EVOLVING FUZZY RULES FOR REACTIVE AGENTS IN DYNAMIC ENVIRONMENTS

Robotics and Autonomous Systems, 1998

This paper concerns tdae learning of basic behaviors in an autonomous robot. It presents a method to adapt basic reactive behaviors using a genetic algorithm. Behaviors are implemented as fuzzy controllers and the genetic algorithm is used to evolve their rules. These rules will be formulated in a fuzzy way using prefixed linguistic labels. In order to test the rules obtained in each generation of the genetic evolution process, a real robot has been used. Numerical results from the evolution rate of the different experiments, as well as an example of the fuzzy rules obtained, are presented and discussed.

Journal of Automation, Mobile Robotics and Intelligent Systems, 2010

2000

Abstract We describe two architectures that autonomously acquire fuzzy control rules to provide ,reactive behavioural competencies ,in a simulated mobile robotics application. One ,architecture is a “Pittsburgh”-style Fuzzy ,Classifier System (Pitt1). The other architecture is a “Michigan”- style Fuzzy Classifier System ,(Mich1). We tested the architectures on their ability to acquire an“investigative” obstacle ,avoidance competency. We found,that Mich1 implemented amore,local

… of the 2005 workshops on Genetic …, 2005

2008

IEEE Transactions on Fuzzy Systems, 2004

An unsupervised learning system, implemented as an autonomous agent is presented. A simulation of a challenging pathplanning problem is used to illustrate the agent design and demonstrate its problem solving ability. The agent, dubbed the ORG, employs fuzzy logic and clustering techniques to efficiently represent and retrieve knowledge and uses innovative sensor modeling and attention focus to process a large number of stimuli. Simple initial fuzzy rules (instincts) are used to influence behavior and communicate intent to the agent. Self-reflection is utilized so the agent can learn from its environmental constraints and modify its own state. Speculation is utilized in the simulated environment, to produce new rules and fine-tune performance and internal parameters. The ORG is released in a simulated shallow water environment where its mission is to dynamically and continuously plan a path to effectively cover a specified region in minimal time while simultaneously learning from its environment. Several paths of the agent design are shown, and desirable emergent behavior properties of the agent design are discussed.

Proceedings of the IEEE, 2001

The behavior of agents in complex and dynamic environments cannot be programmed a priori, but needs to self-adapt to the specific situations. We present some approaches based on evolutionary, reinforcement learning algorithms, able to evolve in real-time fuzzy models that control behaviors. We discuss an application where an agent learns how to adapt its behavior to the different behaviors of the other agents it is interacting with, and another application where a group of agents coevolve cooperative behaviors also by using explicit communication to propose the cooperation and to distribute reinforcement to the others.

International Journal of Approximate …, 1997

Fuzzy Logic Controllers constitute knowledge-based systems that include Fuzzy Rules and Fuzzy Membership Functions to incorporate human knowledge into their knowledge base. The de nition of fuzzy rules and fuzzy membership functions is one of the key question when designing Fuzzy Logic Controllers, and is generally a ected by subjective decisions. Some e orts have been made to obtain an improvement on system performance by incorporating learning mechanisms to modify the rules and/or membership functions of the FLC. Genetic Algorithms are probabilistic search and optimization procedures based on natural genetics. This paper proposes a way to apply (with a learning purpose) Genetic Algorithms to Fuzzy Logic Controllers, and presents an application designed to control the Synthesis of biped walk of a simulated 2-D biped robot.

Proceedings of the Second International Conference on Informatics in Control, Automation and Robotics, 2005

In this work we present the creation of a platform, along with an algorithm to evolve the learning of FLCs, especially aiming to the development of fuzzy controllers for mobile robot navigation. The structure has been proven on a Kephera robot. The conceptual aspects that sustain the work include topics such as Artificial Intelligence (AI), control advanced techniques, sensorial systems and mechatronics. Topics related with the control and automatic navigation of robotic systems especially with learning are approached, based on the Fuzzy Logic theory and evolutionary computing. We can say that our structure corresponds basically to a Classifier System, with appropriate modifications for the objective of generating controllers for mobile robot trajectories. The more stress is made on genetic profile than in the characteristics of the individuals and on the other, the strategy of distribution of the reinforcement is emphasized, fundamental aspects on which the work seeks to contribute.

Adaptive Behavior, 1997

We present an approach to support effective learning and adaptation of behaviors for autonomous agents with reinforcement learning algorithms. These methods can identify control systems that optimize a reinforcement program, which is, usually, a straightforward representation of the designer's goals. Reinforcement learning algorithms are usually too slow to be applied in real time on embodied agents, although they provide a suitable way to represent the desired behavior. We have tackled three aspects of this problem: the speed of the algorithm, the learning procedure, and the control system architecture. The learning algorithm we have developed includes features to speed up learning, such as niche-based learning, and a representation of the control modules in terms of fuzzy rules that reduces the search space, and improves robustness to noisy data. Our learning procedure exploits methodologies such as learning from easy missions and transfer of policy from simpler environments to the more complex. The architecture of our control system is layered and modular, so that each module has a low complexity and it can be learned in a short time. The composition of the actions proposed by the modules is either learned or pre-defined. Finally, we adopt an anytime learning approach to improve the quality of the control system on-line and to adapt it to dynamic environments. The experiments we present in this paper concern learning to reach another moving agent in a real, dynamic environment that includes nontrivial situations such as when the moving target is faster than the agent, and when the target is hidden by obstacles.