I use several methods.
First pdf can be vector or image format. That depends on the software that created it and options used.
If it is image format I use irfanview with the full plugins which support pdf reading. Save it to tif and then in raster design (autodesk) I geo-reference it and save out as geo-tiff.
Then it will work as georeferenced in any of my software.
If it is vector then I have a couple of options now.
the cheapest is scan2cad v8 pro it will allow export to dxf from many image formats and pdf.
The more expensive option is terrago publisher.
It allows direct load of a pdf into arcmap.
The pdf can then be geo-referenced and "hand digitized" into gis data.
In the realm of opensource there is always grass and gdal which have some functionality in converting images into vector. you need a "usable" (normally binary) image for completion of this method.
New in AutoCad: You can use the insert pdf underlay command, move the pdf to the correct location, scale and rotate. Use the pdfimportsettings to determine how to convert. Then use pdfimport to convert the pdf (must be a vector pdf file) into AutoCad lines. Prior to using pdf import you can use pdfshxtext to convert recognizable text to AutoCad text. and the txt2mtxt command will convert those single line text to multi-line.
Autodesk has some help in the knowledgebase here...