Large-scale physics experiments running at high interaction rates place a high demand on the data acquisition system (DAQ) responsible for transporting the data from the detector to the storage. The antiProton ANihilation at DArmstadt (PANDA) at the facility for anti-proton and ion research (FAIR) is one such experiment of the future that will not use fixed hardware triggers; instead, the event selection is based on real-time feature extraction, filtering, and high-level correlations. A firmware framework for such real-time data processing has been developed and tested with hardware setup for a PANDA Forward Tracker (FT) prototype. The solution is applicable for other detector subsystems based on the so-called Trigger Readout Board (TRB) data read-out system.