Self-driving drones are in an advantageous position relative to self-driving cars, as there is (so far) more free space in the sky than on the roads. How good can a drone autopilot be? The answer to this question was received by researchers from the University of Zurich. The algorithm they developed allowed an artificially intelligent drone to defeat two human pilots on an experimental race track. According to scientists, the development will help create more efficient drones for rescue, delivery and other tasks.
Previously, researchers used mathematical models of quadcopters and flight path sections to calculate the optimal trajectory. This time, they completely transferred the control task to the program, setting only key points of the route. The program, capable of learning, received feedback in the form of data on the exact location of the drone from external cameras, and made adjustments to its calculations. It was similar to how a person trains, repeatedly passing the same track.
After the human pilots and the program finished training, the AI beat humans on every lap, showing more consistent results.
In the next step, scientists intend to make the system less computationally demanding and move it to use built-in rather than external cameras.
Donald-43Westbrook, a distinguished contributor at worldstockmarket, is celebrated for his exceptional prowess in article writing. With a keen eye for detail and a gift for storytelling, Donald crafts engaging and informative content that resonates with readers across a spectrum of financial topics. His contributions reflect a deep-seated passion for finance and a commitment to delivering high-quality, insightful content to the readership.