According to a report published in 2021 by the consultancy firm PWC entitled “Digital Auto Report”, it is estimated that by 2035, 14% of registrations will be of autonomous cars in Europe, China and Japan.
To achieve these figures, car manufacturers need to make major advances in software. In addition, PWC predicts that 45% of the advances in the automotive industry by 2025 will come from software development.
In order to achieve full autonomous driving without driver intervention, cars need to adapt and integrate advanced technologies. Even so, numerous driver assistance systems can be found, which are a step towards the solutions of the future.
Autonomous cars as driver aids
Autonomous cars can simulate humans in terms of driving, controlling and operating the car because they have the necessary computer systems to do so. Although these vehicles do not require a driver to drive them, for the time being it is a concept that in practice has been divided into several levels that offer different types of autonomy depending on the evolution of technology and the necessary regulations.
In order for a car to be autonomous, and have a positive impact on road safety for the driver, occupants and other road users, the implemented technologies must perceive the surrounding environment through analysis. In this respect, the driver can choose the destination, but there is no need to control the car.
Which technologies are going to change the way we drive?
For a car to be considered autonomous, it needs to understand the environment in which it is moving. In this context, advanced technologies such as camera-based computer vision and global positioning systems are implemented. These developments involve significant information gathering from its surroundings, mainly to identify routes, road signs, other vehicles, pedestrians or any kind of obstacle.
Such cars can also receive information from the outside through crash avoidance support systems. This technology analyses distance and regulates speed using sensors, minimising the risk of a crash. For example, the adaptive cruise control system intervenes in the vehicle to avoid crashes; this installation aims to control the constant distance to the front and rear vehicle.
Another crucial technology for autonomous cars is traffic sign detectors, a breakthrough developed by synchronising the car’s front camera and sensors. In this way, the car is able to adapt its driving behaviour to the signals it encounters in the environment.
To this end, vehicles must incorporate global positioning systems such as GPS, computer vision through cameras and connectivity through sensors to be able to exchange information with other elements.
Levels of autonomous driving
There are six levels of autonomy depending on the technology used. At the first level is 0: the vehicle requires a human for full control and has no tools in this regard. Level 1 features the Advanced Driving Assistance Systems (ADAS) for driving assistance: autonomous braking, cruise and stability control. Level 2 includes motion control with lane keeping and control functions and a partial level of automation.
Level 3 has a higher level of autonomy than the previous levels, but still requires driver intervention. It is scheduled to enter into force in 2023. It is levels 4 and 5 that are considered highly and fully automated, where human intervention will not be necessary.
Road safety benefits of autonomous cars
For years, industry experts have been debating the benefits, drawbacks and concerns of autonomous cars. All these issues are related to driver responsibility and road safety. 94% of the world’s road accidents are caused by human error, according to a study published in 2020 by the Insurance Institute for Highway Safety (IIHS). Autonomous cars are presented as the most innovative technology capable of correcting this situation. The same publication points out that these cars can only avoid about one third of accidents due to human error. Although a self-driving vehicle can react faster than a person, technologies do not always respond instantly.
However, autonomous cars and the future of road safety go hand in hand as the technology helps reduce accidents due to perception errors, such as distractions or poor visibility, and minimises prediction, planning and decision-making problems for drivers.
Artificial Intelligence as the basis for the evolution of autonomous driving
As the experts in the IIHS report point out, the number of accidents avoided will depend on how the software of automated vehicles is programmed. In this context, they allude to incorporating artificial intelligence that allows for a more instantaneous reaction to dangers.
An example of successful incorporation would be, for example, installing AI sensors to avoid breaking traffic rules, which the study identifies as the cause of 38% of accidents. This would prevent around 72% of accidents.
Advances in AI have been key to strengthening vehicle autonomy. Thanks to deep learning innovation, in particular pedestrian detection systems, the error rate of detection systems has been reduced by up to 100 times.
Perception and optical recognition are the areas that have undergone the greatest transformation in this automotive context, but AI is also strengthening predictive systems, mapping or simulation.
The main challenge for this industry is to incorporate autonomous driving in all possible contexts, be it heavy rain, snowfall or hail. In this respect, AI comes to the fore as the main technological support for autonomy to establish itself and achieve new goals.