Jonas Piccinotti Tutor Amedeo Capozzoli coTutor Claudio Curcio

Jonas Piccinotti Tutor: Amedeo Capozzoli co-Tutor: Claudio Curcio, Angelo Liseno XXIX Cycle - I year presentation Fast GPU implementation of a RCS prediction Tool based on a GO/PO hybrid algorithm accelerated via NUFFT 3 and BVH data structure

Agenda 1 2 3 4 5 Personal background Scientific problem Research activity Products Next year Jonas Piccinotti 2

Personal background Master Science Degree in Telecomunications Engineering with thesis entitled "Mo. M per il calcolo di RCS su GPU", cum laude, december 2011 DIETI group: Electromagnetics – Capozzoli, Curcio, Liseno Worker in the Italian Air Force (Pisa-Rome) Collaborations: IDS Corporation – Ingegneria Dei Sistemi (Pisa) Italian Air Force Flight Test Center (Rome) Jonas Piccinotti 3

Scientific problem 1 Predict the Radar Cross Section (RCS) of an electrically large arbitrarily shaped object within reasonable times and verify the results with experimental measurements. 2 Optimize and accelerate both the prediction and the measurements with hardware and software expedients to fit the tool for optimization problems like antenna placement or evaluation of Low-Observability performances degradation after maintenance on military platforms. Jonas Piccinotti 4

Research Activity (1) Hardware way Accelerating the tool Software way GPU implementation NUFFT 3 3 D BVH Develop and optimize a RCS prediction tool based on a hybrid Geometric Optics/Physical Optics algorithm accelerated by GPU parallel implementation (with Nvidia CUDA language) and Non Uniform Fast Fourier Transform Type 3 3 D (NUFFT 3 3 D) and Bounding Volume Hierarchy (BVH) data structure. Jonas Piccinotti 5

Research Activity (2) Start the validation of the results by comparison with other electromagnetics commercial CADs as well as standard reflectors measurements provided by IDS Corporation. Jonas Piccinotti 6

Research Activity (3) This task required a specific training in IDS to be able to operate the planar scannar (12 m x 8 m) in the semi-anechoic chamber. Jonas Piccinotti 7

Products (1) Promising results were obtained comparing our tool simulations, FEKO simulations and measurements provided by IDS Corporation on standard reflector such as perfectly electric conductor corner reflector and cylinder. Jonas Piccinotti 8

Products (2) The tool showed satisfing outcomes both in accuracy and in speed performances, with the ability to shoot up to 300 milions of rays per seconds. Abovementioned results have been included in a paper entitled “GPU implementation of hybrid GO-PO BVH-based algorithm for RCS predictions” by A. Breglia, A. Capozzoli, C. Curcio, A. Liseno, J. Piccinotti accepted fot the presentation at “ 2015 IEEE International Symposium on Antennas and Propagation and North American Radio Science Meeting” Vancouver, British Columbia, Canada, 19 -25 July 2015. Jonas Piccinotti 9

Next year Credits Summary table: Total 6 bimonth Check 5 bimonth Credits year 3 Estimated 4 bimonth Estimated 3 bimonth Check 2 bimonth 8 8 Summary 1 8 8 4 13 20 - 40 10 0 5 - 10 5 8 7 47 10 - 35 45 8 11 60 60 Modules 20 9 Seminars 5 Research 35 8 8 60 17 8 2 Credits year 2 bimonth Estimated Credits year 1 Check 1 10 - 20 5 - 10 30 - 45 0 0 60 0 - 10 40 - 60 13 0 47 60 Seminars and courses: TBD Jonas Piccinotti 10

Next year 3 Prosecute developing and optimizing the Tool implementing NUFFT 3 3 D and apply the tool to more challenging and complex object of interest for the Ministry of the Defense (e. g. : fighter jet mock-up AT 2000 provided by IDS Corporation) Jonas Piccinotti 11