How coarse was my texture? Richard Wix & Roland Leschinski, ARRB Group

Size: px

Start display at page:

Download "How coarse was my texture? Richard Wix & Roland Leschinski, ARRB Group"

Brett Murphy
6 years ago
Views:

1 How coarse was my texture? Richard Wix & Roland Leschinski, ARRB Group

2 Overview of presentation Section 1 Background What is texture? The quandary A solution Section 2 Results A tale of five lasers Outcomes Where to from here? 2

3 Section 1: What is texture? Texture is categorised by wavelength micro macro mega 3

4 How do we measure texture? Project level manual safety issues subjective 4

5 How can we measure texture faster? Laser profiler automated objective repeatable safe Outputs mean profile depth (MPD) sensor measured texture depth (SMTD) 5

6 Texture verification Various sites & equipment expensive Outcome linear relationships R² = 0.91 Observation variation in relationships

7 The quandary 7

8 In a nut shell. There is a lack of a verification procedure that provides an absolute ground truth measurement of texture that is traceable verifiable precise accurate etc. 8

9 A solution. The ARRB texture jig 9

10 Previously.. 10

11 The disks Disk A B - black silver 11

12 Calculating theoretical texture 12

13 Theoretical texture values Disk Radius (mm) MPD (mm) SMTD (mm) A A A B B B

14 Section 2: A tale of 5 lasers Commonly used laser types for texture measurements 78 khz analog 64 khz analog high power 32 khz digital CCD 62 khz digital CCD 16 khz analog (SMTD only) 14

15 Factors affecting texture measurement Operating speed Internal electrical noise of sensor Signal noise due to vehicle vibration Laser spot size Output data rate 15

16 Effect of speed MPD Speed difference Digital 62 khz Digital 32 khz High power analog Low power analog SMTD Speed difference Digital 62 khz Digital 32 khz High power analog Low power analog 16kHz analog

17 Noise specification All lasers have different noise characteristics Published specifications for lasers include: noise bandwidth However, not included in MPD standards Sensor noise indistinguishable from texture signal 17

18 Texture is a random signal Height (mm) mm Samples..similar to noise 18

19 Flat disk noise measurements Profile (mm) Distance (mm) Digital laser 32 khz Low power analog laser 19

Effect of filtering ISO 13473 recommends low pass filtering

20 Effect of filtering ISO recommends low pass filtering Implemented simple low pass moving average filter using 2,3 and 4 samples 20

21 Accuracy of MPD measurements Disk A Error (mm) No filter 2 x avg 4 x avg MPD = PCD Digital 62 Digital 32 Analog Hi power Analog Lo Power 21

22 Accuracy of MPD measurements No Filter 2 x avg Disk B (black) MPD = 1.02 mm Error (mm) x 110 mm PCD Digital 62 Digital 32 Analog Hi Analog Low Power 22

23 Accuracy of SMTD measurements 1 Error (mm) Disk A Disk B-B Disk B-S Disk A SMTD = 6.11 Disk B SMTD = mm PCD Digital 62 Digital 32 Analog Hi Analog Lo Analog 16 khz 23

24 Conclusions Aim is not to make recommendations on a particular laser. Rather, provide method for objective measurement of: noise performance dynamic performance absolute accuracy speed dependency surface colour effect Lasers still need on-road evaluation 24

25 Future work Designing further test disks. Possibly rounded with lower profile Try more realistic light colour Repeat at various heights throughout the laser s range Increase test speed 25

26 Thank you 26

Working towards a revised MPD standard (ISO )

Working towards a revised MPD standard (ISO 13473-1) a sneak-peek on the current mind set Bo Söderling; LMI Technologies Ltd A look back... 1985-1990 laser sensors are established as THE TOOL for road