Enhancing the Naturalness of Game Character Movements Through the Fusion of Skeletal Rigging and Physics Engine Technology

Abstract

Traditional animation techniques often produce stiff character movements and insufficient physical feedback, especially when simulating virtual fabrics, clothing, and flexible structures. This limitation affects not only visual realism but also the digital representation of wearable antenna substrates, flexible electromagnetic components, and human-body interaction scenarios in electromagnetic wave applications. To address this problem, this paper proposes a character movement driving system based on the fusion of skeletal rigging and physics engine technology. Dynamic constraints are established by setting rotation, elasticity, and damping parameters for bone nodes and combining them with the Inverse Kinematics (IK) algorithm. The physics engine calculates external forces and inertial feedback in real time and transfers the results to the skeletal system for adaptive motion adjustment. A Proportional-Derivative (PD) controller is introduced to coordinate dynamic balance between the animation layer and the physics layer, while a lightweight neural network prediction module is used for posture optimization. In addition, a multi-threaded rendering synchronization mechanism is designed to integrate movement, physical response, and visual output. Experimental results show that the proposed method keeps the standard deviation of joint angular acceleration below 1000 rad/s2, controls the horizontal centre-of-gravity offset within 5.2 cm in collision scenarios, reduces the average action response delay to 35 ms under thrust, and achieves a subjective naturalness score of 7.2–8.5. The method improves the smoothness, stability, and real -time expressiveness of character movement, and provides a simulation reference for flexible textiles and wearable electromagnetic structures.