Validation of lane changing rules

The most important issue for lane changing is that the fundamental diagram should remain plausible, i.e. with a maximum flow of about 2000 veh per hour and lane. This is indeed the case both with the above symmetric and the above asymmetric lane changing rules. A fundamental diagram for a simulation with asymmetric rules is in Fig. 17.5; compare this to a fundamental diagram from (German) reality in Fig. 17.5.

Another quantity of interest is the fraction of vehicles in each lane. For the symmetric rules and 2-lane traffic, this should always be at 50%. For the assymmetric lane changing rule introduced above, lane usage is plotted in Fig. 17.5, which was obtained with a look-ahead distance of $d=16$ cells. Fig. 17.5 shows a plot of the same quantities from (German) reality. Additional rules, which can bring the simulations even closer to reality, are discussed by Nagel et al. (90).

Another validation of lane changing rules concerns vehicles that change lanes in order to be in the correct lane for a turn. Two important questions here are how many vehicles do not reach their desired lane, and how much the lane changing disturbs the throughput. The first question is more critical under congested conditions, and one needs a set-up where the intersection capacity is smaller than the link capacity, caused for example by traffic lights. The second question is most critical near maximum flow; for example, one could test if at a traffic light just turned green, outflow is reduced when there is a lot of last-second lane changing.

Figure 17.3: Multi-lane fundamental diagrams. (a) STCA with $v_{max}=5$, $p_{noise}=0.25$. From Nagel et al. (90). (b) Reality (Germany). From Wiedemann, published in Nagel et al. (90).

[]

[]

Figure: Asymmetric lane usage. (a) Simulation. (b) Reality (Germany). [[from wiedemann, published in ...]]

[]

[]