I have read considerable discussion about the use of automotive O2 sensors
aircraft.  In spite of the poisoning of the sensor by tetraethyl lead many
users report a reasonable life.  The life can be increased somewhat by
mounting the sensor recessed in a cavity instead of being out in the flow
stream.  It will warm up slower and will be less responsive, but that isn't
a concern for aircraft applications.  If one had to replace the sensor every
100 hours at $35 it probably wouldn't be a big deal.

However, the sensors are accurate at only one air/fuel ratio and that is at
stoichiometric.  I would think one effective leaning procedure would be to
lean until the sensor switches state, record the fuel flow and then reduce
the flow by a fixed percentage to arrive at the desired LOP air/fuel ratio.
At stoichiometric the sensor is extremely accurate as it contains a catalyst
that eliminates almost all the effect of CO and NOX concentration.  It
behaves as a high-impedance battery and produces a voltage of about 0.15
volts lean of stoic and 0.85 volts when lean.  A simple comparator set at
about 0.4 volts will accurately record the transition as long as its input
impedance is sufficiently high.  The sensor is more accurate than any
analysis equipment we had a GM so it could be used as a standard.  I'm not
confident in the accuracy of the various "wide range" systems that are being
sold, although maybe they are ok.

Gary Casey