Following Friday’s lesson on what vector coordinates really mean, today I’m having a little more of a look through the literature to see what people who have explained the model in the past reckoned the vector figures represent. (Dictating with a cold today, so it’s going to be interesting to see how Dragon performs)
I suppose understand a little more about the vector quantity is using the simulation. All they free vectors, with all vectors measured from the origin? Are they normalised vectors, or with a length equal to one? Or are they an additional type of vector?
From Conrad Parker’s boid page (http://www.vergenet.net/%7Econrad/boids/pseudocode.html), he interprets “the velocity as how far the boid moves per time step”.
(N.b. Parker also has a section on the bounding the position of the flock, to keep the flock within a certain area)
From Computer Graphics, 21(4), July 1987, pp. 25-34., Craig W. Reynolds, “velocity is a vector quantity, referring to the combination of heading and speed.” — which is something I already knew.
From looking at the data, examples of which are below, the figures involved are very small (I have no idea how the formatting is going to come out below, btw …):
Obs Bird_ID Radius Timestep B_Vel_x B_Vel_y B_Vel_z F_Vel_x F_Vel_y F_Vel_z
1094 0 0.3 910 0.47887 0.02008 -0.02601 0.095783 0.003826 -0.005201
1095 1 0.3 911 -0.01961 -0.09468 -0.47016 -0.004031 -0.018935 -0.094028
1096 2 0.3 912 -0.09798 0.10306 -0.01509 0.000000 0.000000 0.000000
1097 3 0.3 913 0.47946 -0.01399 -0.01719 0.095913 -0.001879 -0.003627
1098 4 0.3 914 -0.02006 -0.09467 -0.47014 -0.003922 -0.018935 -0.094032
1099 5 0.3 915 -0.05677 -0.16655 -0.18719 0.000000 0.000000 0.000000
1100 6 0.3 916 0.47891 0.01929 -0.02601 0.095775 0.004017 -0.005201
1101 7 0.3 917 -0.12513 -0.01598 0.10526 0.000000 0.000000 0.000000
1102 8 0.3 918 0.47955 -0.01016 -0.01798 0.095892 -0.002797 -0.003438
1103 9 0.3 919 0.18955 0.05836 0.09798 0.000000 0.000000 0.000000
1104 . . . . . . . . .
1105 . . . . . . . . .
1106 0 0.3 920 0.47888 0.01995 -0.02601 0.095781 0.003858 -0.005201
1107 1 0.3 921 -0.01969 -0.09468 -0.47016 -0.004013 -0.018935 -0.094029
1108 2 0.3 922 -0.09798 0.10306 -0.01509 0.000000 0.000000 0.000000
1109 3 0.3 923 0.47950 -0.01335 -0.01732 0.095910 -0.002032 -0.003595
1110 4 0.3 924 -0.02000 -0.09467 -0.47015 -0.003937 -0.018935 -0.094032
1111 5 0.3 925 -0.05677 -0.16655 -0.18719 0.000000 0.000000 0.000000
1112 6 0.3 926 0.47890 0.01940 -0.02601 0.095776 0.003990 -0.005201
1113 7 0.3 927 -0.12513 -0.01598 0.10526 0.000000 0.000000 0.000000
1114 8 0.3 928 0.47955 -0.01069 -0.01787 0.095901 -0.002670 -0.003465
1115 9 0.3 929 0.18955 0.05836 0.09798 0.000000 0.000000 0.000000
… Where the figures that begin with ” B” are velocities of individual birds, and those that begin with a “F.” are relevant to the flock as a separate entity. From looking at the data, the differences between velocity values are very small for each bird.
Now incredibly, open office does not have a text columns function built in. As such I attempted to download one from the Internet and install it, but it doesn’t appear to have turned up yet. However the plan of attack was to basically average all of the velocities of individual birds in one particular dimension — e.g. b_vel_x — for 10 time steps. After that compare each bird velocity against the calculated average, and see if this would reveal anything. However I think this is one to try tomorrow, as today’s time is up and I’m going to start researching into particle swarm optimisation.
