Virtual crowds [electronic resource] : steps toward behavioral realism / Mubbasir Kapadia, Nuria Pelechano, Jan Allbeck, Norm Badler.

Kapadia, Mubbasir., author.
San Rafael, California (1537 Fourth Street, San Rafael, CA 94901 USA) : Morgan & Claypool, 2016.
1 online resource.
Synthesis digital library of engineering and computer science.
Synthesis lectures on visual computing ; 2469-4223 # 20.
Synthesis lectures on visual computing, 2469-4223 ; # 20

Location Notes Your Loan Policy


Crowds -- Computer simulation.
Collective behavior -- Computer simulation.
Intelligent agents (Computer software).
System Details:
Mode of access: World Wide Web.
This volume presents novel computational models for representing digital humans and their interactions with other virtual characters and meaningful environments. In this context, we describe efficient algorithms to animate, control, and author human-like agents having their own set of unique capabilities, personalities, and desires. We begin with the lowest level of footstep determination to steer agents in collision-free paths. Steering choices are controlled by navigation in complex environments, including multi-domain planning with dynamically changing situations. Virtual agents are given perceptual capabilities analogous to those of real people, including sound perception, multi-sense attention, and understanding of environment semantics which affect their behavior choices. The roles and impacts of individual attributes, such as memory and personality are explored. The animation challenges of integrating a number of simultaneous behavior and movement demands on an agent are addressed through an open source software system. Finally, the creation of stories and narratives with groups of agents subject to planning and environmental constraints culminates the presentation.
1. Introduction

Part I. Multi-agent collision avoidance
2. Background
2.1 Centralized approaches
2.2 Agent-based approaches
2.2.1 Data-driven approaches
2.2.2 Predictive approaches
2.3 Locomotion synthesis
2.4 Challenges and proposed solutions
2.4.1 Particle-based agent models
2.4.2 Decoupling between steering and locomotion
2.4.3 Generalization and applicability of data-driven approaches
3. Footstep-based navigation and animation for crowds
3.1 Introduction
3.2 Locomotion model
3.2.1 Inverted pendulum model
3.2.2 Footstep actions
3.2.3 Locomotion constraints
3.2.4 Cost function
3.3 Planning algorithm
3.4 Evaluation
3.4.1 Interfacing with motion synthesis
4. Following footstep trajectories in real time
4.1 Animating from footsteps
4.2 Framework overview
4.3 Footstep-based locomotion
4.3.1 Motion clip analysis
4.3.2 Footstep and root trajectories
4.3.3 Online selection
4.3.4 Interpolation
4.3.5 Inverse kinematics
4.4 Incorporating root movement fidelity
4.5 Results
4.5.1 Foot placement accuracy
4.5.2 Performance
5. Context-sensitive data-driven crowd simulation
5.1 Steering in context
5.2 Steering contexts
5.3 Initial implementation
5.3.1 Training data generation
5.3.2 Oracle algorithm
5.3.3 Decision trees
5.3.4 Steering at runtime
5.4 Results
5.4.1 Classifier accuracy
5.4.2 Runtime
5.4.3 Collisions
6. Conclusion
6.1 Footstep-based collision avoidance
6.2 Footstep-based locomotion
6.3 Context-based steering

Part II. Multi-agent navigation
7. Background
7.1 Navigation meshes
7.2 Planning
8. Navigation meshes
8.1 NavMeshes from 3D geometry: NEOGEN
8.1.1 GPU coarse voxelization
8.1.2 Layer extraction and labeling
8.1.3 Layer refinement
8.1.4 NavMesh generation
8.2 Results
8.3 Limitations and discussion
9. Multi-domain planning in dynamic environments
9.1 Multi-domain planning
9.2 Overview
9.3 Planning domains
9.3.1 Multiple domains of control
9.4 Problem decomposition and multi-domain planning
9.4.1 Planning tasks and events
9.5 Relationship between domains
9.5.1 Domain mapping
9.5.2 Mapping successive waypoints to independent planning tasks
9.6 Results
9.6.1 Comparative evaluation of domain relationships
9.6.2 Performance
9.6.3 Scenarios
10. Conclusion

Part III. Perception
11. Background
12. Sound propagation and perception for autonomous agents
12.1 Sound categorization and representation
12.1.1 Sound feature selection and categorization
12.1.2 Sound packet representation (SPR)
12.1.3 SPR selection for hierarchical cluster analysis
12.2 Sound packet propagation
12.2.1 Transmission line matrix using uniform grids
12.2.2 Pre-computation for TLM using a quad-tree
12.3 Sound perception and behaviors
12.3.1 Effect of sound degradation on perception
12.3.2 Hierarchical sound perception model
12.3.3 Sound attention and behavior model
12.4 Experiment results
12.4.1 Applications
13. Multi-sense attention for autonomous agents
13.1 Introduction
13.2 Methodology
13.2.1 Object and action representations
13.2.2 Sense preprocessing
13.2.3 Sensing
13.3 Hierarchical aggregate clustering
13.3.1 Environment-centric clustering
13.3.2 Agent-centric clustering
13.3.3 Aggregate properties
13.4 Analysis and results
14. Semantics in virtual environments
14.1 Incorporating semantics
14.1.1 Lexical databases
14.1.2 Modularized smart objects
14.2 Semantic generation
14.2.1 Hierarchy generation
14.2.2 Semantic modularization
14.2.3 Runtime performance
14.3 Limitations
15. Conclusion

Part IV. Agent-object interactions and crowd heterogeneity
16. Background
17. Parameterized memory models
17.1 Memory system
17.1.1 Memory representation
17.1.2 Sensory memory
17.1.3 Working memory
17.1.4 Long-term memory
17.2 Example and analysis
17.3 Future work
18. Individual differences
18.1 Personality
18.1.1 Personality-to-behavior mapping
18.1.2 User studies on personality
18.2 Roles and needs
18.2.1 Approach
18.2.2 Implementation
19. Conclusion

Part V. Behavior and narrative
20. Background
21. An open source platform for authoring functional crowds
21.1 ADAPT
21.2 Framework
21.2.1 Full-body character control
21.2.2 Steering and path-finding
21.2.3 Behavior
21.3 Shadows in full-body character animation
21.3.1 Choreographers
21.3.2 The coordinator
21.3.3 Using choreographers and the coordinator
21.3.4 Example choreographers
21.4 Character behavior
21.4.1 The ADAPT character stack
21.4.2 Body capabilities
21.5 Character interactions
21.5.1 Characters interacting with each other
21.5.2 Characters interacting with the environment
21.6 Results
21.6.1 Multi-actor simulations
21.6.2 Computational performance
22. Event-centric planning for narrative synthesis
22.1 Problem domain and formulation
22.1.1 State space
22.1.2 Action space
22.1.3 Goal specification
22.2 Planning in event space
22.3 Runtime and simulation
22.3.1 Event loading and dispatch
22.3.2 Handling dynamic world changes
22.3.3 Intelligent ambient character behavior
22.4 Results
22.4.1 Environment design
22.4.2 Object state description
22.4.3 Authored events
22.4.4 Generated narrative
22.4.5 Reacting to user intervention
23. Conclusion
24. Epilogue

Authors' biographies.
Part of: Synthesis digital library of engineering and computer science.
Title from PDF title page (viewed on November 24, 2015).
Includes bibliographical references (pages 219-245).
Pelechano, Nuria., author.
Allbeck, Jan M., author.
Badler, Norman I., author.
Other format:
Print version:
Publisher Number:
10.2200/S00673ED1V01Y201509CGR020 doi
Access Restriction:
Restricted for use by site license.