Rise of the Machine 5 Fingers (RISE)

10 sites
While there is no single scientific project with that exact title, your description aligns with several cutting-edge concepts in advanced robotics and computational geometry. These "hyper-algorithms" and "complex geometrics" are currently being explored to give five-fingered robots human-like dexterity.
Five-Fingered Robotics and Complex Geometry
Modern five-fingered robotic hands, such as the DG-5F-S, use 20 degrees of freedom (DoF) to replicate natural human motion.
Geometric Linkage Systems: To achieve this, robots use series-parallel hybrid designs. These systems use a series of links, cranks, and gears to translate simple actuator movements into complex, coordinated finger motions.
Fibonacci-Based Models: Some researchers use specific mathematical sequences, like the Fibonacci sequence, to model the phalanx lengths and rotation of robot fingers, treating them as hyper-redundant planar manipulators.
"Hyper-Algorithms" for Dexterity
The "hyper-algorithms" you mentioned likely refer to Model Predictive Control (MPC) and Reinforcement Learning, which are the standard for managing "hyper-complex" manipulation tasks.
Real-Time Planning: Frameworks like UniGrasp use deep neural networks to select contact points from an object's geometry, allowing various
-fingered hands to grasp unknown items.
Finger Flow: Systems like "Finger Flow" run parallel threads at high frequencies (e.g., 3 kHz) to generate reactive grasping motions that avoid collisions while approaching objects.
Hyperparameter Adaptation: To handle complex object shapes, robots can dynamically adapt hyperparameters in their modeling algorithms based on the local curvature of the surface they are touching.

https://youtu.be/oLLjHpBQ7mI?si=8SSOMjjWwtHcmkgy