Genetic Optimization of Fuzzy Adaptive Cruise Control for Urban Traffic - Fuzzy Modeling and Control: Theory and Applications - page 256

Fig. 14.7 Evolution of the minimum cost at each generation, along the algorithm
enhancement in the solution, or a minimum dispersion in the population. In this
work, the algorithm ends after 180 generations.
14.4 Results
As was stated in the previous section, to evaluate a particular fuzzy controller a sim-
ulation is carried out, and the cost value is obtained from the results. For comparison
of the controllers it is necessary that their cost values are calculated under the same
conditions, namely, the accelerations and decelerations of the vehicle ahead are the
same in all the simulations.
Figure 14.7 shows the decrease in theminimumcost in each generation throughout
the algorithm. As can be seen, the cost value of the best individual in the final pop-
ulation is 0.00262. In Table 14.1 , the nine values of the consequents of the rules can
be seen, and Fig. 14.8 shows the corresponding control surface. The best individual
was obtained in generation 161.
Figure 14.9 shows the results of the simulation using the controller obtained with
the genetic algorithm, namely, the best individual in the population of the last gen-
eration. In graph (a) the output of the controller throughout the simulation is shown,
which is applied on the accelerator or on the brake, according to expression ( 14.2 ).
In graph (b), the speeds of the cars can be seen. As is observed, the speed of the
car pursued varies in random steps every ten seconds. Although this is an ideal case,