A system of interconnected mobile robot control patterns based on decomposition of the main control problem

Abstract

When a mobile robot switches from goal-directed motion to obstacle avoidance, it must retain the original goal and determine whether the available measurements permit the manoeuvre to continue. This paper describes that interaction through three control patterns: continuous monitoring, motion, and obstacle avoidance. The aim is to give each pattern a common description linking its local goal, input data, control action, and verifiable result. The approach uses decomposition of the main control problem and the kinematics of a differential-drive platform. Pattern states and transitions are specified at the algorithmic level. Continuous monitoring checks measurement freshness and the robot's response to commands. The motion pattern computes linear and angular velocity setpoints. Once an obstacle is confirmed, command generation passes to the avoidance pattern, which retains the goal and the history of the manoeuvre. Nominal motion resumes only after data validity, safe clearance, a free corridor, and renewed progress have each been checked. Failure of any requirement prevents a successful completion status.