Anvendelse af Floating Car Data til fastlæggelse af hastighedskort

Abstract

Denne rapport beskæftiger sig med brugen af Floating Car Data som datagrundlag til udformning af hastighedskort over vejnettet til selvkørende biler. Der tages udgangspunkt i trygge og komfortable hastigheder i 18 kurver på tre forskellige veje i åbent land. Vurderingen af disse trygge og komfortable hastigheder foretages primært ved testkørsel af kurverne, hvor hhv. bilfører og passager vurderer hastighederne. Ud fra indsamlede GPS-data kombineres dette med beregning af ryk i kurverne til kvantificering af komfort. Størrelsen af rykkene stemmer godt overens med vurderingerne, når rykkene vurderes acceptable ved værdier lavere end komfortkravet på 0,5 m/s3. Rykkene er forsøgt beregnet ud fra målte accelerationsdata, men kvaliteten af disse data er utilstrækkelige til formålet. Rapporten konkluderer overordnet, at det ikke er muligt at benytte Floating Car Data til dette formål. Hertil skal det nævnes, at kvaliteten af de konkrete data vurderes at være en så stor fejlkilde, at rapporten ikke kan udelukke, at Floating Car Data generelt set vil kunne enyttes til dette formål. Undersøgelserne har til gengæld vist sammenhæng mellem radius og hastigheder for Floating Car Data, som kan beskrives ved en logaritmisk funktion.

This report deals with the use of Floating Car Data as a mean to determine acceptable speeds on the road network. This is perfomed to provide automated vehicles with better information about safe and comfortable speed choices. The report focuses on 18 curves on three different rural roads. The main purpose is to decide whether or not and in what way Floating Car Data can be used to determine safe and comfortable speeds. To determine safe and comfortable speeds a series of test runs is performed. These were obtained with two Leica 1200GPS receivers and an iPhone that measures acceleration. Each test run was performed with a specific speed which was decreased by 5 km/h between each test run. During the test run the driver and a passenger assessed the safety and comfort. In addition to these, measurements of acceleration were used to calculate jerk as another way of determining comfort. Unfortunately, the acceleration data were not sufficiently precise for this purpose and therefore the calculation of jerk was derived from the GPS data. The calculated jerk supported the assessments of the driver and the passenger, which strengthens the validity of the determined safe and comfortable speeds. The determination of safe and comfortable speeds is compared to the available Floating Car Data. The comparison between them did not show any clear correlation when using the median or quantiles to determine a speed level for the Floating Car Data. Further calculations show that using intervals of ± 10 km/h or ± 20% around certain quantiles it is possible to describe close to 70% of the measured curves. However, this result is not sufficient for this purpose. Contrary to this, a comparison between the safe and comfortable speed and the radius of the curves clearly indicate a correlation that can be described by the following logarithmic function where v is speed and R is radius of the horisontal curve. v = 14 · ln(R) The main conclusion based on the available Floating Car Data in this report is that Floating Car Data cannot be used to determine safe and comfortable speeds when using the median or quantiles. However, the report also concludes that the quality of the data makes it impossible to conclude that Floating Car Data in general cannot be used for this purpose.

A master thesis from Aalborg University

Anvendelse af Floating Car Data til fastlæggelse af hastighedskort

[The Application of Floating Car Data to Determine Speed Maps]