Decoupling instantaneous engine power requirements from instantaneous wheel power requirements does indeed give you a lot more options as to what sort of powerplant can be used.
One reason conventional gasoline engines suck is that they…suck. Literally. They spend most of their time with the throttle plate mostly closed, and part of their power output is devoted to sucking air past that throttle plate. Diesels do better in that regard.
Another reason conventional gasoline engines suck is that they have to use a rich enough mixture to sustain a flame front all the way across the combustion chamber. If it’s too lean, the flame front fizzles out and you get a partial burn, bad for efficiency and emissions. But since you can’t use too much excess air, you get high combustion temperatures which dumps a lot of heat to the cylinder walls, bad for efficiency. Once again, the diesel does better in both regards: it solves the combustion problem by compressing things to the point that they’re so damn hot, any fuel you squirt in there is going to burn, no matter how lean the overall mixture is. And because it’s so lean, peak combustion temperatures are low, and you lose less heat to the combustion chamber walls.
Where diesels suck is in emissions: with so much excess air they produce a lot of NOx, and with incomplete mixing of fuel and air, they produce more soot than gasoline engines.
Best of both worlds? Homogeneous Charge Compression Ignition. Just like a gasoline engine, you mix the fuel and air in the port - but you rely on compression to light it off. You leave out the throttle plate, and control the timing of ignition by how much boost you use, how warm the incoming mixture is (there’s a heater in the intake), and how rich the mixture is. With no throttle plate and lean combustion, you get good efficiency; with premixed mixture, you get low emissions.
But as you might imagine, it’s a bitch to control, and power output can only be controlled within a relatively narrow range. Solution? Reactivity Controlled Compression Ignition. You inject gasoline in the port to create a mixture that’s too lean to ignite on its own when compressed, and then let the cylinder draw it in and compress it. When you finally do want to ignite it, you inject a tiny bit of diesel fuel into the combustion chamber, which ignites quite readily, and in turn lights off the lean gasoline mixture. Now you have efficiency better than a diesel, the emissions of a gasoline engine, and the controllability of a conventional spark ignition system.
I think something like this may be among the most promising technologies to mate with hybrid drivetrains.
This will suffer from the same problems as old-school Wankels: tip seal durability, and a high surface-area-to-volume ratio that results in high unburned hydrocarbon levels in the exhaust.
This isn’t just a conventional 1-cyl engine like your lawn mower: it’s a free-piston engine. With no crankshaft, the piston motion is controlled only by cylinder pressure and actively-managed electromagnetic forces. Power is extracted by the same coils that are managing piston motion, so you don’t even have a separate alternator. Fewer moving parts means less weight and less friction, so yay - but that piston motion is tough to control. If you screw it up, the piston slams into the valves, or into the other end of the bore. someday this could be a good technology to mate up with RCCI, providing the best of both worlds: efficient production of mechanical work through RCCI, and efficient delivery of that work to the wheels/battery through the free-piston mechanism.
These could be good bridge technologies for passenger cars in the next 10-20 years, and they may eventually find a permanent home in long-haul vehicles for which regenerative braking isn’t a huge advantage - but my hope is that battery tech continues to improve to the point that most city-dwelling vehicles are pure electric.