Context-sensitive data-driven crowd simulation

In terms of computation, steering through a crowd of pedestrians is a challenging task. The problem space is inherently high-dimensional, with each added agent giving yet another set of parameters to consider while finding a solution. Yet in the real world, navigating through a crowd of people is very similar regardless of the population size. The closest people have the most impact while those distant set a more general strategy. To this end, we propose a data-driven system for steering in crowd simulations by splitting the problem space into coarse features for the general world, and fine features for other agents nearby. The system is comprised of a collection of steering contexts, which are qualitatively different overall traffic patterns. Due to their similarity, the scenarios within these contexts have a machine-learned model fit to the data of an offline planner which serves as an oracle for generating synthetic training data. An additional layer of machine-learning is used to select the current context at runtime, and the context's policy consulted for the agent's next step. We experienced speedup from hours per scenario with the offline planner and 10 agents to an interactive framerate of 10FPS for 3,000 agents using our data-driven technique.

[1]  Dani Lischinski,et al.  Crowds by Example , 2007, Comput. Graph. Forum.

[2]  Ming C. Lin,et al.  Towards a Digital Makkah – Using Immersive 3 D Environments to Train and Prepare Pilgrims , 2011 .

[3]  Adrien Treuille,et al.  Continuum crowds , 2006, ACM Trans. Graph..

[4]  Helbing,et al.  Social force model for pedestrian dynamics. , 1995, Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics.

[5]  Jessica K. Hodgins,et al.  Reactive pedestrian path following from examples , 2004, The Visual Computer.

[6]  Dinesh Manocha,et al.  Reciprocal Velocity Obstacles for real-time multi-agent navigation , 2008, 2008 IEEE International Conference on Robotics and Automation.

[7]  Dinesh Manocha,et al.  ClearPath: highly parallel collision avoidance for multi-agent simulation , 2009, SCA '09.

[8]  Demetri Terzopoulos,et al.  Autonomous pedestrians , 2007, Graph. Model..

[9]  Rahul Narain,et al.  Aggregate dynamics for dense crowd simulation , 2009, SIGGRAPH 2009.

[10]  Daniel Cohen-Or,et al.  Context‐Dependent Crowd Evaluation , 2010, Comput. Graph. Forum.

[11]  Craig W. Reynolds.  Flocks, herds, and schools: a distributed behavioral model , 1998 .

[12]  Norman I. Badler,et al.  Virtual Crowds: Methods, Simulation, and Control , 2008, Virtual Crowds: Methods, Simulation, and Control.

[13]  Stéphane Donikian,et al.  A synthetic-vision based steering approach for crowd simulation , 2010, ACM Transactions on Graphics.

[14]  Glenn Reinman,et al.  Footstep navigation for dynamic crowds , 2011, SI3D.

[15]  Daniel Thalmann,et al.  Crowd simulation , 2007 .

[16]  Dimitris N. Metaxas,et al.  Eurographics/ Acm Siggraph Symposium on Computer Animation (2007) Group Behavior from Video: a Data-driven Approach to Crowd Simulation , 2022 .

[17]  Petros Faloutsos,et al.  Egocentric affordance fields in pedestrian steering , 2009, I3D '09.

[18]  Nicolas Courty,et al.  Data-Driven Animation of Crowds , 2007, MIRAGE.

[19]  Craig W. Reynolds.  Steering Behaviors For Autonomous Characters , 1999 .

[20]  Norman I. Badler,et al.  Controlling individual agents in high-density crowd simulation , 2007, SCA '07.

[21]  P. Torrens,et al.  Building Agent‐Based Walking Models by Machine‐Learning on Diverse Databases of Space‐Time Trajectory Samples , 2011 .

[22]  Glenn Reinman,et al.  Scenario space: characterizing coverage, quality, and failure of steering algorithms , 2011, SCA '11.