Neural networks trained to guess time allocations for minimum snap problems
The nnet_dts.m file takes as inputs a vector of ranges and angles in the form:
v = [range1;
angle1;
range2;
angle2;
...
rangeN;
angleN];
The function nnet_dts return a set of time allocations:
dt_out = [dt1;
dt2;
...
dt(N-1)];
These time allocations can be used with minimum snap algorithms as an initial guess for the optimal time allocations, instead of using standard time-consuming gradient descents.