Thanks! Have implemented that, trying alpha values up to '3' currently, along with a sigma of 1.
There *seems* to be more braking (and visually seems perhaps more realistic), but I'm not sure I've managed to yet trigger real waves. I've added a measurement to capture 'amount of braking per second' across all the vehicles, as maybe that will give a quantified view of things. Quoting Jakob Erdmann <namdre.s...@googlemail.com>: > Hello, > take a look at MSCFModel_KraussOrig1::moveHelper in > microsim/cfmodels/MSCFModel_KraussOrig1.cpp. This function is also used by > the default model. > I would put in something like: > if (vSafe < vOld) vSafe -= (vOld - vSafe) * alpha > Which would cause a vehicle to overbrake by alpha. > regards, > Jakob > > > 2014-02-17 <vb...@bath.ac.uk>: > >> Hi Jakob, >> >> Ah great, I'll play around with sigma now then. >> >> Regarding model braking, I've had a quick look, is this something in the >> vstop and vsafe methods, in the various cpp files in microsim/cfmodels? Any >> pointers for where to modify appreciated! >> >> Thanks >> >> >> Quoting Jakob Erdmann <namdre.s...@googlemail.com>: >> >> Hello, >>> in the default car-following model (Krauss) there is no concept of >>> overbraking. Vehicles will always brake just hard enough to avoid a crash >>> (which probably is quite unrealistic). However, the Krausss model does >>> have >>> congestian waves due to imperfect acceleration (parameter sigma > 0). You >>> should see a ripple of vehicles braking and due to the nature of imperfect >>> accelerations, this ripple should travel backwards and remain stable for >>> quite a while. >>> Since there is no overbraking the ripple will only get weaker instead of >>> stronger though. So if you want to have jams, you do have to make on >>> vehicle brake quite hard for a start. >>> regards, >>> Jakob >>> >>> PS: modifying existing car-following models is not terribly hard because >>> the part of the code is quite modular (there are already 9 models >>> implemented). Braking, harder than necessary would probably affect 1 line >>> of code and we might eventually add this feature now that I think about >>> it). >>> >>> >>> 2014-02-15 <vb...@bath.ac.uk>: >>> >>> Hi, >>>> >>>> I'm still having troubles with this, more in just creating a >>>> congestion wave / tailback in the first place. I've got a fairly busy >>>> looking flow on a motorway section, and set change lane mode to be >>>> only for strategic. The bulk of vehicles are standard passenger, which >>>> I've defined as: >>>> >>>> <vType id="normalVehicle" length="4.00" maxSpeed="40.00" >>>> guiShape="passenger" speedFactor="1.3" tau="2"/> >>>> >>>> And I'm inserting into fast lane at 40, next lane at 35, and other >>>> lane at 28. So, there's a mix of speeds, going quite fast. >>>> >>>> I have one vehicle selected which I have pre-planned to brake hard - >>>> for about 6 seconds it receives traci command to change speed to 1m/s, >>>> and then resumes to -1. Somewhat exagerated, but I was aiming for that >>>> sort of 'sudden braking causes a ripple of other vehicles braking' >>>> behaviour. >>>> >>>> The problem is.. the rest of the vehicles are just too good! It seems >>>> to me on the roads in the UK, if someone went from 40m/s to 1m/s there >>>> would be someone behind not paying attention, and either crash, or >>>> have to brake even harder.. >>>> >>>> So I'm wondering, is it just that I need to experiment with tau, max >>>> decel? Does sigma play some part? Reduce mingap? Or, is such behaviour >>>> just not possible with current model(s)? >>>> >>>> Thanks! >>>> >>>> Quoting "v.f.baines" <v.f.bai...@bath.ac.uk>: >>>> >>>> > Hi All, >>>> > >>>> > I'd like to implement some sort of repeatable congestion / traffic >>>> > wave scenario. I've created a decent volume of background traffic >>>> > using flows. If a vehicle now brakes suddenly or moves lanes >>>> > suddenly (typical cases I can think of which seem to start ripple >>>> > waves) should/would this impact the flow of the rest of the vehicles? >>>> > >>>> > I'm guessing this sort of scenario might have already been >>>> > implemented a few times so any pointers welcome. >>>> > >>>> > Also, any best practise for measuring the impact? At the moment ive >>>> > placed 5 detectors along the route which give a measure of vehicle >>>> > speed, so in congestion I'd expect this to drop. I guess time >>>> > between detections x detected speed would indicate vehicles gap >>>> > (assuming no lane change), small gap indicating congestion? >>>> > >>>> > Again, any suggestions welcome >>>> > >>>> > Thanks! >>>> > >>>> ------------------------------------------------------------ >>>> ------------------ >>>> > Managing the Performance of Cloud-Based Applications >>>> > Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. >>>> > Read the Whitepaper. >>>> > >>>> http://pubads.g.doubleclick.net/gampad/clk?id=121051231&; >>>> iu=/4140/ostg.clktrk >>>> > _______________________________________________ >>>> > sumo-user mailing list >>>> > sumo-user@lists.sourceforge.net >>>> > https://lists.sourceforge.net/lists/listinfo/sumo-user >>>> >>>> >>>> >>>> >>>> >>>> ------------------------------------------------------------ >>>> ------------------ >>>> Android apps run on BlackBerry 10 >>>> Introducing the new BlackBerry 10.2.1 Runtime for Android apps. >>>> Now with support for Jelly Bean, Bluetooth, Mapview and more. >>>> Get your Android app in front of a whole new audience. Start now. >>>> >>>> http://pubads.g.doubleclick.net/gampad/clk?id=124407151&; >>>> iu=/4140/ostg.clktrk >>>> _______________________________________________ >>>> sumo-user mailing list >>>> sumo-user@lists.sourceforge.net >>>> https://lists.sourceforge.net/lists/listinfo/sumo-user >>>> >>>> >> >> >> ------------------------------------------------------------------------------ Managing the Performance of Cloud-Based Applications Take advantage of what the Cloud has to offer - Avoid Common Pitfalls. Read the Whitepaper. http://pubads.g.doubleclick.net/gampad/clk?id=121054471&iu=/4140/ostg.clktrk _______________________________________________ sumo-user mailing list sumo-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sumo-user
