"Aiming" to dramatically improve his score in Aimlabs' aim training programme, a YouTuber built a physical aim-assist wearable exoskeleton. It uses an attached camera, a built-in Nvidia Jetson computer system, and a series of motors and servos to physically move his arm to improve his aim when he's off target, and the results are impressive.
Aimbots have been a problem in FPS gaming since the multiplayer competitive scene first emerged in the 1990s, but that usually involves exclusively software correction. It's why there's such a cat and mouse game between exploiters looking to cheat and developers trying to prevent it.
The physical device itself is a clever creation, blending 3D-printed components with hinges, that hook on to his forearm, but connect to individual fingers and can adjust the position of the hand, wrist, and all connected digits. The controller tracks the positions of all these individual components of his body and translates that into something the computer can track and understand, and then alter if it needs to.
It was a weeks-long project. There were many hiccups and stumbling blocks along the way, but ultimately Zetta was able to get it working.
Intriguingly, though, there was a learning curve. The first result after testing it in a fully-functioning state saw his score drop by close to 20% over native, augmentation-free gaming. That's because he had to learn to let the assistant do its job and relax his wrist enough to allow the corrective movements to take place.
After that, he managed a 3% improvement over his high-score. That might sound close to margin of error, but Zetta had dialed in his original score before starting, so this jump was noteworthy.
Could it be better, though? It turns out it could. By focusing on tightening up the latency of the camera and image translation part of the process, he brought the reaction time down from around 50 milliseconds camera to arm, to just 17ms!
He also boosted the voltage on the motors to make them stronger and able to maneuver him and his fingers even if he resisted slightly, and the results were really impressive. His first run saw a 12% improvement, then a 28% improvement on another, then 43%, then a 63% boost - giving him the second-place spot on the global leader board.
Sure, it's cheating, but it's an impressive project nonetheless. As he suggests in its conclusion, combined with some neuromuscular elements, this could get even better. That's almost certainly his next project.
Follow Tom's Hardware on Google News, or add us as a preferred source, to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button!
