Evolutionary fuzzy control of two cooperative object-carrying wheeled robots for wall following through multiobjective continuous ACO

Abstract

This paper considers the multiobjective orientation and velocity control of two robots cooperatively carrying an object in completing a convex wall-following task. Fuzzy controllers (FCs) found through a two-stage learning approach is used. The first stage finds FCs that control a single (leader) robot in executing the wall-following task with the two objectives of accurately moving along the wall boundary and moving with a high speed. Based on the FC in stage one, the second stage finds an auxiliary FC to control a follower robot so that the leader and follower robots can cooperatively carry an object in executing the wall-following task. The FC in each stage is optimized through the multiobjective front-guided continuous ant-colony optimization (MO-FCACO) algorithm. Simulations of evolutionary fuzzy control of a single robot and two cooperative robots for wall following are performed to show the control result.