信号级协同计算平台架构及应用思考

Signal Level Collaborative Computing Platform Architecture and Application Thinking

针对智能化作战对军事电子信息系统计算能力提升的迫切需求,结合云计算的资源虚拟化、大数据的分布式计算等技术,提出了基于嵌入式CPU+ALL(DSP、FPGA、PPC、GPU、AI处理器等)的异构处理的信号级协同计算平台架构,包括弹性、轻量级异构资源虚拟化模型、分布式实时计算框架和智能计算框架等,形成了一套架构统一、资源共用、使用简便的协同计算和智能计算环境。通过战术级无线电认知和智能信号与信息处理两个典型应用场景,探讨了该信号级协同计算平台可能带来的颠覆性效用。

For the urgent demand of improving computing power of intelligent operations for military electronic information system,by combining cloud computing resource virtualization,distributed computing technologies such as large data,this paper proposes a heterogeneous processing signal-level collaborative computing platform architecture based on the embedded CPU + ALL(DSP processor and FPGA,PPC,GPU,AI etc),including flexible,lightweight virtualization model,heterogeneous resources distributed real-time computing framework and intelligent computing framework,etc.,thus forming a set of unified architecture,resource sharing,simple use collaborative computing and intelligent computing environment.Through two typical application scenarios of tactical level radio cognition and intelligent signal and information processing,the potential disruptive effects of this signal-level cooperative computing platform are discussed.