Triggers won't help you here. There's nothing to make it fire!

What you need is a (scheduler) job. So your process is:

- Create a procedure to transfer the data from the current to the history table
- Submit a job which runs daily/weekly/monthly/... calling that procedure

The procedure could be a simple insert+delete. For example:

create table t (
  x int,
  up_dt date
);

create table t_hist as
  select * from t;

insert into t
  select rownum, sysdate-400+rownum from dual connect by level <= 400;
  
create index i on t(up_dt);

create or replace procedure arch_t is
begin
  insert into t_hist
    select * from t
    where  up_dt <= add_months(sysdate, -12);
    
  delete t
  where  up_dt <= add_months(sysdate, -12);
end arch_t;
/
select count(*) from t;

COUNT(*)  
400       

select count(*) from t_hist;

COUNT(*)  
0         

exec arch_t;

select count(*) from t;

COUNT(*)  
366       

select count(*) from t_hist;

COUNT(*)  
34  

If your table is "large" so you're deleting millions of rows+ on each run you may want to look at partitioning. If you partition the primary table on update_date, instead of a delete can do a "drop partition". This will be faster and generate less redo.

Once you're happy with the procedure, you can schedule it with something like:

BEGIN
    DBMS_SCHEDULER.CREATE_JOB (
            job_name => 'ARCH_JOB',
            job_type => 'STORED_PROCEDURE',
            job_action => 'ARCH_T',
            number_of_arguments => 0,
            start_date => NULL,
            repeat_interval => 'FREQ=WEEKLY',
            end_date => NULL,
            enabled => FALSE,
            auto_drop => FALSE,
            comments => '');
    
    DBMS_SCHEDULER.enable(
             name => 'ARCH_JOB');
END;
/

And Oracle will automatically run it according to the repeat_interval (weekly here).

For more about the scheduler, see:

http://docs.oracle.com/database/121/ADMIN/schedover.htm#ADMIN033
https://oracle-base.com/articles/10g/scheduler-10g