Autonomous vehicles (AVs) have made driving easier than ever. However, it causes problems while driving at night. This obstacle is stopping the automakers to go with a full autonomy design.
The pedestrian fatalities are increasing at an alarming rate at night when the sun is down, confirmed by the Insurance Institute for Highway Safety. It is surprising that around three-quarters of deaths are happening in the dark hours. You can take the example of Uber crash in Arizona. Therefore, you cannot dream of a fully autonomous vehicle without resolving this serious issue.
Role of Current Sensors
The problem with the fully automated vehicles is that these models lack required sensing technologies to deal with every scenario. Many of them use radar, LiDAR, and cameras. But these sensors have perception problems. This issue limits the autonomy and needs a human driver to take control whenever required. These sensors are not that effective at night.
You can take the example of radar. It is able to detect any object at a long range. But it cannot identify the objects clearly. Cameras are appreciated for accuracy. It can identify objects only at a close distance. Many automakers consider the combination of both to identify the object and detect them from a long range.
Some vehicles are also equipped with LiDAR sensors. Like the previous one, LiDAR can detect the objects from the long distance by sending out the signals and then using the reflections to evaluate the distance. But the difference is that radar uses signals and LiDAr uses light waves or lasers.
When LiDAR is able to cover wide range than the radar, radar is more capable to get the adequate detection with the right coverage. In the current conditions, these options are expensive. A few companies want to cut down the cost. For this, they will have to use the lower-resolution sensors. These sensors cannot offer the coverage required for the Level-5 autonomy.
What Are The Challenges of Night Driving?
To support the night driving, many automakers are using multiple sensor types. They believe that a great combination will be able to compensate for the weakness of a single sensor. They are trying to create a redundant network to deal with sensing problems. This combination will help to cover up the weakness of one another. However, all these combinations might not serve the purpose.
By combing all these, the vehicles cannot get Level-5 autonomy. The problem is with classification. All the sensors might be able to detect the data of the surrounding, but they will not be that effective to classify the objects. This is more important while driving at night. Without proper classification, it can be serious and dangerous.
Thermal Camera is The Only Solution
You can use a new type of sensor known as thermal sensors to get the best protection while driving at night. It uses far infrared technology (FIR) to get a reliable and complete coverage to make your AVs functional and safe both in the day and night time.
Unlike other sensors, FIR cameras are able to sense signals from the objects through the radiating heat. This passive technology scans the infrared spectrum to produce a new layer of information while detecting the objects that will not be perceptible to other sensors like radar, camera, and LiDAR.
In addition to the temperature, the camera can capture the emissivity of the object. All the objects have a different emissivity that enables the thermal cameras to sense different objects. More importantly, a thermal camera can immediately detect and classify the objects.
By using this information, the camera will create a visual painting of the surroundings both the far and near range. Another benefit is that it can detect lane marking even the position of the pedestrians. It can determine the pedestrian is about to cross the road or going off the sidewalk. The vehicle will use this information to avoid an accident. Thermal cameras can help you driving at night in both urban and rural areas regardless of environment type.
AdaSKy is the company that develops infrared technology for AVs. This Israeli startup company has recently developed a high-resolution thermal camera, viper. Viper can passively collect the signals and convert them into the high-resolution VGA video. In addition, it offers deep learning algorithms to analyze the surroundings.
from Ageloire Agency http://www.ageloire.com/autonomous-vehicles-self-driving-car-technology/