| Popis: |
With the advancement of factory logistics into the autonomous era comes the need to validate the safety and performance characteristics of Autonomous-Unmanned Ground Vehicles (A-UGVs) working in these application spaces ASTM Committee F45 has been developing standards for A-UGV performance measurement in various domains. The object detection and obstacle avoidance performance of A-UGVs in factories needs to be managed carefully as the action may cause severe damages, particularly when obstacles are either not detected or erroneously detected. In this paper, the grid-video measurement method is proposed to measure the small (e.g., short and/or thin) obstacle avoidance performance of A-UGVs. First, this paper describes the need for measuring the A-UGV performance through examples of small obstacles and the required A-UGV capability to avoid them. Next, the grid-video measurement method is introduced as a low cost, standard method to measure the small obstacle avoidance performance of A-UGVs. An experiment using blocks demonstrates how the grid-video measurement method can be used effectively to measure the A-UGV obstacle avoidance performance, and it shows that the performance changes upon A-UGV specification, obstacle sizing, and environmental conditions quantitively. The method and experimental results proposed in this paper will be used to support ASTM F45 standard development. |
