AI-powered and autonomous drones are vulnerable to jamming and spoofing from enemy signals, which can destroy expensive weaponry. Long-range attack drones often depend on unreliable GPS and camera systems. A Brooklyn engineer developed a vision-based Visual Positioning System that uses electro-optical or infrared cameras feeding an onboard NVIDIA edge computer. The system performs real-time terrain matching, celestial navigation, and terminal guidance against pre-loaded maps. Evasive maneuvers are being added to protect against enemy kinetic interceptors. The technology is already operating in Ukraine, with plans to update American drones to reduce costs and improve resilience.
"Long-range Shahed-style attack drones costing $50,000 are built to fly thousands of miles, carrying hefty warheads and complete complicated missions. But, they're often reliant on glitchy and dated GPS and camera systems. While the technology is advancing rapidly, both manned and autonomous drones are extremely vulnerable to getting jammed or spoofed by enemy signals, which can result in the pricy weaponry getting destroyed."
"Streem, is building a smarter camera with guidance and navigation systems in the drone itself. He has also created a vision-based navigation system (Visual Positioning System or VPS) that relies on electro-optical or infrared cameras feeding an onboard NVIDIA edge computer. The system performs real-time terrain matching, celestial navigation, and terminal guidance against pre-loaded maps."
"The company is also adding evasive maneuvers to the VPS to protect against enemy kinetic interceptors. The technology is already working in Ukraine, and Streem believes updating American drones with this kind of equipment will save a fortune for t"
#drone-navigation #anti-jamming-and-anti-spoofing #computer-vision #edge-ai-computing #military-guidance-systems
Read at New York Post
Unable to calculate read time
Collection
[
|
...
]