I wouldn't suggest it. There was a recent post where somebody mentioned the wonderful gas mileage, but the engine blew something like 4 months later. I believe that as an HHO experiment, but the same things happen there with O2 sensor hacking.
If it was that easy (run lean) to boost a cars mileage, it would be a very popular aftermarket flash for the ECU, rather than an extra circuit to change what the ECU "sees".
I didn't mean to start nor do I welcome a debate on whether or not HHO is a good idea. I have done the research on the technology and am taking many precautions to keep it safe. I know the dangers, pitfalls, etc. I will use a temp gauge to make sure my exhaust temp doesn't go too high.
I have successfully completed several standard versions of EFIE's in which I had a schematic to go by; that's really all I was looking for, a schematic for a new type of EFIE design.
Those narrowband lambda sensors are only accurate over a very narrow range; right around stochiometric. Attempting to operate them outside of that range will give you very mixed results. Besides, as those sensors age, they become slower and slower to transition over that 450mV midpoint.
You'd be much better off to replace narrowband lambda sensors with a wideband version, and use a comparator to switch the ECU's input between 0.1v and 0.9v.
An engine running lean burns its valves and pistons instead of the normal amount of fuel, creates a very high amount of the NOX polutant so it will not pass an emissions test and the exhaust stinks and burns your eyes like a car from the 60's.
If you live in Europe where governments rip off people with high taxes on gasoline then replacing the valves and pistons is less expensive that buying more taxed gasoline.