Online Pattern Recognition for the ALICE High Level Trigger (tracking and compression techniques)

The ALICE High Level Trigger has to process data online, in order to select interesting (sub)events, or to compress data efficiently by modeling techniques. Focusing on the main data source, the Time Projection Chamber (TPC), we present two pattern recognition methods under investigation: a sequential approach (cluster finder and track follower) and an iterative approach (track candidate finder and cluster deconvoluter). We show, that the former is suited for pp and low multiplicity PbPb collisions, whereas the latter might be applicable for high multiplicity PbPb collisions of dN/dy>3000. Based on the developed tracking schemes we show that using modeling techniques a compression factor of around 10 might be achievable.