The answer is to let them just mass cross, let the insects go where they will, collect the seed, and next season just sow a truckload of seed, and select really heavily. This is a long term project if you are after individual traits - you never know the exact parentage of the plants you are growing, or whether you have homozygous or heterozygous individuals,. You just have to take more time, grow for more seasons, and hope that you selection processes are driving the population in the right direction.
An alternate approach is to not be too hung up on individual traits, and if something is good eating or attractive for some other reason, just select them and continue to grow and eat them.
A chance for a bit of landracing emerged this summer. A pile of Senposai-like individuals sprouted in a bed I was growing melons in. I let them go - the melons were an experiment, so i wasn't trying to maximize the melon crop, so i left the Senposai things alone. I love senposai, a stable cross between asian mustards and asian cabbage that is just a dream vegetable - easy to grow, doesn't bolt, tasty, never woody, and still delightful at the size of dinner plates.
I didn't give them much water over summer and our really dry autumn - some just bolted and got pulled. About half of them then got smashed by cabbage white butterfly caterpillars, but others were untouched. I figure there was a bit of resistance in the untouched ones. This is when my interest was aroused. Then a little bit of powdery mildew came along, closely followed by aphids, which are busy sucking away at the moment. The bed, and the plants look a bit untidy, but I can let the selection process run for a while longer, I don't need that bed at the moment.
A couple of the better survivors, and the whole short trial bed at the bottom. They are looking very Kale-like - I suspect some kale genes might have jumped over the greenhouse from the Tronchuda bed.
