Velomobil-Simulator

[labella-baron]

Zum Verständnis:

• encodeURIComponent konvertiert die Sonderzeichen in der URL
• dann wird das an die URL des Velomobil-Simulators angehängt

Da kommt bei mir auch bei einer kurzen einfachen Strecke immer das heraus:
//christoph-moder.de/fahrrad/vm-simulator.php?brouter=about%3Ablank
Heißt das die Konvetierung schlug fehl, oder woher kommt 'about:blank'

 

[Christoph Moder]

Heißt das die Konvetierung schlug fehl, oder woher kommt 'about:blank'

Vielleicht funktioniert das auf Android einfach nicht. Ich kannte Bookmarklets nur vom PC, und da funktionieren sie meist recht gut. Dein Lesezeichen sieht jedenfalls richtig aus.

Ich habe das vorhin mal auf meinem Android-Telefon ausprobiert, im Firefox-Browser, und da ging es auch nicht. Habe aber nicht näher recherchiert, warum, oder ob es mit einem anderen Browser ginge.

Notfalls musst du es per Hand machen:

• Beispiel: die vorige BRouter-Route
• davon brauchst du nur den hinteren Teil, ab lonlats
• dann ersetzt du:
  • alle "=" durch "%3D"
  • alle "&" durch "%26"
• und hängst das an die VM-Simulator-URL an.

Ergebnis:

christoph-moder.de/fahrrad/vm-simulator.php?brouter=lonlats%3D11.871071,47.987165;11.764984,48.050317;11.822319,48.081973%26profile%3Dvm-forum-velomobil-schnell

 

[Aethelwulf]

Good question; I have to investigate that. There is a Javascript error that the interpolation fails. Maybe there are multiple data points at the same location, and I have to remove all but one.

When I try some more routes now I can barely get anyone to work. They get stuck on the spinning wheel-symbol, maybe because of a javascript error(?). Even when the route is only 100m long on many different locations. Just as an information. Do you think it's better suited with another browser than chrome?

 

[Christoph Moder]

When I try some more routes now I can barely get anyone to work. They get stuck on the spinning wheel-symbol, maybe because of a javascript error(?). Even when the route is only 100m long on many different locations. Just as an information. Do you think it's better suited with another browser than chrome?

It should now work as before. I was just testing some things, and probably made it worse.

Yes, the reason is a Javascript error. No, a different browser does not help; the problem is in my code.

Maybe there are multiple data points at the same location, and I have to remove all but one.

This is the problem. Generating the braking/accelerating curves is complex, and this lead to multiple points at the same location, so the interpolation between them fails (because the distance is 0).
=> Maybe I can fix it tonight.

 

[Riese]

Ich habe das vorhin mal auf meinem Android-Telefon ausprobiert, im Firefox-Browser,

Hab's heute morgen mal im Android Firefox probiert. Per Lesezeichen-Aufruf ging's nicht. Wenn ich aber den Inhalt in die Adresszeile des Tabs mit Brouter kopiere und dann aufrufe bekomme ich ein wildes Diagramm, das richtig sein könnte.
Nur funktioniert der Zoom und scrollen nicht richtig auf Firefox, daher konnte ich es so noch nicht richtig angucken.

 

[labella-baron]

Komme auch nicht recht weiter. Habe jetzt mal händisch was einfaches versucht, nämlich

//christoph-moder.de/fahrrad/vm-simulator.php?brouter=https%3A%2F%2Fbrouter.de%2Fbrouter-web%2F%23map%3D15%2F49.3983%2F11.0130%2Fosmmapnikgerman_style%26lonlats%3D11.023364%2C49.401413%2C3B11.023622%2C49.394739%26profile%3Dvm-forum-velomobil-schnell

Da wird das ohne Kurven gemalt:

 

[labella-baron]

Da wird das ohne Kurven gemalt:

Na ja, was soll auch passieren, wenn sonst nichts angegeben ist !

 

[Christoph Moder]

This is the problem. Generating the braking/accelerating curves is complex, and this lead to multiple points at the same location, so the interpolation between them fails (because the distance is 0).

Ich hoffe, ich habe das jetzt repariert.

 

[Aethelwulf]

I hope I've fixed this now .

It works like a charm now! Is it possible to upload a route (file?) with more elevation points than is included in the brouter segments? At least on my route many peaks and lows are smoothed out because they are both inside a segment.

 

[Christoph Moder]

It works like a charm now!

Great!

