Shop floor analysis for industrial human-aware mobile robots 2019-01-1854
The long-established automated guided vehicles work following predefined paths in factories. Currently, more flexible autonomous mobile robots are being deployed in industry, capable of adjusting their routes using simultaneous location and mapping algorithms. To properly adapt to manufacturing environments, where people and vehicles also operate, mobile robots need to comply with the rules that govern their interactions within the shop floor environment. These rules, which can be implicit or not, span from interpersonal distance to critical logistical zones that cannot be obstructed. These rules have not been thoroughly studied in production environments. Since these rules and constraints can change from one factory to another, a systematic approach to collect data was designed in this work. Firstly, data was extracted by surveying factory truck drivers and pedestrians about how they would deploy a mobile robot. Secondly, a portable sensor pack consisting of a light detection and ranging planar sensor and an inertial measurement unit was used to collect information about how trucks and pedestrians move. Finally, an autonomous mobile robot was manually controlled and the sensor readings were stored. This approach was utilised to collect data in a major aerospace factory. The values were then utilised to define the metrics and boundaries that the autonomous mobile robots should respect while navigating through the shop floor. The results allow to develop a human-aware navigation algorithm, either using a model or learning based approach, capable of adapting to the shop floor. In addition to that, the system used to collect data could be further applied in other factories to tune the navigation according to the requirements of different plants.
Marc Auledas Noguera, Amer Liaqat, Ashutosh Tiwari