PAVEsense - IRI Correlation: Difference between revisions

From ASAsense Documentation (staging)
← Older edit

Thomas (talk | contribs)

4 user groups
427 edits
VisualWikitext
pavesense v8 eIRI update documentatie
English source text for translation
 
(5 intermediate revisions by 2 users not shown)
Line 1: Line 1:
[[Bestand:Port of Antwerp Groundtruth vs eIRI measurements of ASAsense.jpg|miniatuur|eIRI ASAsense vs IRI reference]]
<languages/>
''Geldig vanaf PAVEsense versie 8 (november 2024)''
<translate>
<!--T:1-->
[[File:Port of Antwerp Groundtruth vs eIRI measurements of ASAsense.jpg|thumb|eIRI ASAsense vs IRI reference]]


De eIRI-waarden van ASAsense worden berekend op basis van het wegprofiel dat wordt opgemeten aan de hand van de accelerometer in de sensorbox. We combineren hierbij de metingen van de verschillende passages en nemen daarbij het '''gemiddelde''' van de eIRI (over de geselecteerde passages) van alle passages die voldoen aan het [[Roads AI#Convergentie|convergentiecriterium]].
<!--T:2-->
''Valid from PAVEsense version 8 (November 2024)''


==== Vergelijking met referentie metingen in de haven van Antwerpen ====
<!--T:3-->
De correlatie werd gedaan met behulp van referentiedata die we van de haven van Antwerpen mochten ontvangen. Met behulp van enkele correcties voor de GPS-drift merken we een correlatie (R² factor) van 70% op. Gezien het feit dat we meten na de demping van de wagen, over verschillende passages heen, terwijl de referentiedata eenmalig wordt opgemaakt en telkens langs één zijde van de wagen, beschouwen we dit als een goede correlatiefactor. Belangrijker nog is echter wat het effect is op de analyse. Er is een zeer goede gelijkenis tussen de referentiedata en de ASAsense data. De slechtere wegen in de referentiedata komen ook in de ASAsense data terug als slecht, de goede als goede.  
The ASAsense eIRI values are calculated from the road profile measured with the accelerometer in the sensor box. We combine the measurements of the different passes and take the '''average''' eIRI of all passes that meet the [[PAVEsense/en#Convergentie|convergence criterion]].
[[Bestand:Distributie eIRI vs referentie.png|miniatuur|eIRI ASAsense t.o.v. PoA distributie]]
Als we de totale distributie van de data bekijken zien we ook een zeer sterke overeenkomst.


We hebben hieronder de “driving quality” indicator ook vergeleken met de referentie metingen, dit analyseren we met een “confusion matrix”, wat een standaard aanpak is voor classificatie problemen. De kwaliteit wordt voornamelijk voorspeld in naburige kwaliteitslabels, wat zichtbaar is in de confusion matrix. Bijvoorbeeld, de wegen die “zeer goed” zijn als echt label, daarvan wordt 96% voorspeld als “goed” of “zeer goed”.
==== Comparison with reference measurements in the Port of Antwerp ==== <!--T:4-->


==== Ontwikkeling ====
<!--T:5-->
The correlation was made using reference data that we kindly received from the Port of Antwerp. After applying some corrections for GPS drift, we observe a correlation (R² factor) of 70%. Given that we measure after the damping of the vehicle and across several passes, whereas the reference data is recorded once and always along one side of the vehicle, we consider this a good correlation factor. More important, however, is the effect on the analysis. There is a very good similarity between the reference data and the ASAsense data: the worse roads in the reference data also come out as bad in the ASAsense data, and the good ones as good.


===== PAVEsense versie 8 =====
<!--T:6-->
Het aligneren van de gemeten data werd verbeterd. Dit heeft als effect dat pieken nu beter samenvallen op de juiste locatie. Hierdoor is het gemiddelde over verschillende passages nemen een effectievere manier om tot het eindresultaat te bekomen. Meer nog, dit is een simpelere manier en werkt ook wanneer er weinig passages over dezelfde locatie zijn. In de volgende tabel is de R^2 afgebeeld voor de nieuwe versie ten opzichte van de vorige. We merken ook dat het voorspellen van rijkwaliteit “''Driving Quality”'', ook significant verbeterd is ten opzichte van de vorige versie. De tabel is gecreëerd met dezelfde data voor v7 en v8, wat vergelijking mogelijk maakt.
[[File:Distributie eIRI vs referentie.png|thumb|eIRI ASAsense compared with the PoA distribution]]
 
<!--T:7-->
When we look at the overall distribution of the data, we also see a very strong agreement.
 
<!--T:8-->
Below we have also compared the "driving quality" indicator with the reference measurements. We analyse this with a "confusion matrix", a standard approach for classification problems. Quality is mainly predicted in neighbouring quality labels, as the confusion matrix shows. For example, of the roads whose true label is "very good", 96% are predicted as "good" or "very good".
 
