The best way to do that is doing CSE:
You put fentanyl only in the spinal and place the epidural catheter and allow her to ambulate.
This should give several hours of analgesia until she gets uncomfortable, then you bolus the epidural and she stops walking.
My advice: Don't mess around with epidural bupivacaine and walking, and better yet don't do walking epidurals.
Have you noticed any problems with decels and fentanyl-only spinals in pregnant women? Last night my attending was discussing how nobody can figure out why solo fentanyl intrathecal causes decels in patients, yet it's less common (but still occurs) when fentanyl is given with bupivicaine. Epidural clonidine supposedly has a risk of 30% of patients developing fetal decels. That makes some sense if clonidine passes to the fetus. I haven't looked any of this up myself, but this particular attending is like a walking encyclopedia of medicine, so I suspect his assertions are well founded. This came up after I suggested doing a CSE on an uncomfortable woman ~5 cm dilated (multiparous) having scattered late decels. My attending wanted me to do a straight epidural so future decels wouldn't be blamed on our spinal injection (I would have done 2.5 mg bupi with 20 mcg fent). The literature suggests this has something to do with fetal head engagement vs disengagement- if it's not engaged, decels are more common. The mechanism of this phenomenon is something of a mystery.
In response to the walking epidural issue- all our patients are on bedrest once the epidural goes in. It makes sense that you would be wide open to a lawsuit if a patient fell with a so-called walking epidural.