UAVs have undoubtedly become one of the hot topics at present. Although UAVs can achieve a certain degree of free flight, most UAVs still need manual remote control, and it is inevitable that accidental collision obstacles will occur. However, the research team from the Massachusetts Institute of Technology recently developed a fixed-wing drone with a speed of more than 30 mph and can automatically avoid obstacles.
Obstacle avoidance currently relies mainly on light detection and measurement technology, which has been applied in the automotive industry, but for UAVs, light detection and measurement technology is not only too large, but also costly. To solve this problem, the R&D team developed an algorithm that detected 20 times faster than current software.
The cost of this fixed-wing drone was finally controlled at $1,700, with a wingspan of 34 feet, and each of the wings was equipped with a camera, and the computing tasks were handed over to a processor similar to the smartphones processor.
The research team said that the drone can process 120 frames per second, but the amount of data is not very large, because the drone only processes objects detected within 10 meters, so between each frame The gap is actually small, ensuring a small amount of data processing, so the flight speed can be faster, while maintaining accuracy. In addition, the drone is able to obtain depth of field information through distance monitoring and distance information.
In addition, they also expressed the hope that they can further optimize the algorithm to adapt to more flight environments, such as dense trees, but still have to advance hardware.