A study of L1 track trigger algorithms for the CMS detector at HL-LHC

In order to take advantage of the planned High Luminosity LHC upgrade, CMS tracker along with its L1 trigger must be upgraded. An approach based using Associative memories (AMs) for the CMS L1 trigger has been investigated, by extending current studies to the whole tracker with an updated geometry. Yhe AMsallows to perform track recognition over a coarse-grained representation of the tracker, in which tracker layers are divided in sections called "superstrip" tracks are stored onto banks of patterns and subsequently used by the AM to perform massive parralel pattern recognition, br matching stored tracks with simulated ones in O(200) PU interactions. An original new "flower" design has been developed and compared to the old one, and performance improvements have been demonstrated in the forward region, where the modules are arranged in disks and also in the barrel region wheretilted module are present. Flower design manages to reduce inefficiency in forward region from 5% to 3% ì, while reducing the number of total stubs combinations to pass to fitter stage. Future studies will focus on perfomance with extreme PU events.