SAPFIS: a parallel fuzzy inference system for air combat situation assessment

Situation assessment is an important basis for achieving autonomous decision-making in air combat. The ever-increasing multi-source fusion information perceived by situation assessment system poses a computational challenge to current airborne equipment. Fuzzy inference method introduced in situation assessment could effectively adapt to the incompleteness and uncertainty of situational information, but still struggling to meet the high-performance requirements under limited hardware resources on airborne equipment. Leveraging hardware accelerators (GPUs, FPGAs, etc.) to accelerate intensive computation like situation factor evaluation has become paramount. In this work, we present a novel air combat situation assessment architecture with multi-level Parallel fuzzy inference system (SAPFIS), which designs the first-ever fuzzy inference accelerator directed by our proposed situation assessment model to accelerate inference computation. Experimental results show that our fuzzy inference accelerator implemented on FPGA achieves 230.48 times of performance speedup and 2053.76 times of inference efficiency ratio improvement over the software solutions on four general computing platforms(e.g., Intel and Phytium platforms) with multi-dimensional test datasets. Moreover, SAPFIS with 128-core accelerator delivers up to 17.45 times performance improvement in the simulation of air combat situation assessment application.