AN AUTONOMOUS FOREST ROBOT THAT USES A HIERARCHICAL, FUZZY LOGIC CONTROLLER

Abstract

This article presents a hierarchical fuzzy logic controller that arbitrates recommendations from lower-levelmodules in order to autonomously guide a robot down forest trails. This robot would be used to move logs or trees from wherethey are cut to a location where they could be further processed or loaded onto trucks. The lower-level modules contributerecommendations derived from lower-level tasks or behaviors. Each lower-level module provides its own recommendationbased on its limited perception of the environment. A fuzzy quality measure is used by the hierarchical controller to determinehow best to fuse the individual control recommendations of lower-level modules. The quality measure provides an approximatedetermination of the value of each lower-level module recommendation. These quality measures, as well as other fuzzy setsused in the controller, were defined with a set of parameters. A computer simulation was used to tune subsets of the fuzzy setparameters to improve the performance of the controller. After these parameters were tuned, they were used in the fuzzy logiccontroller, which was tested in a vehicle. The results of these vehicle tests are also discussed in the article.