# Re: [sumo-user] Can you clarify how density is computed for lanes and edges

```I've tried to make the description more clear in the wiki.
regards,
Jakob

2018-03-07 12:37 GMT+01:00 eegea <ee...@protonmail.com>:```
```
> Hello Jakob,
> thanks a lot, now it is more clear.
> The problem is that I was assuming that sampledSecond is just the number
> of seconds there were vehicles on the lane, not the sum of the number of
> seconds of every vehicle on the lane. I was assuming that if your measure
> interval is, say 10 s, then sampledSeconds<=10 always.  Even though it is
> on the wiki, I think it is not clear enough.
>
> Best regards,
> Esteban
>
>
> Sent with ProtonMail <https://protonmail.com> Secure Email.
>
> ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
> On 7 March 2018 10:10 AM, Jakob Erdmann <namdre.s...@googlemail.com>
> wrote:
>
> Hello,
>
> Consider an aggregation period of 1s. In that case each vehicle on the
> lane results in 1 sampled second so sampleSeconds gives the number of
> vehicles on the lane.
> Since the length of the lane is in m we get the density #veh/m by dividing
> by myLaneLength.
> To get #veh/km we multiply by 1000
>
> For longer periods, if the number of vehicle on the lane stays constant,
> it should be obvious that the sampleSeconds is a multiple of the number and
> the period and the formula is correct as well.
>
> If the number of vehicles changes during the aggregation period, the
> formula gives the appropriate average.
> regards,
> Jakob
>
> 2018-03-06 18:59 GMT+01:00 eegea <ee...@protonmail.com>:
>
>> According to the wiki (http://sumo.dlr.de/wiki/Simul
>> ation/Output/Lane-_or_Edge-based_Traffic_Measures)
>> density is given as "#veh/km"
>> However, looking at the source in MSMeanData_Net line 179 and 208 it is
>> computed as
>>
>>
>> dev.writeAttr("density", sampleSeconds / STEPS2TIME(period) * (SUMOReal)
>> 1000 / myLaneLength)
>>
>> I cannot see why the above gives the number of vehicles per Km in the
>> lane during the period. As far as I see, sampleSeconds is updated as
>> expected
>> sampleSeconds += timeOnLane
>> so it seems to be what the wiki says "Number seconds vehicles were
>> measured on the edge/lane".
>> It seems that we have the fraction of time the lane has been occupied
>> multiplied by 1000  and divided by the lane length.
>> Is this an error or is this coming from some formula I do not know of?
>>
>> _______________________________________________
>> sumo-user mailing list
>> sumo-user@eclipse.org
>> from this list, visit
>> https://dev.eclipse.org/mailman/listinfo/sumo-user
>>
>>
>
```
```_______________________________________________
sumo-user mailing list
sumo-user@eclipse.org