of the wheels. In this category there are also a large number of vehicles on the market.

At level 2 there are fewer models contained in this category. They are systems oriented to use on the motorway, much more elaborate that allow automatic reaction to traffic conditions, which can even stop in traffic jams.

The most representative brands with systems of this nature could be Mercedes-Benz (Drive Pilot), Nissan (ProPilot) , Volvo (Pilot Assist),  Audi or BMW . Finally, if the parking aid system additionally controls the throttle, they are also included in this level.

In level 3 of autonomous driving, the cases of brands or models that incorporate it are discrete. The best known is Tesla with its Autopilot, that allows you to predict traffic conditions on the road and even change lanes to overtake without driver intervention, without limiting its functionality to motorways.

For level 4 autonomy it is extremely difficult to find commercial models, since, although many of them are under development, they are not yet commercialized. The most important is that of Google, which has accumulated many kilometers driving autonomously and is increasingly closer to being a reality.

As in the previous case, there are only prototypes of vehicles with level 5, and the most advanced is again that of Google. There is a variant of this model without a steering wheel or pedals, which has been tested for short tours and in ideal conditions.

Future forecast on autonomous cars

It is difficult to forecast how the evolution of these systems will advance, since each brand shares their horizons in a very ethereal way. What is clear is that the implementation of autonomous driving has to go hand in hand with an awareness of drivers, to avoid accidents because the driver is not attentive to possible events.

Now, it is expected that by 2025 the vast majority of models with up to level 3 autonomy will be released, which would help lay the foundations for the new legislation that has to go hand in hand with these advances.

Since this summer the UN has signed a regulation to limit the use of level 4 and 5 autonomous vehicles and prevent them from operating near pedestrians. According to this document, for now these systems (level 4 and 5) can only be activated on roads where pedestrians, cyclists or similar traffic is not allowed.

There must also be a physical separation between directions of movement, with a speed limited to 60km / h. Finally, it will be mandatory that all passengers are correctly seated and with their seat belt fastened.

Do you want to implement AGVs in any of your Projects? Contact us!