Research Directions of the Center

Our overall goals of the Center for Human Modeling and Simulation are the investigation of computer graphics modeling, animation, and rendering techniques. Major focii are in behavior-based animation of human movement, modeling through physics-based techniques, applications of control theory techniques to dynamic models, illumination models for image synthesis, and understanding the relationship between human movement, natural language, and communication.

• Human Occupants in 3-Dimensional Virtual Environments: This major project involves simulating the movements of one or more human figures, determining their mobility, seeing their viewable and reachable spaces, checking for collisions with other occupants or the environment, scaling the bodies to specific population statistics, modeling strength capabilities, and animating actual tasks to be carried out in that environment. This effort has produced a software system called Jack which uses inverse kinematic positioning from multiple 3-D joint position constraints and biomechanical principles to model appropriate human behaviors. Jack is used in applications such as human factors analysis, workplace evaluation and assessment, product design evaluation, logistics and maintenance training, virtual environment marksmanship, and real-time dismounted soldiers and medical corpsmen in Distributed Interactive Simulation (DIS). These efforts are primarily supported by the Army Research Office, the Army Research Lab (Aberdeen), the Air Force Armstrong Labs (Logistics Research Division), Sandia National Laboratories, the Naval Air Warfare Center Training Systems Division, the Defense Modeling and Simulation Office, and the Advanced Research Projects Agency.
• Animation Control Techniques: A variety of methods for specifying and controlling animation, primarily of human figures, are studied. Techniques include interactive systems, posture positioning and interpolation, constraint systems, collision avoidance, locomotion models, strength-guided motion, dynamics, behavioral simulation, motor-control methodologies, motion learning, reactive planning, task description methodologies, task performance measurement, artificial intelligence models, and natural language directives. Gestural motions, facial expression, lip movement and coarticulation,, head motion, eye movements, speech and speech intonation are generated from a dialogue planner. These efforts are also supported by the Army Research Office, the Army Research Lab (Aberdeen), the Air Force Armstrong Labs (Logistics Research Division), the Naval Air Warfare Center Training Systems Division, the Defense Modeling and Simulation Office, and the Advanced Research Projects Agency. Face animation and gesture modeling is supported by the National Science Foundation.
• Language-Based Interfaces: The language-related effort includes motion verb representations, semantics of verbs and adverbial modifiers, reactive planning, and formulation of executable animation primitives. Action simulation requires agent capability, resource, and responsibility models, tool use understanding, functionality representation, knowledge-base partitioning, and object-specific reasoning. This research is primarily supported by the Army Research Lab (Aberdeen), the Air Force Armstrong Labs (Logistics Research Division), and the Advanced Research Projects Agency.
• Physics-Based Modeling, Animation and Control: This effort deals with the use of physics-based techniques for modeling, animation and control of dynamic models. Techniques include: modeling of complex dynamic objects through blending of parametric and spline or finite element models, systematic techniques for converting geometric degrees of freedom to physical degrees of freedom, dynamic animations through the Lagrange equations of motion, deformable model simulations, modeling of elastic or viscoelastic materials, finite element techniques, computationally efficient dynamic constraint algorithms, collision detection and impact force computation algorithms, fluid simulations, and control of dynamic models through the use of control theory algorithms. This work is supported by the National Science Foundation and the National Library of Medicine.
• Biomedical Applications and Scientific Visualization: Our major focii are physics-based modeling and visualization of human internal organs, skeleton, and muscles, and their motions and deformations. Specific instances where such models can be applied are crash simulations, virtual surgery, blunt or penetrating trauma assessment, model shape recovery from CT, MRI and/or SPAMM data, recovery of human motion from video or range data, and assessment of an athlete's fitness, heart modeling, heart motion description, blood flow simulation, blood flow estimation from SPAMM or phase-shift techniques. This work is supported by the Children's Hospital of Philadelphia and the National Science Foundation and the National Library of Medicine.
• Rendering Techniques for Complex Environments: This research deals with the simulation of and user interaction with environments of extreme geometric complexity. Initial investigations focus on the creation of automatic and adaptive techniques for progressive refinement rendering methods. This work is funded by the Army Research Office Center of Excellence in Artificial Intelligence at the University of Pennsylvania.