Graph Neural Network‐Based Reinforcement Learning for Decentralized Multi‐Robot Manipulation
Robot arms lifting a large object face a trade‐off: centralized controllers explode in parameters, while decentralized ones cannot coordinate. A GNN resolves this—each arm runs its own network but acts on the full team state, achieving centralized‐level coordination with decentralized execution. Trained across team sizes, a single policy scales to four‐
Tong Chen +3 more
wiley +1 more source
A reinforcement learning enhanced pseudo-inverse approach to self-collision avoidance of redundant robots. [PDF]
Hong T, Li W, Huang K.
europepmc +1 more source
Clarke Transform: A Lingua Franca for Continuum Robotics
Using the Clarke transform, both robot‐dependent mappings are known exactly for kinematic models, leading to a complete chain of transformations between spaces with associated abstraction. This eases the derivation. Therefore, proposing the Clarke transform as a lingua franca and utilizing a practical fact sheet can pave the way for more systematic ...
Reinhard M. Grassmann, Sami Haddadin
wiley +1 more source
3D Closed‐Loop Navigation of Magnetic Microswimmers Using Vision‐Derived Depth Feedback
A planar magnetic microswimmer is navigated in three dimensions under a standard microscope. Image defocus supplies a real‐time depth cue, and model‐predictive control with A* planning steers the body on surfaces and in free swimming. Axial position is regulated by propulsion alone, without magnetic field gradients or extra cameras.
Yasin Cagatay Duygu +2 more
wiley +1 more source
Interpersonal Coordination between Female Soccer Players: Leader-Follower Roles within a Collision-Avoidance Task. [PDF]
Fernandes CA +6 more
europepmc +1 more source
Shape Sensing, Mapping, and Obstacle Avoidance of the Continuum Manipulator in Confined Environments
An integrated sensing framework combines capacitive sensors, IMU, and laser array sensor. The shape reconstruction is finished by backpropagation neural network. Accurate shape estimation (relative error of 1.04%) is maintained under a 110 g tip load, while a miniaturized VL6180X array enables local environmental reconstruction and reactive obstacle ...
Pengyuan Wang +9 more
wiley +1 more source
Six‐DoF Strawberry Harvesting Pose Selection and Adaptive Obstacle Clearing in Occluded Environments
An occlusion‐aware 6‐DoF harvesting strategy integrates γ‐angle pose selection, ripeness‐weighted risk evaluation, and adaptive obstacle clearing. The approach enables collision‐free grasping orientation selection and improves strawberry harvesting reliability under complex occlusion, achieving higher success rates and reduced fruit damage. In complex,
Tianxiao Zhu +4 more
wiley +1 more source
Multi‐Stage Social Value Orientation‐Based Lane‐Changing Driving for Autonomous Vehicles
A multi‐stage SVO framework is developed to dynamically infer the social preferences of surrounding vehicles from human driving data. By combining offline learning, online SVO estimation, temporal intention tracking, and interaction‐aware planning, the framework enables autonomous vehicles to adapt lane‐changing decisions and trajectories to evolving ...
Nongrui Qin +4 more
wiley +1 more source
Advancing Digital Twin-Based Collision Avoidance: A Comprehensive Analysis of Communication Networks for Safety-Critical Applications in Industry 4.0. [PDF]
Moldovan C +4 more
europepmc +1 more source
Related searches:
Among the many models for flocking systems of interacting particles, the one introduced by Cucker and Smale has attracted attention due to the fact that a convergence to flocking (i.e., to a common velocity) could be established depending on conditions on the initial state of the system.
Felipe Cucker, Jiu-Gang Dong
openaire +2 more sources

