Subsribe to our RSS Feed

Virtual Reality Research Topics

Crowd Simulation

  • Overview


    A crowd simulation mimics synthetically the movement of a large number of pedestrians in a given environment. If a real world system is modeled, the simulation can be useful in gaining insights about the real world so long as the pedestrians behave as faithfully as possible to the real movements of pedestrians in the real world. This is why related psychological and social findings and observations are highly valued in our research work. Simulated crowd behavior is directly affected by an agent's interaction with other agents and surroundings. Individual and crowd level agent motions together determine how that simulated crowd behaves. Individual level agent motion focuses on in-place individual desired movements and actions by using inverse kinematics, baked animations, and/or motion capture methods. Crowd level agent motion on the other hand is determined by the interaction between agents forming a virtual crowd which is determined by the perception, cognition, and decisions made by agents when performing path planning, group behavior, and steering behavior.


  • Goal of the Research


    Our work improved the realism of group behavior in a crowd simulation in regards to collision avoidance, navigation, and social behaviors.


  • Research Participants


    Current Members

    • Jin Ki Jung
    • Hyeop Woo Lee
    • Chur Woong Park
    Past Members

    • Francisco Rojas
    • Jin Hyoung Park



  • Research

  • Title Summary Movie

    Minimizing Collision among Social Groups in Wide-Open Spaces




    A novel approach is introduced for collision avoidance (CA) among social groups that uses the collective gaze-movement angle (CGMA) between groups as the primary approach for collision prediction. Rules are used for handling the CA steering such that it mimics how real human groups travel together in wide-open walkable areas. Reciprocal velocity obstacle (RVO) based CA is applied to groups as an alternative, and this alternative is used as a baseline for comparison to the CGMA approach for group-to-group CA. In an immersive evaluation using a head-mounted display, a group of people determined that the CGMA based CA approach was much more believable than the RVO based CA approach for groups.


    Safe Navigation of Pedestrians in Social Groups in a Virtual Urban Environment

    (Distinguished Paper Award)




    The interpolation approach is used for the group agent. By interpolating the current and desired slot positions of the group agent using formation templates, dynamic social group formations occur and adapt to the width of passageways by using the proposed ray casting technique. This approach is extended by slot-locking which keeps subgroups in a group shoulder-to-shoulder regardless of the current formation state assuming sufficient surrounding space exists. When this happens, hand holding between adjacent members may occur. Individual characteristics such as age and body shape impact the overall group speed. In an immersive evaluation using an Oculus Rift HMD, participants validated the realism of dynamic social group behavior as virtual pedestrians safely navigated via traffic light crosswalks and sidewalks with vehicular traffic on the streets.


    Immersive Human-in-the-Loop HMD Evaluation of Dynamic Group Behavior in a Pedestrian Crowd Simulation that uses Group Agent-based Steering




    Two approaches for social group movement are offered for a slot-based group agent: the basis of the first is an articulated object with hinges at the inner slots which allows for dynamic formation transitions, and the second approach which has no hinges uses interpolation to do the same. In the articulated object approach, we accommodate sub-groups of people in the group who are more strongly bonded (e.g. couple, parent and child, talkative friends) so that they can continue to socialize or hold hands so long as the surrounding environment conditions permit it. Helped by many participants one-at-a-time, each conducted an immersive human-in-the-loop HMD evaluation of the dynamic group behaviors from a first-person view.


    Group Agent-based Steering for the Realistic Corner Turning and Group Movement of Pedestrians in a Crowd Simulation









    Group agent-based steering keeps pedestrians moving around essentially shoulder to shoulder in a small group as in real life, leaving it up to the individual members for performing collision avoidance while keeping up with the group as a whole. This is useful for coherent corner turning. Furthermore, we also propose a way to handle traffic congestion by groups of agents corner turning through the use of group agent path decision making with the help of triggers. Experiments show that congestion is reduced and collision avoidance handling is simplified because there is more room for maneuvering.


    A Collision Avoidance Behavior Model for Crowd Simulation based on Psychological Findings












    We created a collision avoidance behavior model for crowd simulation based on psychological findings of human behaviors such as gaze movement angle (GMA), side stepping, gait motion, and personal reaction bubble to have better results in the crowd simulation. By calculating the GMA between agents, a collision can be predicted and avoided without knowing the exact trajectories of the agents. The proposed model consists of four phases: (1) GMA-based collision prediction for mid/long range by using speed-variant information process space, (2) collision avoidance steering, (3) gait-based locomotion generation, and (4) space keeping based on the personal reaction bubble.

    VR training simulation authoring tool for ship crew


    VR training simulation authoring tool for ship crew



    currently in progress