Here is how I would approach this problem using XML::Twig.

This reads in the XML file, and for each 'Vehicle' element, calls a subroutine named 'vehicle'. This sub opens a new output file whose name is the value of 'AuctionID' (since that seems unique) with a '.csv' extension. It outputs a comma-separated header line, then a CSV line. I only showed 4 values, but you could easily extend the code to add more values, in the order you require. This could be something you can build upon.

The code creates 2 output files: 25019.csv and 25020.csv.

Note: the XML you posted is invalid because it is missing a closing 'Vehicle' tag.

use strict;
use warnings;
use XML::Twig;

my $twig= new XML::Twig(
    twig_handlers => { Vehicle => \&vehicle }
);
$twig->parsefile('in.xml');
exit;

sub vehicle {
    my ($twig, $car) = @_;
    return if $car->first_child('Summary');
    my $file = $car->first_child('AuctionID')->text() . '.csv';
    open my $fh, '>', $file or die "can not open $file: $!";
    print $fh "AuctionID,Ref,VehicleID,Manufacturer\n";
    print $fh join ','
        , $car->first_child('AuctionID'   )->text()
        , $car->first_child('Ref'         )->text()
        , $car->first_child('VehicleID'   )->text()
        , $car->first_child('Manufacturer')->text()
    ;
    print $fh "\n";
    close $fh;
}
[download]

Here are the contents of output file 25019.csv:

AuctionID,Ref,VehicleID,Manufacturer
25019,1480714,156171,TOYOTA
[download]

Since some of your data does contain commas, it might be prudent to also use a CSV-type CPAN module such as Text::CSV_XS.