The present invention discloses a neuromorphic processor system (100) for deterministic real-time processing of input data with bounded latency and its method. The system comprises sensor interface circuits (102), an event encoder module (104), neuromorphic compute cores (NCCs) (106), a hypersonic event fabric (HEF) (108), a deterministic scheduling controller (110), an adaptive learning engine (112), a context management and task coordination controller (114), and an output interface circuit (118). The sensor interface circuits (102) is configured to receive multi-modal input signals. The event encoder module (104) is configured to convert the input signals into time-encoded spike event streams. The HEF (108) is configured to: transmit event packets between the NCCs, and perform priority-aware routing and congestion controlled communication to ensure bounded end-to-end latency. The deterministic scheduling controller (110) is configured to assign priority levels to event streams and regulate routing and processing across the hypersonic event fabric based on latency constraints.
Jena et al. (Fri,) studied this question.