Research Areas

This section describes each area of the project.

Agent Based Simulation

A crowd simulation framework using modular and object-oriented design is developed in the project. The framework is flexible and efficient to support agent-based crowd modeling and simulation in a multi-layer virtual environment. The framework consists of three basic functional modules, which are Environment Construction module, Simulation module, and Visualization and Animation module. In the first module, we developed an ontology-based virtual environment denotation tool that allows the users to setup a multi-layer environment quickly. The tool allows the users to define and annotate the environment based on a set of 2D maps, to define objects based on a set of ontology in the XML files, and to identify the relationships between different regions in the environment. In the second module, we developed a flexible and extensible architecture based on RePast (a simulation engine) and Drools (a rule engine) for the modeling of human agents with behavior models and of intelligent environment with active-smart objects respectively. Under this architecture, different functional components are built to support efficient crowd simulation, including multi-layer environment retrieval model, active-smart objects, behavior and cognitive model, motion and path planning, and distributed simulation add-in. The environment retrieval model consists of different data structures, such as Hash Map, QuadTree, and Graph, for fast information retrieval. The active-smart objects interact with human agents dynamically based on a rule engine, and in the development prospective they reduce the complexity of programming the environment-related behaviors of agents. In the last module, a 2D interactive interface is developed to visualize the simulation results, and to support real-time interaction with the simulation. A log generation component is built to output results for 3D animation.

Behavioural Modeling

To model the agent's behaviors in our crowd simulation, we propose a generic framework for human behavior modeling, which generally follows the "perceive-decide-act" paradigm involved in real human behaviors. The term "human-like" is emphasized in our design, as we aims to reflect the way real human makes decisions and performs actions in real-life situation. In the perceiving stage of the framework, a sensory system is designed to imitate human's vision, hearing and other sensing abilities. The human's attention is also considered for filtering out irrelevant sensed information. In the deciding stage of the framework, we incorporate two important factors in human decision making: experience and emotion. In our framework, rather than relying on deliberate rational analysis, an agent makes its decisions by matching past experience cases to the current situation. We provide a clear representation of experience and investigate the mechanisms of situation assessment, experience matching and experience execution. To incorporate emotion into our framework, we introduce an emotion appraisal process in situation assessment for emotion elicitation. The decision making process of an agent may be affected by its emotional states when: 1) deciding whether it is necessary to do a re-match of experience cases, 2) categorizing the situation and 3) determining which experience to match with the current situation. In the acting stage of the framework, the high-level actions, such as seeking to a goal, moving as a group and searching along a direction are realized by following some simple qualitative rules that a real human usually follows. The computational algorithms for these actions are developed to drive the agent's motion. In general, our design focuses on the general mechanisms for the modeling of human behaviors and we suggest a context-oriented approach to build behavior model using the proposed framework.

Motion Planning

We present a rule-based motion planning system for agent-based crowd simulation, consisting of sets of rules for both collision avoidance and collision response. In order to avoid an on-coming collision, a set of rules for velocity sampling and evaluation is proposed, which aims to choose a velocity with an expected time to collision larger than a predefined threshold. In order to improve the efficiency over existing methods, the sampling procedure terminates upon finding an appropriate velocity. Moreover, the proposed motion planning system does not guarantee a collision-free movement. In case of collision, another set of rules is also defined to direct the agent to make a corresponding response.

Multi Resolution Modeling

Macroscopic and microscopic approaches for crowd simulation take a trade-off between efficiency and accuracy, but neither of them achieves the two goals at the same time. In the multi-resolution model, models with different level of resolutions are combined together to simulate the crowd  behavior and its movement in the high density situation. The simulator works under macroscopic and microscopic mode alternately, using the flow-based and agent-based models respectively. The switch of working mode is taken by either aggregation or disaggregation  operation between the two models. Thus, the simulation can take advantages from both models to improve the simulation performance as well as improving the accuracy.

 Distributed Simulation

 (...)

 3D Visualization

We present an approach to create a bridge between a federated agent-based crowd simulation architecture and a game engine to drive the visualization and animation of virtual crowd in a 3-dimentional space. In our simulation bridge framework, we present a method for data acquisition to create the virtual environment. An interfacing methodology is introduced between the 3D virtual environment maps and the federated agent-based crowd simulator. Our method comprises of converting the 3D maps to multi-level 2D maps to feed the 2D agent-based crowd simulator’s virtual world representation. A tool was implemented for defining objects in the virtual environment based on 2D maps and a color coder scheme. Our bridge framework makes use of a game engine as the visualization medium. The 3D virtual environment maps are reconstructed within the game engine and human avatars are created to simulate the agents. Motion, expression and behavioral animations are attached to each virtual agent. At run-time, data is fed from the 2D crowd simulator to our framework via an event-driven file buffer. Virtual agents are created at run-time based on the social aspects of the crowd composition. Behavioral animation is triggered on each agent based on the commands represented in the simulation results. Our framework also provides visualization of emergency situations and makes use of particle dynamics to generate more realistic visuals of disaster situations and human crowd behavior in an emergency situation. Rendering and visualization performances using GPU rendering optimization techniques are also investigated. Another research area in this section is the extensibility of our framework to Military COTS packages for visualization or interaction.

