Radioactive decay is a small energy source compared to fission. Look at the binding energies [1]: the energy accessible by fissioning heavy elements is about 200 MeV -- 1 MeV per nucleon. Radioactive decays are two orders of magnitude less, or smaller.
Pu-238 is one of the nicest ones because it is "clean" -- emitting only alpha radiation, which is easy to shield. You can transmute on the order of 1 atom of it per 100 fission events [2]. A fission event is 200 MeV; a Pu-238 decay is 5.6 MeV. So at steady state, you can support less than 1/3,000th as much Pu-238 power as you can fission power. If you have a Pu-238 source, you also have a vastly larger nuclear fission power source, so why not just use that?
More broadly, it's not a likely idea on physics, even if you accept all radioisotopes (difficult-to-shield gamma emitters) and not just "nice" ones. The high-Z radioactive decay chains end quickly at the element lead [3], so only release a fraction of the energy that is accessible by fission. (Look back to the binding energy curve [1]). And many of these decays are unusably, geologically slow. The naturally-occurring radioisotopes have half lives of >10^8 years (otherwise they wouldn't exist anymore -- they were created in ancient history in supernovae). And trying to transmute them doesn't improve much -- i.e. in a fission reactor, the most common [4] high-Z product is Pu-239 (half life >10^4 years), whose next decay is to U-235 (half life >10^8 years). Others are similar. So I don't see any way forwards.
[1] http://en.wikipedia.org/wiki/File:Binding_energy_curve_-_com...
[2] Through the several processes which create Np-237 from uranium -- a relatively rare process [4]. Np-237 can be isolated and transmuted to Pu-238 separately through neutron irradiation
[3] http://en.wikipedia.org/wiki/File:Radioactive_decay_chains_d...
[4] Small table here: http://en.wikipedia.org/wiki/Minor_actinide
However, it is likely that a "home" device could be made simpler/less expensive.
People don't like things called 'nuclear'.
Those are 10 watt Sr90 RTGs, sitting around. The used to use them for lighthouses.
It's theoretically feasible, but the cost would be, ah, prohibitive.
However plutonium would not be easy to put in every house since you have to make it in nuclear reactors. You get about half a kilowatt per kilogram of it, I don't think production of it could scale that high.
Strontium is naturally occurring and relatively common. Sr 90 is a trace isotope though. Off the top of my head I would suspect you would be facing similar challenges that separation of Uranium isotopes face.
* Radioisotope power sources (like this, Cassini, New Horizons, etc.)
* Critical fission reactors (popular in the early space race before solar PV, still the best option for high-power deep-space missions like (cancelled) Jupiter Icy Moons Orbiter/Project Prometheus)
* Nuclear pulse propulsion -- setting off nuclear weapons, riding the shockwaves (e.g. Project Orion)
Pulse prolusion is banned, quite pointlessly, under the Partial Test Ban Treaty of 1963. (Note it extends not only to atmospheric explosions, but ones in deep space). The others aren't. Contained, critical nuclear reactions are legal -- NASA chose a nuclear reactor a few years ago for the Jupiter Icy Moons Orbiter, before they cancelled it. And no one cares about radioisotope sources except confused hippies.
http://en.wikipedia.org/wiki/Partial_Nuclear_Test_Ban_Treaty
Obscure history: a running Soviet nuclear reactor crashed from space into the Canadian wilderness in 1977. In case anyone's curious of the consequences.
More info here: http://www.motherjones.com/politics/1997/09/cassini-controve...