The cyclist crept imperceptibly Bashny.Net

Autonomous cars on the road — a unique achievement progress! Moreover, the developers have prevented the danger of collisions with other cars, pedestrians, even squirrels and birds, which is especially important for American cities. But still formed a group of vehicles that were left vulnerable due to their criteria. This bike is quiet, light and agile movement.

The existing system of self-driving mediocre job of detection close to themselves, but also hard to predict their actions. Although cyclists on the roads are relatively few, but do not ignore this problem.
 



According to some researchers, today, unmanned vehicles are the most difficult to recognize on the road cyclists. This is due to their uncharacteristic speed, small size and dissimilarity with each other. Cars are a lot more alike than these two-wheeled road users.

If you walk around any city and look closely at the bikes, their diversity becomes apparent. Here is a young man in a green sport, and that drove a girl urban pink ladies bike. So this woman goes to the country, by the iron horse in addition to the hanging baskets for another 2-3 package with seedlings. My mother carries a child in belokrilie on the trunk, but this father takes daughter in vilacrosse, but installed right behind the wheel.

Any drone in the headlights will sarabit from this palette and cultural diversity!



Electronic driver collects all the necessary information about cyclists using three devices: a camera, radar and laser locator. Then he estimates the parameters of an object using previously obtained data on these participants. To autopilot definitely know about the intentions of the cyclist detail to measure the dimensions of the head and hands, the distance between them, the distance from the head of the cyclist to the pavement and the angle of the bend of the elbow. If successful, the perspective of the on-Board computer is able to determine and model of bike.

It is now possible to trace the relationship between the problem of detecting cyclists and creating new generation vehicles. Scientists are trained system of autopilots for a variety of vehicles, which show different cars, but almost no bicycles. Therein lies the flaw. Not all unmanned machines learn to recognize bicycles, and therefore they have difficulties with their detection on the roads. Fortunately, this problem is temporary.

Recently in the automotive industry began to use an algorithm Deep3DBox. It owes its origin to researchers from George Mason University in tandem with the company to develop unmanned taxi Zoox. In test mode, two-dimensional images, the algorithm is able to identify 89% of the cars. Equally important, he knows how to predict which way heading other vehicles. Moreover, the algorithm even draw around each object a zone is like a box. Today Deep3DBox is considered one of the best algorithms for recognition of the vehicles. But, unfortunately, in a comparative analysis he is able to see only 74% of bicycles, and only 59% of them guess for further trajectory.



Note that this is the result of one of the best algorithms, to say nothing of the less sophisticated designs. Very sad situation for cyclists: for unmanned vehicles they are likely to look like an unpredictable moving unidentified object. And this contributes not only to "neobychainoi" algorithms, but the compactness and delicacy of bikes compared to cars. Because of this, in recent years, automakers had to increase the scanning frequency of the road with radars and lidars to allow the system to confidently detect bicycles, to determine the distance to them and their trajectory. Together with the increase of datasets for learning algorithms that allows to increase the proportion of correct recognitions.

To save the lives and health of cyclists is called for a unique design — 3D maps ultra-high detail. The computer sees all the objects on the road and near her, line marking and traffic signs. Such awareness will help to recognize and map the bike.

Perhaps, then, on the other begin to treat the point of view according to which the computer is driving better inconsiderate and always hurrying people. Many cyclists get into accidents because of the carelessness of the drivers. So for someone the appearance of the car, managed not scattered and tired people, and unfailingly attentive computers that still besides learning to recognize gestures cyclists — a real golden age!

For example, today unmanned taxi Google equipped with sensors that allow you to discern and to decipher the intentions of cyclists. When turning cyclists must submit the necessary signs according to the rules of the road. Drones, Google has learned to read and recognize these signs in order to properly maneuver or change speed. Also, robotaxi able to see two-wheeled road users in the dark.

But there are difficulties posed by cyclists. As the developers Deep3DBox, to anticipate the maneuvers of a car is much easier than the bike. The vast majority of car drivers are more predictable, while the bike owners usually have a very vague idea about the rules of movement, and each second the cyclist behaves on the road as impulsive and charismatic personality. If this person decides to suddenly appear out of nowhere among the stream of cars, even the smartest algorithm may not be able to react.



This characteristic unpredictability is today taken into account by the developers of the algorithms to detect bicycles. For example, today cyclists less afraid of drones Jaguar, equipped with a detection system of the cyclists. And Volvo in 2013 he presented the technology of emergency braking when it detects a cyclist (AEB) in front of the car. The system will stop the machine before the driver to avoid a tragic accident. However, AEB is not perfect: the algorithms it is difficult to predict the trajectory of the cyclist. So it is highly recommended not to use this feature Volvo on the streets of Beijing: the daily flow in 9 million bicycles can cause damage to the electronic brain.

Moreover, cyclists can cool to make fun of electronic drivers. A couple of years ago in Austin there was an incident with a cyclist and unmanned taxi Google. The car was done by a man confused by surplas. The cyclist had to stop the car without removing the foot from the pedal action. The car moved, but noticing that the bike has moved 1 inch, and braked sharply. In an experiment, the Texan repeated the trick several times, much to the amused passengers by taxi. And now imagine such a car on the streets of Russian cities. Knowing the propensity of our citizens to humor and fun, not hard to guess that far the drones have not gone.

Diversity is a feature of the modern world, which carries over into the road. In many cities, more and more citizens choose the Bicycle. And if the cyclists of Sweden for many years angrily waving careless drivers, in connection with the advent of driverless cars policy should act in a different way. For example, in the United States is in full swing testing of an unmanned taxi service Uber. Naturally, the transition to this level of service has significant benefits for the customer and for the company. A taxi ride without the driver needs to be cheaper, the computer is not rude to customers and not to violate traffic rules, keeps to work from morning to evening etc the Most Uber, of course, easier to work with machines whose actions can be improved and controlled. In some countries, have been quite a crime with the participation of the company's drivers, as they do not undergo a special selection process when applying for a job.

While there is testing, travel and drones Uber is completely free. As for the rides, you can sit back and recorder to record the entire route. So far, the machines are not perfect, but Uber does not get upset about this, considering such violations of working time. Carelessness or is it a constant innovation — this question provokes many debates between supporters of the newest technologies and a conservative.

In San Francisco Uber one drone was caught passing a red light, while a few others unfolds on bike paths. But in that moment there could pass people.



A number of experts condemn the zeal with which Uber aspires to market an unmanned taxi. It's hard to argue with the assertion that such a crude technology must first be carefully tested before to carry passengers. It is impossible to do business on the technologies that have such a high probability of failure. Let's hope that other companies will be more responsible approach to the introduction of drones. In the meantime, cyclists, beware! published

 

P. S. And remember, only by changing their consumption — together we change the world! ©

Source: geektimes.ru/company/asus/blog/287108/