Publication List

1. Malcolm Yoke Hean Low, Wentong Cai, and S.P. Zhou, “A Federated Agent-based Crowd Simulation Architecture”, in Procs. of 21st European Conference on Modelling and Simulation (ECMS 2007)
   
2. S.P. Ting and Suiping Zhou, “Quartz: An Autonomous Navigation System for MOUT Simulations”, in Procs. of International Conference on Computer Animation and Social Agents (CASA 2007).
3. Suiping Zhou, Linbo Luo, W. L. Koh, S.P. Ting, “Human behavior modeling for crowd simulation”, in Proceedings of International Conference and Symposium on Computer Games; Animation, Multimedia, IPTV and Edutainment (CGAT’08)
4. S.P. Ting, Suiping Zhou, “SNAP: A time critical decision making framework for MOUT simulations”, in Computer Animation and Virtual Worlds, also in Proceedings of the International Conference on Computer Animation and Social Agents (CASA 2008)
5. Linbo Luo, S.P. Zhou, Wentong Cai, Malcolm Yoke Hean Low, F. Tian, Yongwei Wang, X. Xiao, and D. Chen, “Agent-based Human Behavior Modeling for Crowd Simulation”, in Computer Animation and Virtual Worlds, also in Proceedings of the International Conference on Computer Animation and Social Agents (CASA 2008)
6. Yuanxi Liang, Wentong Cai, Stephen J Turner, Georgios K. Theodoropoulos, and Rob Minson, “Interfacing PePast with HLA Using a Generic Architecture for COTS Simulation Package Interoperability”, accepted by 2009 Spring Simulation Interoperability Workshop (S-SIW 2009), San Diego, California, USA, Mar 2009.
7. Yongwei Wang, Wentong Cai, Malcolm Yoke Hean Low, Suiping Zhou, Feng Tian, Linbo Luo, Darren Wee Sze Ong, and Benjamin D. Hamilton, “A Framework of Evaluating  Partitioning Mechanisms for Agent-based Simulation Systems”, accepted by 42nd Annual Simulation Symposium (ANSS 2009), San Diego, California, USA, Mar 2009.
8. Muzhou Xiong, Wentong Cai, Suiping Zhou, Malcolm Yoke Hean Low, Feng Tian, Chen Dan, Darren Wee Sze Ong, and Benjamin D. Hamilton, “A Case Study of Multi-Resolution Modeling for Crowd Simulation”, accepted by 2009 Agent-Directed Simulation Symposium (ADS 2009), San Diego, California, USA, Mar 2009.
9. Kabilen Sornum, Yuanxi Liang, Wentong Cai, Malcolm Yoke Hean Low, Suiping Zhou, “3D Visualization and Animation of Crowd Simulation Using a Game Engine”, accepted by 2009 International Conference & Symposium on Computer Games, Animation, Multimedia, IPTV, Edutainment & Security (CGAT 2009), Singapore, May 2009.
10. Muzhou Xiong, Michael Lees, Wentong Cai, Suiping Zhou and Malcolm Yoke Hean Low. “A Rule-Based Motion Planning for Crowd Simulation” In Proceedings IEEE International conference on Cyberworlds 2009 (CW 2009)
11. Linbo Luo, Suiping Zhou, Wentong Cai, Malcolm Yoke Hean Low and Michael Lees “Toward A Generic Framework for Modeling Human Behaviors in Crowd Simulation” in proceedings of the 2009 IEEE/WIC/ACM International conference of Intelligent Agent Technology (IAT 2009) 
12. Yongwei Wang, Michael Lees, Wentong Cai, Suiping Zhou and Malcolm Yoke Hean Low. "Cluster Based Partitioning For Agent-Based Crowd Simulations". In Proceedings of the 2009 Winter Simulation Conference (WSC 2009).
13. Yuan Wei Chua and Malcolm Yoke Hean Low. “Predictive Algorithms for Aggregation and Disaggregation in Mixed Mode Simulation”. In Proceedings of the 2009 Winter Simulation Conference (WSC 2009).

 Completed FYP Projects

1.  “A Multi-agents based Crowd Simulation”, by Seet Yew Siang, May 2007.
2. “HLA-based Adaptors for Multi-player Networked Games”, by Nguyen Ngoc Nam, May 2007.
3. “A Knowledge-based Environment and Behaviour Representation for Game System”, by Sean Ng Boon Kiat, May 2007.
4. “Simulation of Crowd Behaviours in Military Operations on Urbanized Terrain”, by Hu Nan, May 2007.
5. “Multi-agents based Crowd Simulation using PECS”, by Toh Qi Xian, May 2008.
6. “Investigation of Context Adaptability in Multi-agents based Crowd Simulation”, by Lee Zhiwei, May 2008.
7. “Cognitive Behaviour Modelling in a Dynamic Environment”, by Wu Jian Liang, May 2008.
8. “Crowd Simulation with Smart Objects and Inference Engine”, by Tan Su Li Debbie, May 2009 
9. “Aggregation and Disaggregation Issues in Distributed Simulation”, by Benjamin Chua Yuan Wei, May 2009

Completed MSc Thesis

1.  “Modelling and Simulation of Crowd Behaviour”, by Chua Beng Hwee, May 2007

On-going FYP Projects

1. “An Analysis of Path-finding Algorithms for Agent-based Crowd Simulation”, by Hua Khac Nam, May 2010
2. “Agent Simulation of Crowd Behaviour”, by Seow Wei Quien, May 2010
3. “Developing Methods for Modelling the Human Sensory System in Crowd Simulation”, by Vaisagh Viswanathan, May 2010