Robots with wheels at the ends of the legs, capable of both walking and rolling, are as old an idea as the oldest mecha anime, or even older. More recently, you can see them in many real world uses -- practically all robots during the DARPA Robot Challenge 2015 had both wheels and legs, and the most recent robot revealed by Boston Dynamics, nicknamed Handle, is also a wheeled+legged robot.
Sure, there are problems with it. The legs become much more complex and heavier, there are problems with power transmission, the wheels themselves are often too small to really be efficient, etc. Plus, we really don't have a lot of experience with such robots and the ways they could utilize their hybrid transportation system better.
In this project, I'm going to be experimenting with a number of different configurations for robots that can both roll (to efficiently travel over flat surfaces) and walk (to negotiate obstacles and narrow passages), and maybe sometimes even both at the same time. I will be using the walking robots that I have already built as a base, but I foresee a lot of modifications for them to experimentally see what they are capable of.
Why does it matter? If we want to have robots that actually do useful work for us, and not just stand there bolted down to the factory floor, or beep for help like a stuck roomba, we need to give them means of locomotion that is both energetically efficient and versatile in both of the terrain that can be covered. Hopefully this will lead to a relatively simple and cheap construction that is nevertheless a good base for a home robot.