SQL & PL/SQL

I am trying to convert the names of attributes that use capitalization sort of like camelcase to distinguish multiple words, e.g. VehicleColor to use underscores instead, eg. Vehicle_Color.

I have a regular expression that does this, however I have a problem when an abbreviation consisting of multiple upper case characters is present, e.g. AverageMPG becomes Average_M_P_G. I am trying to come up with a pattern that only adds the underscores to the first occurrence of a capital letter in a series which should result in the abbreviation MPG becoming Average_MPG.

SQL> select * from v$version where rownum = 1;

BANNER
-----------------------------------------------------------------------------
Oracle Database 11g Enterprise Edition Release 11.1.0.7.0 - 64bit Production

SQL> with test_data as
  2      (
  3      select 'VehicleColor' str from dual union all
  4      select 'WeightClass' str from dual union all
  5      select 'AverageMPG' str from dual union all
  6      select 'HighMPG' str from dual union all
  7      select 'LowMPG' str from dual union all
  8      select 'ABS_System' str from dual
  9      )
 10  select
 11      str,
 12      regexp_replace(str, '([A-Z])', '_\1', 2) result
 13  from
 14      test_data;

STR          RESULT
------------ ------------------------
VehicleColor Vehicle_Color
WeightClass  Weight_Class
AverageMPG   Average_M_P_G
HighMPG      High_M_P_G
LowMPG       Low_M_P_G
ABS_System   A_B_S__System

6 rows selected.

SQL>

These are the results I would like, but I don't know how to modify the pattern to only have the replace act on the first capital letter in a series of capitals or if it is possible.

STR          RESULT
------------ ------------------------
VehicleColor Vehicle_Color
WeightClass  Weight_Class
AverageMPG   Average_MPG
HighMPG      High_MPG
LowMPG       Low_MPG
ABS_System   ABS_System