I am trying to process a dataset from GEO.

rawdata <- getGEO("GSE32992", GSEMatrix =TRUE, getGPL=FALSE)

class(rawdata)
------------------------------------------------------------------------------------------------
"ExpressionSet"

The problem is when I try to normalize it:

oligo::rma(raw_data)
----------------------------------------------------------------------------------
#Error in (function (classes, fdef, mtable)  : 
  #unable to find an inherited method for function ‘rma’ for signature ‘"ExpressionSet"’

Is not working even:

oligo:rma(raw_data, normalize=FALSE)
-----------------------------------------------------------------------------------
#Error in (function (classes, fdef, mtable)  : 
 #unable to find an inherited method for function ‘rma’ for signature ‘"ExpressionSet"’

What can I do with my object to trasform it to a proper one?

When you use getGEO, you by default get the summarized data that were uploaded to GEO. If you want the raw data, you have to use getGEOSuppFiles. In addition, the oligo package is intended for use with Affy and Nimblegen arrays, and the Capitalbio Human 22K oligo array v1.0 doesn't seem to be either. So even if you got the raw data you probably wouldn't want to use oligo. Looking at the information for one of the samples, it looks like that's some sort of two color array, in which case you would probably want to read raw data in using the limma package, after determining if there is a pre-existing method (or more likely using the generic method and specifying which columns to read in).

Or you could take the easy way and just use the data you have in hand, after reading on GEO how it was processed, so you know how to proceed from here.