An Adaptive Approach for Reactive Actor Design
نویسندگان
چکیده
To address the complex and dynamic conditions of a virtual environment, computer animation researchers are applying methods similar to the ones used in artificial life to create reactive actors. A reactive actor is a control entity whose behavior is based on the sensory information it receives from the environment. The system presented within this paper, RAVE (Reactive Actors in Virtual Environments), demonstrates the successful use of a reinforcement learning model to automatically generate controllers for typical 2D navigational tasks. This is an improvement to existing methods because it requires no programming, can be used for a variety of tasks, and the control algorithms adapt during run-time. Collective Learning Systems (CLS) theory is integrated with a hierarchical controller to create control modules that quickly converge on optimal navigational strategies. Five different worlds are created to train and evaluate the actors. Performance metrics and results are presented for three different navigational tasks: obstacle avoidance (avoid), heading towards a goal object (goto), and moving away from a threat (retreat). CR Categories: G.3 [Probability and Statistics]: Probabilistic Algorithms; I.2.6 [Artificial Intelligence]: Learning; I.3.7 [Computer Graphics]: Three-Dimensional Graphics and Realism Animation, Virtual Reality. Additional Keyword: Behavioral Modeling, Collective Behavior, Virtual Actors, Learning Automaton.
منابع مشابه
Active distributed framework for adaptive hypermedia
Navigation through large hypermedia information spaces is complex and is an important application area for adaptive hypermedia systems. User navigation can be best supported when the design of the hypermedia system is embedded in an evolutionary process model that takes into account the decentralization of data sources and the variety of users. The paper deals with distributed frameworks for op...
متن کاملIntelligent Auto pilot Design for a Nonlinear Model of an Autonomous Helicopter by Adaptive Emotional Approach
There is a growing interest in the modeling and control of model helicopters using nonlinear dynamic models and nonlinear control. Application of a new intelligent control approach called Brain Emotional Learning Based Intelligent Controller (BELBIC) to design autopilot for an autonomous helicopter is addressed in this paper. This controller is applied to a nonlinear model of a helicopter. This...
متن کاملDesign of an Adaptive Fuzzy Estimator for Force/Position Tracking in Robot Manipulators
This paper presents a stable new algorithm for force/position control in robot manipulators. In this algorithm, position vectors are measured by sensors and then used in the control law. Since using force sensor has some issues such as high costs and technical problems, an approach is presented to overcome these issues. In this respect, force sensor is replaced by an adaptive fuzzy estimator to...
متن کاملAn economic-statistical model for production and maintenance planning under adaptive non-central chi-square control chart
Most of the inventory control models assume that quality defect never happens, which means production process is perfect. However, in real manufacturing processes, the production process starts its operation in the in-control state; but after a period of time, shifts to the out-of-control state because of occurrence of some disturbances. In this paper, in order to approach the model to real man...
متن کاملAn adaptive modified fuzzy-sliding mode longitudinal control design and simulation for vehicles equipped with ABS system
In order to improve the safety and longitudinal stability of a vehicle equipped with standard ABS system, this paper, analyzes the basic principles of vehicles stability and proposes a control strategy based on fuzzy adaptive control which will adjust PID gain parameters, using genetic algorithm. A linear three-degree-of-freedom (DOF) vehicle model was set up in Simulink and the stability test ...
متن کامل