The revolution of autonomous vehicles with simulation
The revolution of autonomous vehicles with simulation
Just as unmanned aerial vehicles (UAVs) are expected to transform logistics, with companies like Amazon investing a considerable amount of their budget in the use of drones during the delivery process, unmanned ground vehicles, also known as autonomous vehicles, promise revolutionize the automotive market. “
“Will you embrace AI fast enough”, by A.T. Kearney, innovation and intelligence systems have already changed the way the industry operates and seeks investments. In general, to ensure a good investment, it is necessary to seek strategies centered on artificial intelligence (A.I.). The industry invested more than $ 2.5 billion in artificial intelligence between 2014 and 2017. Autonomous car technologies received the largest percentage of that capital, around 20%. In addition, the analysis revealed some factors that have stimulated the development of A.I., especially in the area of machine learning. Amazon, Apple, Facebook, Alphabet, IBM, Microsoft and other major players in the digital age see autonomous vehicles and artificial intelligence as opportunities to transform not only the technology sector, but a good portion of the economy. In late 2017, the company Lyft fell on the radar of Google's holding company, Alphabet, and can now receive more than 1 billion for projects related to autonomous cars. This contribution will give impetus to a direct competition against the giant Uber, rival of Lyft and Waymo, Alphabet's, autonomous vehicle division.
And security? For the general public, autonomous vehicles can cause a feeling of insecurity due to the feeling of lack of command of the situation, but researches even point out that about 90% of deaths in traffic accidents are caused by human error, either by distraction, influence of alcoholic and nootropic substances or imprudence.
According to the World Health Organization, about 1.2 million people die in car accidents every year. Autonomous vehicles rely on deep learning algorithms that recognize patterns and respond like humans. A vehicle must perceive its surroundings, determine the necessary action and perform the necessary movement within an ever-changing environment. And one of the main responsible for autonomous vehicle operations is, in a way, automotive radars.
The simulation architecture of an autonomous vehicle is very complex. Requires integration of physics, embedded systems, software development and code generation for an accurate simulation. Autonomous vehicles require the continuous evolution of the vehicle's sensors - the eyes and ears of the control system that perceive the vehicle's operational characteristics and the environment around it.
Four major classes of vehicle sensors provide most of the environmental sensory data for an autonomous vehicle - visual spectrum cameras, laser devices (LIDARs), ultrasound sensors and radio frequency sensors (radars). Automotive radar employs millimeter frequencies to detect obstacles and long-range objects, as well as to track the speed and direction of various scene actors, such as pedestrians, other vehicles, guardrails, etc.
Three main categories of radar systems are typically employed in automotive security systems: • Short-range radar (SRR) for collision warning and safety proximity; • Medium range radar (MRR) to observe the vehicle's corners and perform blind spot detection; • Long-range radar (LRR) for sensors aimed at early collision control and other detection functions. But how to ensure safety in autonomous vehicles? How to subject an autonomous vehicle to all the situations and adversities that it may encounter on a daily basis in order to ensure that it responds properly? How to trust only radars and sensors for car navigation? No vehicle can reach the market without exhaustive tests simulating real situations. The only viable answer is computer simulation.
Building, testing and validating ideas virtually before testing on physical prototypes is a matter of economy and safety. In addition to the flexibility to understand how small ideas can change a whole in a short time, simulation is an easy and accurate way to accelerate the product development line by reducing risks and ensuring the best possible use of time and resources.
In Brazil, one of the companies that is at the forefront of research and development of simulation for automotive cars is ESSS, a multinational based in Florianópolis (SC) that develops and sells simulation software throughout Latin America and Iberia.
“Although in Europe some companies are already undergoing physical tests on the roads, such as Google, Tesla and Uber, here in Brazil the development is still timid. The major players and companies that form part of the productive chain of the Brazilian automotive sector are already responding positively to the creation of new techniques and parts to meet this new trend. And simulation is the most viable way, as it allows several tests to be carried out without the need to place a vehicle on the street, which, in addition to bringing more safety, is also the most economical solution for the companies' R&D sector ”- says Juliano Mologni, an engineer specializing in electromagnetic simulation at ESSS.
Through computer simulation, engineers can:
• Test the prototype virtually and adjust the antenna topologies quickly, without requiring the manufacture of the product;
• Test antennas effectively to understand their behavior under a variety of structural and environmental conditions;
• Optimize elements and matrices with the least effort and cost.
• Build only a single prototype to test at the end.
Automotive radar systems play a central role in security systems and must be tested with vehicle control systems and algorithms to validate safe operation.
It may take a while for autonomous vehicles to become a reality for the final consumer in Brazil. In Europe this is already ahead, with physical tests taking place on the streets and roads of the old continent. But for national industries, preparing now is to stay ahead of a market that has everything to be very promising.