Is it possible to upload a route (file?) with more elevation points than is included in the brouter segments?

Not yet, but I thought about it, and it is quite easy to implement. I just have to do it.

 

[labella-baron]

Ich hab momentan einen Knoten im Hirn: Wie wählt man die Muskelleistung? Ähnlich wie im Velomobil-Simulator?

 

[Christoph Moder]

Ich hab momentan einen Knoten im Hirn: Wie wählt man die Muskelleistung? Ähnlich wie im Velomobil-Simulator?

Da gibt es mehrere Möglichkeiten:

• In der gelben Kurve Punkte hinzufügen und ziehen (also in die Nähe der Kurve klicken, dann entsteht dort ein neuer Punkt).
• Wenn die gesamte Linie verschoben werden soll, die Endpunkte ziehen.
• Mit dem URL-Parameter default_power.

 

[Aethelwulf]

Another improvement suggestion would be to somehow change the curve of max speed before a turn or roundabout. If I want to take a turn 90 after a 5km straight the simulator shows did I want to brake the last km approximately.

 

[Christoph Moder]

Another improvement suggestion would be to somehow change the curve of max speed before a turn or roundabout. If I want to take a turn 90 after a 5km straight the simulator shows did I want to brake the last km approximately.

You mean the braking curve should be steeper?

In the first attempt I just assumed a deceleration of 1 m/s^2, because I thought: No one will pedal until directly before the intersection, and then brake hard. You would rather start coasting something like 250 m before, and only brake in the last 30 m or so.

A braking curve with two slopes would make things quite complicated, since braking curves can intersect arbitrarily. So I would like to stick with one single deceleration. What would be realistic (= cause the least amount of manual correction)?

Or maybe set the muscle power to 0 well in advance of the intersection. But this will fail when going uphill (one will come to a halt before the intersection and never reach it).

The problem is that the max speed curve is calculated at the very beginning, before the power calculations are done. So it must be independent of them.

Another possibility would be to somehow adjust the muscle power curve afterwards. E.g. setting it to 0 before points where braking starts from more than e.g. 30 km/h.

 

[Aethelwulf]

You mean the braking curve should be steeper?

In the first attempt I just assumed a deceleration of 1 m / s ^ 2, because I thought: No one will pedal until directly before the intersection, and then brake hard. You would rather start coasting something like 250 m before, and only brake in the last 30 m or so.

A braking curve with two slopes would make things quite complicated, since braking curves can intersect arbitrarily. So I would like to stick with one single deceleration. What would be realistic (= cause the least amount of manual correction)?

Or maybe set the muscle power to 0 well in advance of the intersection. But this will fail when going uphill (one will come to a halt before the intersection and never reach it).

The problem is that the max speed curve is calculated at the very beginning, before the power calculations are done. So it must be independent of them.

Another possibility would be to somehow adjust the muscle power curve afterwards. Eg setting it to 0 before points where braking starts from more than eg 30 km / h.

I guess I would do a little bit of both, coasting maybe 200m if I have good speed and then brake later. But in the picture above it hits maximum speed curve and reduces speed linearly at 1.2 km away from the turn. And that I would not do: D

So yes, in most cases I see in the graph I would like the brake curve steeper. I guess it's individual though. I've just modded the max speed curve to my liking. On a simple route like mine that's not much work at all.

Maybe introduce a choice could be a solution? Braking power: 100-2000W. And maybe speed to keep through 90 degree turn: stop/5-40 km/h or something.

 

[Christoph Moder]

I guess I would do a little bit of both, coasting maybe 200m if I have good speed and then brake later. But in the picture above it hits maximum speed curve and reduces speed linearly at 1.2 km away from the turn. And that I would not do: D

Yes, I noticed that there is something wrong. Braking should not start that early.

=> I found a bug that causes the braking curve to always start at the last point, instead of inserting a new point if the last point is far away. I’ve fixed it.

So yes, in most cases I see in the graph I would like the brake curve steeper.

For now I’ve kept the deceleration at 1 m/s^2; with the bug, the curves were sometimes much shallower.

Maybe introduce a choice could be a solution? Braking power: 100-2000W. And maybe speed to keep through 90 degree turn: stop/5-40 km/h or something.

Yes, I could make these things configurable.

However, specifying the braking power is difficult, because it depends not only on the deceleration, but only on the speed, which is only known after the simulation has run.

 

[Aethelwulf]

That bug-fix surely fixed it all. Now it looks a lot more realistic. Nice work!