Jinhua Cai

Dr. Jinhua Cai


Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 September 2025 in Sensors
Verified
Journal Article
Open access
PDF file
JCJinhua Cai
Jinhua Cai
HDHongchang Ding
Hongchang Ding
PWPing Wang
Ping Wang
XGXiaoqiang Guo
Xiaoqiang Guo
HHHan Hou
Han Hou
XQXiaoli Qiao
Xiaoli Qiao
In response to the growing demand for automation in aerospace harness manufacturing, this study proposes a digital twin-based intelligent monitoring system for robotic wiring operations. The system integrates a seven-degree-of-freedom robotic platform with an adaptive servo gripper and employs a five-dimensional digital twin framework to synchronize physical and virtual entities. Key innovations include a coordinated motion model for minimizing joint displacement, a particle-swarm-optimized backpropagation neural network (PSO-BPNN) for adaptive gripping based on wire characteristics, and a virtual–physical closed-loop interaction strategy covering the entire wiring process. Methodologically, the system enables motion planning, quality prediction, and remote monitoring through Unity3D visualization, SQL-driven data processing, and real-time mapping. The experimental results demonstrate that the system can stably and efficiently complete complex wiring tasks with 1:1 trajectory reproduction. Moreover, the PSO-BPNN model significantly reduces prediction error compared to standard BPNN methods. The results confirm the system’s capability to ensure precise wire placement, enhance operational efficiency, and reduce error risks. This work offers a practical and intelligent solution for aerospace harness production and shows strong potential for extension to multi-robot collaboration and full production line scheduling.

17 July 2026 in Sensors
Verified
Journal Article
Open access
PDF file
XGXiaoqiang Guo
Xiaoqiang Guo
HDHongchang Ding
Hongchang Ding
HDHongchang Ding
Hongchang Ding
HHHan Hou
Han Hou
JCJinhua Cai
Jinhua Cai
Force control is crucial for robotic contact-rich tasks such as assembly, grinding, and polishing, directly affecting task accuracy, interaction stability, and safety. Yet in unstructured environments, environmental uncertainty, contact oscillations, and friction disturbances make high-performance contact control and reinforcement learning policy optimization difficult. To address this issue, this paper proposes a deep reinforcement learning-based variable-damping impedance control method that integrates parameterized Hindsight Experience Replay (TO-HER) and Importance Sampling (IS). Within the impedance control framework, a residual parameterized policy enables online damping adjustment, improving dynamic adaptability across contact phases. To overcome the training instability of conventional HER in contact-intensive tasks, a parameterized goal relabeling mechanism is introduced to improve relabeled sample quality and sample efficiency. In addition, an importance sampling scheme based on density ratio estimation mitigates the distribution mismatch between relabeled and real samples, enhancing training quality and stability. Experimental results show that the proposed method outperforms baseline methods in convergence, final return, and training stability, while achieving higher tracking accuracy and better dynamic response in typical contact scenarios, demonstrating strong effectiveness and robustness for robotic contact control in unstructured environments.

9 September 2026 in Sensors
Verified
Journal Article
Open access
PDF file
JCJinhua Cai
Jinhua Cai
TJTao Jiang
Tao Jiang
HHHan Hou
Han Hou
PGPeng Gao
Peng Gao
NLNa Liu
Na Liu
Aircraft wire-harness routing on fixture boards requires not only efficient large-workspace robot motion but also smooth and process-safe passage around densely distributed pins. Conventional robot trajectory optimization generally focuses on generic motion indices and cannot explicitly represent local over-bending risk and coupled wiring-process constraints. This paper proposes a process-safety-aware multi-objective trajectory optimization method for a seven-axis wiring robot composed of a linear rail and a six-DOF manipulator. An executable reference trajectory is parameterized using quintic B-splines, and execution time, normalized joint jerk, and a pin-neighborhood weighted bending-safety proxy are jointly optimized under robot-motion, pin-passing, board-clearance, and non-target-pin constraints. A constraint-guided adaptive DE-NSGA-II algorithm is developed to improve feasible-solution search in the resulting narrow and strongly coupled feasible region. Physical experiments show that, compared with MOPSO, the proposed method reduces normalized joint jerk by 35.8% and the bending-safety proxy by 77.0%, while completing all ten repeated trials without controller alarms. Vision-based measurements further show a strong correlation between the proposed proxy and the actual harness bending radius (R2=0.991). Additional routing layouts confirm the applicability of the method to more complex fixture-board configurations. These results demonstrate that the proposed method provides a practical balance among routing efficiency, trajectory smoothness, and process reliability.