There is a crossroads in the automobile industry. On the one hand, driving enthusiasts want models with higher performance, rigid chassis and direct steering that maximize driving fun. On the other hand, pragmatists say that driving is an annoying procedure, in which you have to avoid traffic jams and look for parking spaces.
It is this last group that dreams of freeing up the hours behind the wheel through robotic driving and being able to allocate them to productivity, sleep or, according to a study by the English publication ‘Annals of Tourism Research’, sex. “It is likely that the autonomous car will end up linked to prostitution in the future, whether legal or illegal,” they say. But, for these situations to exist, two conditions are necessary for driverless cars to become a reality on our roads: legislation that allows it and vehicles with a satisfactory response capacity in daily traffic. For the R&D director of
Cesvimap,
Rodrigo Encinar Martin«There is still time to see autonomous cars in a widespread way, at least until 2050».
His is the autonomous car project that his organization, the
Mapfre Accident and Road Safety Experimentation Center, presented in Ávila last Tuesday. To replace the driver, every car needs to be able to accurately detect its surroundings and make decisions in a matter of hundredths of a second, without compromising the safety of its occupants or those around them. For this, the three essential elements are the sensors, a powerful microprocessor and the computer commands that determine what to do in each situation.
This last batch was developed by the team from the Carlos III University (UC3M), which, together with another from the Polytechnic University of Madrid (UPM), were in charge of unifying the physical sensors. Of the latter, the most important —and one of the most limiting factors for the development of the autonomous car— is lidar: a kind of sophisticated radar that works with lasers instead of electromagnetic waves. The two main problems that the use of lidar presents is that they are fragile systems when faced with impacts and with a prohibitive cost for mass production, around 4,000 euros per unit.
However, the robotic driving models will begin to arrive gradually: Mapfre has already ensured several autonomous buses that circulate – always with the same route – through the Timanfaya National Park (Canary Islands). According to
Jose Maria Cancerthe general director of Cesvimap, small autonomous vehicles will soon arrive to carry out home deliveries.
Autonomous in Avila
The self-driving powerhouses are Waymo — which has the computational artillery of its parent company, Google — and the Volkswagen Group, which aims to have an operational service in Hamburg by the middle of the decade. Both operations have in common that they have a commercial objective once their product is ready, as well as that they have multi-million dollar budgets. This is not the case for the Cesvimap car.
Conceived as “a platform to test the latest advances in sensors and computing”, the Cesvimap vehicle is comparatively more rustic than those of Google and Volkswagen, but it reaches autonomy level 4, just like them, in a closed and controlled environment. Where driverless cars work best are in closed environments, because they are easier to map and because they do not have the difficult to calculate variable of other vehicles. For this reason, the section chosen in Ávila was about 300 meters long, with a roundabout at each end.
Cesvimap’s vehicle used its front camera to detect the straight line of the street and its position in it, while when facing the circular shape of the roundabout it needed to give way to GPS positioning -with an accuracy of two centimeters, instead of two meters, like the one available on phones—to calculate your position and trajectory. The UPM and UC3M programmers had to manually enter the position of a zebra crossing along the route —their sensors are not capable of detecting traffic signals— and the lidars are responsible for identifying if there is a pedestrian waiting for the crossing.
Calibrating the sensitivity of the lidar is tricky, as being too strict can cause the car to brake on a small object, like a pigeon; while being too lax would make it miss pedestrians. As for driving, this, without being as fluid as that of a human, does allow attention to be devoted to matters other than the road, which is the ultimate goal of autonomous driving.
Now, it only remains to implement these vehicles in a situation of real coexistence with others, even with people behind the wheel, but the technology is still incipient and the legislation will not allow it for at least two more decades.
#Driving #hands #Spain #longer #science #fiction
