Speaker
Description
Cherenkov Telescope cameras stream about 1 Billion frames per seconds and are dominated by night-sky background, yet the γ-ray air-shower patterns of interest appear only occasionally.
Filtering is thus paramount for guaranteeing that science-grade data are recorded without saturating the downstream read-out.
In this work we present TDSCAN (Trigger Distributed Spatial Convolution Area Network) meets this challenge by replacing full cluster reconstruction with a fixed 3-D neighbourhood vote implemented as a small convolution kernel on the camera’s native hexagonal grid.
The algorithm is deployed on three Kintex UltraScale FPGAs each analyse a third of the camera, and a fourth device merges their votes, collectively handling the 1 Billion frame per second stream in real time while keeping trigger latency safely below one microsecond regardless of event complexity.