We present a novel dense crowd simulation method. In real crowds of high density, people manoeuvring the crowd need to twist their torso to pass between others. Our proposed method does not use the traditional disc-shaped agent, but instead employs capsule-shaped agents, which enables us to plan such torso orientations. Contrary to other crowd simulation systems, which often focus on the movement of the entire crowd, our method distinguishes between active agents that try to manoeuvre through the crowd, and passive agents that have no incentive to move. We introduce the concept of a focus point to influence crowd agent orientation. Recorded data from real human crowds are used for validation, which shows that our proposed model produces equivalent paths for 85 percent of the validation set. Furthermore, we present a character animation technique that uses the results from our crowd model to generate torso-twisting and side-stepping characters.

Simulating a crowded scene like a busy shopping street requires tight packing of virtual characters. In such cases, collisions are likely to occur, and the choice in collision detection shape will influence how characters are allowed to intermingle. Full collision detection is too expensive for crowds, so simplifications are needed. The most common simplification, the fixed-width, pose-independent cylinder, does not allow intermingling of characters, as it will either cause too much empty space between characters or undetected penetrations. As a possible solution to this problem, we introduce the bounding cylinder hierarchy (BCH), a bounding volume hierarchy that uses vertical cylinders as bounding shapes. Because the BCH is a generalization of the single cylinder, we expect that this representation can be easily integrated with existing crowd simulation systems. We compare our BCH with commonly used collision shapes, namely the single cylinder and oriented bounding box tree, in terms of query time, construction time, and represented volume. To get an indication of possible crowd densities, we investigate how close characters can be before collision is detected and finally propose a critical maximum depth for the BCH.

While walking through a crowd, a person balances several desires, such as reaching some goal position, avoiding collisions with others, and conserving energy. Crowd models generally try to mimic this behaviour by planning short paths that avoid collisions. However, when the crowd density increases, choosing a collisionfree path becomes more difficult. In such highdensity crowds, one can observe torso twists; people rotate their upper body to decrease their width perpendicular to the motion path, in order to squeeze through narrow spaces between other crowd members. In this paper we investigate this behaviour, by recording and analysing dense crowds. We show that the paths chosen by the participants can be predicted by generalized Voronoi diagrams, and identify relations between instantaneous speed and look-ahead distance, and between the participants' torso orientations and goal positions.

With the growth in available computing power, we see increasingly crowded virtual environments. In densely crowded situations, collisions are likely to occur, and the choice in collision detection technique can impact the perceived realism of a real-time crowd. This paper presents an investigation into the accuracy of human observers with regard to the recognition of collisions between virtual characters. We show the result of two user studies, where participants classify scenarios as "colliding" or "not colliding"; a pilot study investigates the perception of static images, whereas the main study expands on this by employing animated videos. In the pilot experiment, we investigated the effect of two variables on the ability to recognize collisions: distance between the character meshes and visibility of the inter-character gap. In the main experiment, we investigate the angle between the character paths and the severity of the (near) collision. On average, respondents correctly classified 72% (static) and 68% (animated) of the scenarios. A notable result is that the maximum uncertainty in determining existence of collisions occurs when the characters are overlapping and that there is a significant bias towards answering "not colliding." We also discuss differences in bias in the recognition of upper-and lower-body collisions.

No abstract