arXiv subjects

Zeming Wu

Publications and source records attributed to Zeming Wu.

6 recordsLinked to original sources

arXiv2025

Collision-free Control Barrier Functions for General Ellipsoids via Separating Hyperplane

This paper presents a novel collision avoidance method for general ellipsoids based on control barrier functions (CBFs) and separating hyperplanes. First, collision-free conditions for general ellipsoids are analytically derived using the concept of dual cones. These conditions are incorporated into the CBF framework by extending the system dynamics of controlled objects with separating hyperplanes, enabling efficient and reliable collision avoidance. The validity of the proposed collision-free CBFs is rigorously proven, ensuring their effectiveness in enforcing safety constraints. The proposed method requires only single-level optimization, significantly reducing computational time compared to state-of-the-art methods. Numerical simulations and real-world experiments demonstrate the effectiveness and practicality of the proposed algorithm.

cs.RO↗

arXiv2024

GPU-Accelerated Optimization-Based Collision Avoidance

This paper proposes a GPU-accelerated optimization framework for collision avoidance problems where the controlled objects and the obstacles can be modeled as the finite union of convex polyhedra. A novel collision avoidance constraint is proposed based on scale-based collision detection and the strong duality of convex optimization. Under this constraint, the high-dimensional non-convex optimization problems of collision avoidance can be decomposed into several low-dimensional quadratic programmings (QPs) following the paradigm of alternating direction method of multipliers (ADMM). Furthermore, these low-dimensional QPs can be solved parallel with GPUs, significantly reducing computational time. High-fidelity simulations are conducted to validate the proposed method's effectiveness and practicality.

cs.RO↗

arXiv2023

Collective flow and the fluid behavior in p/d/$^3$He+Au collisions at $\sqrt{s_{NN}} = 200$ GeV

By varying the intrinsic initial geometry, the p/d/$^3$He+Au collisions at the Relativistic Heavy Ion Collider (RHIC) provide a unique opportunity to understand the collective behavior and probe the possible sub-nucleon fluctuations in small systems. In this paper, we employ the hybrid model iEBE-VISHNU with Trento initial conditions to study the collective flow and the fluid behavior in p/d/$^3$He+Au collisions. With fine-tuned parameters, iEBE-VISHNU can describe the $v_2(p_T)$ and $v_3(p_T)$ data from the PHENIX and STAR collaborations. However, for some certain parameter sets with initial sub-nucleon fluctuation, the hydrodynamic simulations have already beyond their limits with the average Knudsen number $\langle K_n \rangle$ obviously larger than one. Our calculations demonstrate that, for a meaningful evaluation of the fluid behavior in the small systems, model simulations should also pay attention to the validity range of hydrodynamics.

nucl-th↗

arXiv2018

Universal scaling of the sigma field and net-protons from Langevin dynamics of model A

In this paper, we investigate the Kibble-Zurek scaling of the sigma field and net-protons within the framework of Langevin dynamics of model A. After determining the characteristic scales $\tau_{kz},l_{kz}$ and $\theta_{kz}$ and properly rescaling the traditional cumulants, we construct universal functions for the sigma field and approximate universal functions for net-protons in the critical regime, which are insensitive to the relaxation time and the chosen evolving trajectory. Besides, the oscillating behavior for the higher order cumulants of net-protons near the critical point is also drastically suppressed, which converge into approximate universal curves with these constructed Kibble-Zurek functions.

nucl-th↗

arXiv2018

Applications of deep learning to relativistic hydrodynamics

In this proceeding, we will briefly review our recent progress on implementing deep learning to relativistic hydrodynamics. We will demonstrate that a successfully designed and trained deep neural network, called {\tt stacked U-net}, can capture the main features of the non-linear evolution of hydrodynamics, which could also rapidly predict the final profiles for various testing initial conditions.

nucl-th↗

arXiv2018

Applications of deep learning to relativistic hydrodynamics

Relativistic hydrodynamics is a powerful tool to simulate the evolution of the quark gluon plasma (QGP) in relativistic heavy ion collisions. Using 10000 initial and final profiles generated from 2+1-d relativistic hydrodynamics VISH2+1 with MC-Glauber initial conditions, we train a deep neural network based on stacked U-net, and use it to predict the final profiles associated with various initial conditions, including MC-Glauber, MC-KLN and AMPT and TRENTo. A comparison with the VISH2+1 results shows that the network predictions can nicely capture the magnitude and inhomogeneous structures of the final profiles, and nicely describe the related eccentricity distributions $P(\varepsilon_n)$ (n=2, 3, 4). These results indicate that deep learning technique can capture the main features of the non-linear evolution of hydrodynamics, showing its potential to largely accelerate the event-by-event simulations of relativistic hydrodynamics.

nucl-th↗