Vehicle localisation by altitude data matching in spatial domain and its fusion with dead reckoning

Abstract

In this paper, localisation algorithm based on altitude profiles is proposed. It is a kind of landmark-based localisation algorithm in which altitude profile which is obtained while driving a vehicle are treated as landmarks. The author proposes an algorithm for measuring the location of vehicles by using altitude data in locations where GNSS positioning is difficult. Specifically, altitude is estimated from the pressure information obtained from a barometer, in other words, pressure sensor. This is a position estimation algorithm that attempts to identify the current position by matching altitude reference data prepared in advance with altitude data currently acquired. A sensor fusion with wheel pulse speed data is also proposed as an extension to the original altitude similarity matching algorithm to make it more applicable to the cases when variation in altitude is little. The effectiveness of the algorithm is presented with several experimental results obtained with real data.