Design of a UAV for Autonomous RFID-Based Dynamic Inventories Using Stigmergy for Mapless Indoor Environments

Abstract

Unmanned aerial vehicles (UAVs) and radio frequency identification (RFID) technology are becoming very popular in the era of Industry 4.0, especially for retail, logistics, and warehouse management. However, the autonomous navigation for UAVs in indoor map-less environments while performing an inventory mission is, to this day, an open issue for researchers. This article examines the method of leveraging RFID technology with UAVs for the problem of the design of a fully autonomous UAV used for inventory in indoor spaces. This work also proposes a solution for increasing the performance of the autonomous exploration of inventory zones using a UAV in unexplored warehouse spaces. The main idea is to design an indoor UAV equipped with an onboard autonomous navigation system called RFID-based stigmergic and obstacle avoidance navigation system (RFID-SOAN). RFID-SOAN is composed of a computationally low cost obstacle avoidance (OA) algorithm and a stigmergy-based path planning and navigation algorithm. It uses the same RFID tags that retailers add to their products in a warehouse for navigation purposes by using them as digital pheromones or environmental clues. Using RFID-SOAN, the UAV computes its new path and direction of movement based on an RFID density-oriented attraction function, which estimates the optimal path through sensing the density of previously unread RFID tags in various directions relative to the pose of the UAV. We present the results of the tests of the proposed RFID-SOAN system in various scenarios. In these scenarios, we replicate different typical warehouse layouts with different tag densities, and we illustrate the performance of the RFID-SOAN by comparing it with a dead reckoning navigation technique while taking inventory. We prove by the experiments results that the proposed UAV manages to adequately estimate the amount of time it needs to read up-to 99.33% of the RFID tags on its path while exploring and navigating toward new zones of high populations of tags. We also illustrate how the UAV manages to cover only the areas where RFID tags exist, not the whole map, making it very efficient, compared to the traditional map/way-points-based navigation.