Presentation is loading. Please wait.

Presentation is loading. Please wait.

Autonomous Virtual Humans Tyler Streeter. Contents Introduction Introduction Implementation Implementation –3D Graphics –Simulated Physics –Neural Networks.

Similar presentations


Presentation on theme: "Autonomous Virtual Humans Tyler Streeter. Contents Introduction Introduction Implementation Implementation –3D Graphics –Simulated Physics –Neural Networks."— Presentation transcript:

1 Autonomous Virtual Humans Tyler Streeter

2 Contents Introduction Introduction Implementation Implementation –3D Graphics –Simulated Physics –Neural Networks –Genetic Algorithms Results Results Future Work Future Work

3 Introduction Project goal: to create motor control systems for virtual humans to allow them to control their behaviors without using pre-scripted animations Project goal: to create motor control systems for virtual humans to allow them to control their behaviors without using pre-scripted animations Simulations take place in a graphical 3D environment with simulated physics Simulations take place in a graphical 3D environment with simulated physics Control systems are artificial neural networks evolved with genetic algorithms Control systems are artificial neural networks evolved with genetic algorithms

4 Implementation (1/4) C++, Windows & Linux C++, Windows & Linux Graphics: OpenGL Graphics: OpenGL Physics: Open Dynamics Engine Physics: Open Dynamics Engine –Simple ODE demo –Limp virtual human demo

5 Implementation (2/4) Artificial neural network control system Artificial neural network control system Body part positions, head velocity, etc. Joint torques (“muscles”)

6 Implementation (3/4) Genetic algorithms Genetic algorithms –Start with random population of solutions –Evaluate each individual’s “fitness” –Throw away bad solutions –“Mate” good solutions to produce offspring –Randomly mutate new offspring Algorithm is stochastic hill-climbing, so it can find ways to get out of local maxima Algorithm is stochastic hill-climbing, so it can find ways to get out of local maxima

7 Implementation (4/4) My genetic algorithm “learning” system My genetic algorithm “learning” system 1.2. Bad Good 3. 4.

8 Results (1/4)

9 Results (2/4)

10 Results (3/4)

11 Results (4/4) Video Demonstrations Video Demonstrations –Standing Video –Jumping Video –Walking Video

12 Future Work (1/2) Try more complex behaviors Try more complex behaviors –Staying balanced when pushed –Walking across uneven terrain –Carrying objects –Jumping over obstacles –Operating virtual machinery Use A* to guide paths Use A* to guide paths Competitions between virtual humans Competitions between virtual humans

13 Future Work (2/2) Create a hierarchy of neural nets to perform tasks with multiple steps Create a hierarchy of neural nets to perform tasks with multiple steps New sensory inputs to neural net New sensory inputs to neural net –Sense of touch –Sense of sight More complex bodies More complex bodies Try simulated annealing instead of ANNs Try simulated annealing instead of ANNs

14 References K. Sims. Evolving Virtual Creatures. ACM Computer Graphics (SIGGRAPH '94). pp. 15-22, 1994. Bentley, P. Digital Biology. Simon & Schuster, 2002. J. Klein. BREVE: a 3D Environment for the Simulation of Decentralized Systems and Artificial Life. Artificial Life VIII: The 8th International Conference on the Simulation and Synthesis of Living Systems. MIT Press, 2002. T. Reil, Husbands. Evolution of Central Pattern Generators for Bipedal Walking in a Real-Time Physics Environment. IEEE Transactions in Evolutionary Computation. pp. 159-168, April 2002. D. Thalmann, C. Babski, T. Capin, N. Magnenat Thalmann, I. S. Pandzic. Sharing VLNET worlds on the WEB. Computer Networks & ISDN Systems. Vol. 29, No. 14, pp. 1601-1610, 1997.


Download ppt "Autonomous Virtual Humans Tyler Streeter. Contents Introduction Introduction Implementation Implementation –3D Graphics –Simulated Physics –Neural Networks."

Similar presentations


Ads by Google