You're making the mistake of talking about evolution as though it proceeds in discrete steps. First there is Australopithecus, then it dies off and H. erectus takes its place, then the same with the other intermediate species, and then finally the last one dies off and H. sapiens arises. It doesn't work that way.
Mutations occur continuously in individuals within a species. Occasionally one mutation or a cluster of mutations makes the individuals who have it somewhat better suited to their environment, so they tend to reproduce and survive at a higher rate than the individuals who don't have it. Eventually most of the individuals in the population have those mutations, but it's still the same species. The individuals accept each other as the same species and, in a pack-social species like humans, they trust and care for each other as pack mates. We can't even clearly distinguish the fossilized bones of the mutated individuals from the others. In fact we often don't usually even have intermediate fossils, since the conditions for fossilization are very specific and don't occur as often as paleontoligsts would like. The creationists who ask why we don't have more fossils simply don't understand how remarkable it is that we have any at all.
Another mechanism is the genetic bottleneck. Due to a disaster or sheer luck, of the entire species, in one particular generation only the individuals with the mutation survive. The mutation itself may not even have been a survival advantage, it was just fate. This has happened a couple of times in our species within the past quarter million years. We all have one single male ancestor, Y-Chromosome Adam, and one single female ancestor, Mitochondrial Eve, from two different eras. The descendants of their peers all died out.
This happens several times, and eventually you end up with a different species. There has been a more-or-less slow and smooth transition from one to the other, and it's not possible to say clearly at which point the speciation occurred. The previous unmutated population may still be in existence, but the earlier population lacking the last ten or twenty mutations is long gone.They did survive. They just changed slowly. If you take a snapshot of the "average" human at, say, twenty-thousand year intervals, you will see a series of barely discernable differences. It's hard to tell one from the next, but over the millennia the changes are significant.
Nobody died out. Their descendants were just ever so slightly different.
The average human twenty thousand years ago was substantially shorter than the average human today. But not so short that he would necessarily stand out in a crowd. Keep going back and you find more hair, changes in skull shape, etc. Eventually after you've gone back a quarter of a million years you say, "Hey, these people are definitely not our species." But it's hard to pick a date when that speciation occurred. We're guided by the fossils we have available. There is not a flip-book, with a fossil recovered for every thousand years. We work with what we've got.