Research on the Dragon Dance Formation Based on Spiral Motion Model and Collision Detection

Jiawei Han, Yinuo Su, Ziheng DingHighlights 2025 · DOI

Abstract

This study presents a mathematical model of the motion of a dragon dance team moving along an Archimedean spiral path. The motion of 224 points representing the dragon dance team is described using parametric equations, with the velocity of each point determined through the Runge-Kutta method. The study investigates the dynamics of the dragon head and body segments as they spiral inward, with a focus on maintaining fixed distances between consecutive points. Additionally, collision detection between the dragon's benches is implemented to prevent overlap. The termination time of the motion is determined by identifying the first collision between the dragon head and the body segment. The results demonstrate the application of numerical methods to model complex dynamic systems and ensure smooth movement without collision. The first collision occurs at 432.18 seconds, during the dragon's thirteenth loop.

Introduction

The traditional "bench dragon" dance is a complex multi-agent coordination problem. It involves 224 rigid segments (benches) connected at hinges, navigating an Archimedean spiral. The primary challenge is modeling the motion of these individual points as they spiral inward while maintaining a fixed distance between consecutive segments. As the spiral path tightens, the risk of "self-collision"—where the dragon's head overlaps with body segments—becomes inevitable. This research provides a high-fidelity kinematic model and a robust collision detection algorithm to simulate this movement.

Kinematic Modeling and Numerical Integration

The path is defined by an Archimedean spiral where the pitch (distance between successive turnings) is a critical parameter. To ensure the dragon segments behave as rigid benches rather than elastic strings, the model enforces a specific kinematic constraint: for any two consecutive points, the difference in their position vectors must be perpendicular to the difference in their velocity vectors.

Vector diagrams illustrating the perpendicularity constraint between position and velocity differences to maintain rigid segment lengths.

Mathematically, this is expressed as $(\vec{r}_i - \vec{r}_{i+1}) \cdot (\vec{v}_i - \vec{v}_{i+1}) = 0$. By applying the fourth-order Runge-Kutta (RK4) method, the researchers iteratively calculated the position and velocity of all 224 points at one-second intervals, ensuring numerical stability throughout the simulation.

Collision Detection Strategy

To detect overlaps, the benches were modeled as rectangles. Instead of using computationally expensive trigonometric checks, the authors employed a Boolean logic approach simplified by De Morgan’s Law. By defining the conditions under which two rectangular bounding boxes do not overlap, they derived an efficient "overlap" formula. This allowed the algorithm to traverse the segments and identify the exact moment of intersection between the dragon's head and any other body segment.

Results and Discussion

The simulation successfully tracked the formation from $t=0$ to $t=300$ seconds with high precision. While the dragon head maintained a constant speed of 1 m/s, segments further down the body exhibited a slight, smooth velocity decay as they navigated the increasing curvature of the inner spiral.

The model identified the critical failure point at 432.18 seconds. At this time, during the 13th loop of the spiral, the dragon head collided with the 6th segment of its own body.

Visualization of the full dragon formation along the Archimedean spiral, highlighting the point of first collision between the head and the 6th body segment.

Conclusion

This study demonstrates that even with perfect coordination, the geometry of a tightening spiral makes collision mathematically certain. The combination of the perpendicularity constraint for rigid body modeling and efficient Boolean collision detection provides a robust framework for simulating complex multi-point systems. Future work will look to incorporate dynamic forces, such as centrifugal force and mass, and explore active avoidance algorithms to delay collision times in swarm robotics or automated traffic management.

Summary updated

Abstract

This study presents a mathematical model of the motion of a dragon dance team moving along an Archimedean spiral path. The motion of 224 points representing the dragon dance team is described using parametric equations, with the velocity of each point determined through the Runge-Kutta method. The study investigates the dynamics of the dragon head and body segments as they spiral inward, with a focus on maintaining fixed distances between consecutive points. Additionally, collision detection between the dragon's benches is implemented to prevent overlap. The termination time of the motion is determined by identifying the first collision between the dragon head and the body segment. The results demonstrate the application of numerical methods to model complex dynamic systems and ensure smooth movement without collision. The first collision occurs at 432.18 seconds, during the dragon's thirteenth loop. This study provides a valuable theoretical and practical basis for the motion analysis of complex dynamic systems by accurately simulating the movement of a dragon dance team along an Archimedean spiral and proposing a novel collision detection method.
BibTeX
@inproceedings{han2025research,
  author = {Jiawei Han and Yinuo Su and Ziheng Ding},
  title = {{Research on the Dragon Dance Formation Based on Spiral Motion Model and Collision Detection}},
  booktitle = {Highlights},
  year = {2025},
  doi = {10.54097/bkkh5m76},
}
Figure