Positioning of Electric Vehicles for Inductive Charging Systems: Magnetic Field based Localization of the Charging Coil using Trilateration

摘要

The IVK actually develops a new electric vehicle positioning prototype in order to ensure an efficient and safe inductive charging process. The system has been patented and is based on a square-wave magnetic signal which is generated by a coil located in the ground of a parking place. This signal is sampled and evaluated by special magnetic field sensors located at the electric vehicle's underbody. Thereby always two pulse half-waves are correlated in order to calculate the differential magnetic flux density and to map it to the corresponding distance. However, the high inductance of the coil causes a long pulse rise time. This fact can lead to high measurement errors because it can happen that several sampling instances catch this transient interval. This paper shows how the sampling instances can be synchronized with the stationary interval of the pulse half-waves without the need of a broadband wireless communication path between the electric vehicle and the pulse signal source. Further an algorithm is presented which considers the motion of the electric vehicle. The motion is detected by calculating a line trough the sampled stationary magnetic flux density values using simple linear regression. If its slope is above a certain limit, it is used to predict the next magnetic flux density value for the next sampling cycle leading to higher measurement accuracy and better utilization of the low frequency pulse signal. The algorithms test results are provided within this paper. As mentioned above several sensors measure the differential magnetic flux density and map it to a relative distance they have regarding the coll. With these distances the location of the coil can be determined using trilateration. In this context an already available robust closed-form algorithm for the least-squares trilateration problem was implemented and tested on a single-core microcontroller regarding accuracy and runtime. The objective was to find an adequate algorithm parametrization to get a good trade-off between these two criteria.