Seems that a config of: <vType id="normalVehicle" length="4.00" maxSpeed="40.00" guiShape="passenger" speedFactor="1.3" tau="3" sigma="1.65" accel="4.6" minGap="1.5"/>
(as well as the hardcoded modification to MSCFModel_KraussOrig1.cpp setting vSafe -= (oldV - vSafe) * 3) Visually looks really good, I get a 'standing wave' of vehicles slowing down, but vehicles leaving the wave go on to resume what looks like normal behaviour and gaps. Quoting "v.f.baines" <v.f.bai...@bath.ac.uk>: > Hi, > > OK just two bits to check then, when you say speed reduced by > accel*sigma*random - you definitely mean accel? Just seems strange, > I thought that value would just be used in increasing speed rather > than decreasing. > > And when you say high values, how high is high? In the docs it says > sigma is between 0 and 1? So it can be bigger than that you mean? > And for accel, high but feasible? Or just super high? Looks like > default is 2.6, something like 6.5 is feasible and 3 times bigger if > thats high enough? > > Will experiment around with some numbers anyway, I just dont want to > accidentally break validity of models > > Thanks > > <div>-------- Original message --------</div><div>From: Jakob > Erdmann <namdre.s...@googlemail.com> </div><div>Date:24/02/2014 > 13:13 (GMT+00:00) </div><div>To: vb...@bath.ac.uk </div><div>Cc: > sumo-user <sumo-user@lists.sourceforge.net> </div><div>Subject: Re: > [sumo-user] Creating congestion waves and measuring them </div><div> > </div>Hello, > I have to retract an earlier statement due to a review of the > computations involving sigma. > > in the default car-following model (Krauss) there is no concept of > overbraking. Vehicles will always brake just hard enough to avoid a crash. > > This was wrong. The speed of a vehicle is reduced by sigma * accel * > random in every simulation step regardless of whether the vehicle is > braking or accelerating (accel is the value of the vType attribute > accel, random is from [0,1]).This means that a braking vehicle will > sometimes drive slower than it needs to. > The fact that it scales with accel is up to debate but you should be > able to see a wave of increasing deceleration if you use high values > of sigma and accel. > > Vehicles should never crash as long as tau > step-size and unless > you play with speed mode or lane change mode. > If you really want a crash you can set the speed mode of a vehicle > so as to ignore maximum deceleration and then braker harder than it > normally could (and thus harder than the follower expected). > regards, > Jakob ------------------------------------------------------------------------------ Flow-based real-time traffic analytics software. Cisco certified tool. Monitor traffic, SLAs, QoS, Medianet, WAAS etc. with NetFlow Analyzer Customize your own dashboards, set traffic alerts and generate reports. Network behavioral analysis & security monitoring. All-in-one tool. http://pubads.g.doubleclick.net/gampad/clk?id=126839071&iu=/4140/ostg.clktrk _______________________________________________ sumo-user mailing list sumo-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sumo-user
