HCHao Chen

Hao Chen

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Hao Chen has a total of 8 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    RL-accelerated OSQP for real-time asymmetric corotational beam contact simulation in continuum robots

    by Hao Chen, Jian Chen, Shuai Wang, Mingcong Chen, Zhongkai Zhang, Hongbin Liu

  • Open Access

    RL-accelerated OSQP for Real-time Asymmetric Corotational Beam Contact Simulation in Continuum Robots

    by Hao Chen, Jian Chen, Shuai Wang, Mingcong Chen, Zhongkai Zhang, Hongbin Liu

    Real-time simulation of continuum robots within narrow anatomical cavities is crucial for navigating minimally invasive surgeries, yet it remains computationally challenging due to extensive beam-surface multi-point contacts and significant bending deformations. While the corotational finite element method (FEM) effectively captures large nonlinear displacements, its inherent asymmetric stiffness matrix severely impairs the convergence of standard optimization algorithms. To address these challenges, this paper proposes a consistent, implicit 3D corotational beam formulation tailored for real-time dynamic contact simulations. The HHT-α method is employed to predict motion trajectories, effectively mitigating spurious contacts induced by non-convex geometries. The underlying contact problem is formulated into a variational inequality (VI) and resolved using a novel adaptation of the Operator Splitting Quadratic Program (OSQP) algorithm, specifically modified for asymmetric systems. Additionally, reinforcement learning (RL) is integrated to dynamically update the OSQP parameters, significantly accelerating algorithmic convergence. By parallelizing the parameter updates and collision detection on a graphics processing unit (GPU), the proposed framework achieves robust, high-fidelity, real-time simulation of continuum robots in complex environments, such as pulmonary bronchi. Experimental results demonstrate that the proposed algorithm achieves real-time simulation performance without significantly compromising accuracy, reaching 285 FPS in a 200-node simulation of a transbronchial continuum robot.

  • A Reinforcement Learning Based FEM Solver for Accelerating Contact-Influenced Simulation of Continuum Robots

    by Hao Chen, Jian Chen, Zhongkai Zhang, Hongbin Liu

    Continuum robots exhibit exceptional flexibility and multi-degree-of-freedom maneuverability, offering significant advantages for navigating confined luminal spaces. However, rapid simulation of their contact-influenced behavior remains challenging. This paper presents RLFEM, an innovative finite element method (FEM) featuring a reinforcement learning-enhanced quadratic programming solver, specifically designed for efficient continuum robot simulation updates. To address complex robot dynamics, we employ a quasi-static FEM formulation. Our core contribution integrates an accelerated solver scheme within this framework, leveraging reinforcement learning to rapidly resolve FEM contact problems without accuracy compromise. RLFEM exhibits significant gains in runtime performance for continuum robot simulations, as evidenced by numerical experiments showing a 16.20-fold acceleration compared to the baseline at the 150-node discretization level.


  • Multi-Interface Strain Transfer Modeling for Flexible Endoscope Shape Sensing

    by Xinran Liu, Jian Chen, Jian Hu, Hao Chen, Yuanrui Huang, Hongbin Liu

    Robot-assisted minimally invasive surgery (MIS) using flexible endoscopy has emerged as a groundbreaking technology for improving traditional surgical approaches. However, a major challenge in advancing this technology is the lack of shape sensing, which leads to inaccurate navigation and control of ultra-long flexible endoscopy within the narrow and tortuous lumen environment. The unique characteristics of flexible endoscopy, including large slenderness ratio, high bending angles, and non-symmetric, multi-channel configurations, pose significant challenges to accurate shape sensing. To address this challenge, we propose a novel shape-sensing scheme based on distributed fiber optic strain measurement, which incorporates a complete and applicable multi-interface strain transfer model adapted to large deformation and multiple sensing points. To validate the theoretical model, a shape sensor with a diameter of 2.84 mm and a length of 500 mm is fabricated. Both 2D and 3D shape sensing experiments are conducted on predefined templates, and the results highlight a significant improvement in the precision of sensor measurements through the utilization of the proposed model. Specifically, the 3D experiment results show a mean absolute error (MAE) of 5.38 mm for complex geometrical shapes and the proposed model reduces the MAE by approximately 52.9% compared to the unmodified case.