Folks,
For the first time in so many days, I can post something without some hesitation - I'm qualified to post
The frequency of the signal is 100 Hz - Do what you want, measure the frequency - It will be constant 100 Hz. peak voltage remains the same. What changes is the width of the pulses. If the output is low, the pulses turn skinny, but keep repeating at 100 times a sec - 100 Hz. Take the signal higher, the pulse width increases - pulses go fatter, ever at 100 Hz.
The most effective way to interpret this signal easily is through averaging it using a simple leaky capacitor - look into a digital electronics book for basics, or google. it would take a diode, at most two resistors and a capacitor. The result can then be seen across the capacitor by using a simple voltmeter. it will show something between 0.5V to the peak (12V in this case), depending on what the limiting pulse widths (duty cycles) in % are.
Have fun. If you want the component specs and circuit diagram (hardly anything there, but I would make one and check to be sure), just post the limiting duty cycles (would require some sleuthing / a waveform check in limiting conditions on an oscilloscope)
If, on the other hand, the multimeter handles the signals in an averaging manner (Inertia of the analog multimeter needle would mostly average out the pulses) it would give an average, proportional signal directly, in V. Tying it to the MPG rating is a different matter altogether, depends on what relation the pulse widths have to the exact MPG value at the instant.
If nothing, it would give a much needed feedback about instantaneous fuel consumption. Not bad for FREE!