Localization errors of low-cost GPS devices in railway worksite-like scenarios

Abstract

Many well-known and efficient technologies for satellite-based localization in open space there exist; however, their accuracy and availability depend on several factors, such as the characteristics of the devices used, the surrounding environment, and the distance from reference stations. In the ALARP (A railway automatic track warning system based on distributed personal mobile terminals) system, it is required to accurately localize workers in railway worksites using low-cost GPS devices. In this paper we present an experimental campaign aimed to quantify the systematic and random contribution to localization errors of cheap GPS devices in railway worksite-like scenarios. The evaluation is performed comparing data from lowcost GPS devices to data collected using a highly accurate reference system. Results provide information to understand the feasibility of GPS-based solutions for the ALARP requirements on localization, and constitute a basis for the definition of a further, more extensive testing campaign.

GPS and Electronic Fence Data Fusion for Positioning within Railway Worksite Scenarios

“…The target to be localized was a Garmin 18 LVC [26]. In [9] the experiments were run with no fence information. Thus, we emulate the fence information, mapping the ground truth from the Trimble measurement to the output of a hypothetical fence, placed within the worksite.…”

“…The experimental GPS data and the related measurement process are described in [9]. As the goal of the fusion algorithm is to identify the correct zone in the worksite, we introduce the corresponding performance metrics which we subsequently evaluate for the experimental GPS input data.…”

“…The drawback is the highly expensive price of such systems that range in the order of hundreds or thousands of Euros [8]. Given the above-mentioned premise of a low cost system, the alternative relies on code-phase measurements [9]. These techniques cost few tens of Euros, however, accuracies are often in the order of supra-meter level.…”

Experimental assessment of low-cost GPS-based localization in railway worksite-like scenarios

“…This paper is an extended version of Bondavalli et al [13], where the measurement setup was presented, some criteria about the choice of the experimental conditions were discussed, and some preliminary results were shown, which however did not permit to come out with sound results that may help answer the question above. The paper includes a new set of experimental data, which represent the re-execution of some experiments under different weather conditions, and presents the outcomes of a deep analysis on the previous and the new data (Section 5), which provides some important findings and a quantitative assessment of the behavior of low-cost GPS devices in the considered scenarios.…”

“…As in Bondavalli et al [13], the measurements are performed comparing data from the low cost GPS devices [6,19] candidate for the ALARP system with those provided by accurate DGPS devices [7,8] acting as reference. In the new experiments, five identical GPS devices are used, in order to perform a concurrence analysis of the results [23].…”