Fuzzy behavior-based control (3rd report, module learning and behavioral coordination by an adaptive fusion unit with external sensor information)

In order to obtain more effective coordinated behaviors against changes in environments or unlearned environments, we here propose a new fusion unit using external sensor information for a fuzzy behavior-based control system trained in the framework of module learning. The present fusion unit falls into a class of adaptive priority-based architecture in which the fusion between behavioral elements is carried out by a cooperative or competition strategy, depending on whether an objective point or obstacle can be detected or not. Some experimental results of a miniature mobile robot Khepera show that the proposed method is superior to the conventional method that applies a fusion unit based on a fixed priority-based architecture.