Introducing Anisotropic Fields for Enhanced Diversity in Crowd Simulation

Yihao Li,Junyu Liu,Xiaoyu Guan,Hanming Hou,Tianyu Huang

2024-09-24

Abstract:Large crowds exhibit intricate behaviors and significant emergent properties, yet existing crowd simulation systems often lack behavioral diversity, resulting in homogeneous simulation outcomes. To address this limitation, we propose incorporating anisotropic fields (AFs) as a fundamental structure for depicting the uncertainty in crowd movement. By leveraging AFs, our method can rapidly generate crowd simulations with intricate behavioral patterns that better reflect the inherent complexity of real crowds. The AFs are generated either through intuitive sketching or extracted from real crowd videos, enabling flexible and efficient crowd simulation systems. We demonstrate the effectiveness of our approach through several representative scenarios, showcasing a significant improvement in behavioral diversity compared to classical methods. Our findings indicate that by incorporating AFs, crowd simulation systems can achieve a much higher similarity to real-world crowd systems. Our code is publicly available at <a class="link-external link-https" href="https://github.com/tomblack2014/AF" rel="external noopener nofollow">this https URL</a>\_Generation.

Multiagent Systems

What problem does this paper attempt to address?

The problem this paper attempts to address is the lack of behavioral diversity in existing crowd simulation systems. Specifically, traditional crowd simulation methods often result in homogeneous simulation outcomes, lacking the complexity and diversity of real crowd behaviors. To solve this problem, the authors propose a new approach by introducing Anisotropic Fields (AFs) into the crowd simulation system to better reflect the complex behavior patterns of real crowds.

### Main Issues:
1. **Lack of Behavioral Diversity**: Existing crowd simulation systems usually lack behavioral diversity, leading to overly homogeneous simulation results.
2. **High Determinism**: Traditional methods rely too heavily on deterministic behavior patterns in local navigation and global path planning, making it difficult to capture the uncertainty and complexity of real crowds.
3. **Excessive Manual Intervention**: Traditional methods require a lot of manual intervention and parameter adjustments to achieve complex crowd behaviors, which is inefficient and results in unnatural outcomes.

### Solutions:
1. **Introduction of Anisotropic Fields (AFs)**: By introducing AFs, uncertainty can be incorporated into the scene, causing different agents at the same location to exhibit different movement tendencies, thereby increasing behavioral diversity.
2. **Behavioral Inertia Mechanism**: To reduce frequent behavioral fluctuations caused by randomness, a behavioral inertia mechanism is introduced, making agents more inclined to maintain their current movement state when generating movement tendencies.
3. **Flexible Generation Methods**: AFs can be generated through intuitive sketching or extracted from real crowd videos, making the crowd simulation system more flexible and efficient.

### Main Contributions of the Paper:
1. **Introduction of a New Anisotropic Layer**: A new anisotropic layer is proposed, which does not belong to traditional global path planning or local navigation structures, effectively addressing the issue of insufficient behavioral diversity in traditional methods.
2. **User-Friendly Interactive Interface**: A series of user interfaces are designed, allowing users to modify AFs in real-time using virtual brushes, enabling users to create crowd simulation systems that meet their expectations.
3. **Similarity Measurement Method**: A method based on AFs and information entropy is proposed to measure the similarity of crowd systems, and experiments have verified the similarity relationship between simulated crowd systems and real crowd systems.

In summary, this paper aims to improve the diversity and realism of crowd simulation systems by introducing Anisotropic Fields, making them closer to real-world crowd behaviors.