本文旨在以低成本的紅外線感測器與低運算演算法於移動式機器人建構出具有避障功能的跟隨行為。藉由低成本的紅外線距離感測器偵測跟隨目標的距離,並藉由陣列式架構的紅外線接收與發射器得到跟隨目標的方向,並以額外測程導航法來對應紅外線接收發射器訊號喪失時跟隨行為的實現。同時藉由人類移動時的避障行為,建構出以低運算量且不影響跟隨功能的基本避障能力。最終藉由上述發展的演算法與感測器建構,測試在不同環境與距離下的跟隨行為的狀況,並驗證系統的架構與運作狀況。
We report on the sensory setup, algorithm development of motion generation with obstacle avoidance, and system-level integration for a mobile robot that follows a human. The target's direction and distance are detected with a low-cost, compact and coded infrared (IR) transmitter/receiver pair and a range finder, free from environmental lighting conditions. Based on the characteristics that the human would avoid obstacles while moving, instead of using a complex algorithm suitable for autonomous navigation by mobile robots, a simple obstacle avoidance algorithm with low computation power is developed. The robot implemented with the developed sensor and algorithm can reactively track and follow the target at various distances, and moves in a complex environment. In addition, the added odometry navigation allows the robot to continue track the target when its transmitting signal is occasionally obstructed. Experimental validation of following humans is executed to evaluate the performance of the proposed system setup.