==== Development ==== <!--T:9-->
 
===== PAVEsense version 8 ===== <!--T:10-->
 
<!--T:11-->
The alignment of the measured data was improved. As a result, peaks now coincide better at the correct location. This makes averaging over different passes a more effective way to obtain the final result. Moreover, it is a simpler method and also works when there are few passes over the same location. The following table shows the R² of the new version compared with the previous one. We also see that the prediction of ''Driving Quality'' has improved significantly compared with the previous version. The table was created with the same data for v7 and v8, which makes comparison possible.
 
<!--T:12-->
{| class="wikitable"
{| class="wikitable"
|'''eIRI version'''
! eIRI version
|'''R2 (dtw-aligned)'''
! R² (dtw-aligned)
|-
|-
|PAVEsense v7 en vroeger
| PAVEsense v7 and earlier
|0.61
| 0.61
|-
|-
|vanaf PAVEsense v8
| from PAVEsense v8
|'''0.7'''
| '''0.7'''
|}
|}
[[Bestand:Driving quality confusion matrix.png|miniatuur|eIRI driving quality confusion matrix genormalizeerd op ware label]]


===== PAVEsense versie 7 en vroeger =====
<!--T:13-->
In PAVEsense v7 en vroeger werd het 85<sup>ste</sup> percentiel gebruikt om verschillende passages samen te nemen tot een meting. Deze versie onder-rapporteerde de gemeten eIRI, waardoor vele wegen in een betere categorie terechtkwamen dan in werkelijkheid.
[[File:Driving quality confusion matrix.png|thumb|eIRI driving quality confusion matrix, normalised on the true label]]
[[Bestand:PAVEsense v7 driving quality confusion matrix.png|miniatuur|PAVEsense v7 (oudere software) driving quality confusion matrix genormalizeerd op ware label]]
 
===== PAVEsense version 7 and earlier ===== <!--T:14-->
 
<!--T:15-->
In PAVEsense v7 and earlier, the 85<sup>th</sup> percentile was used to combine different passes into one measurement. This version under-reported the measured eIRI, so many roads ended up in a better category than in reality.
 
<!--T:16-->
[[File:PAVEsense v7 driving quality confusion matrix.png|thumb|PAVEsense v7 (older software) driving quality confusion matrix, normalised on the true label]]
</translate>

Latest revision as of 12:07, 14 October 2026

Other languages:
eIRI ASAsense vs IRI reference

Valid from PAVEsense version 8 (November 2024)

The ASAsense eIRI values are calculated from the road profile measured with the accelerometer in the sensor box. We combine the measurements of the different passes and take the average eIRI of all passes that meet the convergence criterion.

Comparison with reference measurements in the Port of Antwerp

The correlation was made using reference data that we kindly received from the Port of Antwerp. After applying some corrections for GPS drift, we observe a correlation (R² factor) of 70%. Given that we measure after the damping of the vehicle and across several passes, whereas the reference data is recorded once and always along one side of the vehicle, we consider this a good correlation factor. More important, however, is the effect on the analysis. There is a very good similarity between the reference data and the ASAsense data: the worse roads in the reference data also come out as bad in the ASAsense data, and the good ones as good.

eIRI ASAsense compared with the PoA distribution

When we look at the overall distribution of the data, we also see a very strong agreement.

Below we have also compared the "driving quality" indicator with the reference measurements. We analyse this with a "confusion matrix", a standard approach for classification problems. Quality is mainly predicted in neighbouring quality labels, as the confusion matrix shows. For example, of the roads whose true label is "very good", 96% are predicted as "good" or "very good".

Development

PAVEsense version 8

The alignment of the measured data was improved. As a result, peaks now coincide better at the correct location. This makes averaging over different passes a more effective way to obtain the final result. Moreover, it is a simpler method and also works when there are few passes over the same location. The following table shows the R² of the new version compared with the previous one. We also see that the prediction of Driving Quality has improved significantly compared with the previous version. The table was created with the same data for v7 and v8, which makes comparison possible.

eIRI version R² (dtw-aligned)
PAVEsense v7 and earlier 0.61
from PAVEsense v8 0.7
eIRI driving quality confusion matrix, normalised on the true label
PAVEsense version 7 and earlier

In PAVEsense v7 and earlier, the 85th percentile was used to combine different passes into one measurement. This version under-reported the measured eIRI, so many roads ended up in a better category than in reality.

PAVEsense v7 (older software) driving quality confusion matrix, normalised on the true label
Retrieved from "https://docs-staging.asasense.com/index.php?title=PAVEsense_-_IRI_Correlation&oldid=1136"