A lot of the time the problem they face is that, at that time of the year, the riders don't want anything too tough, hence the shorter stages.
The thing is, they do have some obstacles, they just don't use them. They could take some of the smallish climbs they have and put them close to the finish to allow some riders to make attacking worthwhile so that you get the possibility of it not being a sprint.
What I'd be tempted to do, given what ACF's just said about Mount Lofty, would be to do something that wouldn't work in most stage races but because of the limited number of obstacles and easy racing of the TDU might be able to work, is take something from the world of wintersports.
Instead of the altogether extremely tedious Adelaide crit on the final day, do a split stage 6a/6b thing. In the morning, a 10-15km time trial, so that everybody is on separate times... then in the afternoon, a 'catch-me-if-you-can' pursuit, where everybody sets off on their own race time, racing from Adelaide to Mount Lofty - if that's only going to be 15-20km you then have the possibility of a group of riders on close time chasing down those ahead, or the climbers who've lost a minute or two, might be able to stay in a group with some rouleurs and then catch time up on the sprinters (the TT will probably be short enough that most sprinters won't lose too much time and will still be up there in the mix) on the climb... that way, the finishing positions on the stage would also be the finishing positions of the overall Tour.
This wouldn't work in most stage races because the gaps would be too big - imagine doing it in the Tour, when most of the front runners would have finished a 100km stage by the time the last stragglers are allowed to set off. But in a short stage race that doesn't threaten to break the péloton up too much, it could work. It would also be something unique to the TDU, which would be one more defining characteristic than it has now.