An Indoor Low-Cost and High-Accuracy Localization Approach for AGVs

Abstract

Demanding for warehousing automation and intelligence is critical with the rapid development of information and big data teleology. In most storehouses, cargos are mainly carried by Automatic Guided Vehicles (AGVs) or even manpower. Most AGVs follow predefined paths that are normally paved by some marks detected by AGVs. It largely limits the application of AGVs. Some AGVs relies on highly accurate Light Laser Detection and Ranging (LiDAR) devices for navigation. The paper presents a low-cost localization approach for indoor AGVs with the localization accuracy up to centimetres. The proposed method not only allows an AGV to move precisely without any predefined paths, but also reduces the cost largely. The Ultra Wide Band (UWB) technology is used in our approach. In the paper, a gradient decent method cooperated with a least square method is developed to deal with the nonlinearity of UWB ranging data. An optimal localization result is achieved within a small amount of iterations. The positioning results are also compared to a high-accuracy Real-Time Kinematic (RTK) satellite system. Meanwhile, the approach is able to diagnose the original UWB data and will discard any data corrupted by non-ignorable noises. Thus, the robustness is guaranteed.