智能系统学报
智能繫統學報
지능계통학보
CAAI TRANSACTIONS ON INTELLIGENT SYSTEMS
2015年
4期
592-598
,共7页
雷达成像%压缩感知%近似观测模型:消息传递%置信传播
雷達成像%壓縮感知%近似觀測模型:消息傳遞%置信傳播
뢰체성상%압축감지%근사관측모형:소식전체%치신전파
radar imaging%compressed sensing%approximate observation model%approximate message pass-ing%belief propagation
针对大观测矩阵引起的压缩感知雷达成像高计算量的问题,提出了一种基于近似观测模型的近似消息传递算法,通过引入逆线调频变标算子,压缩感知雷达成像中的大观测矩阵能够得到很好的近似,从而能够有效地运用现有的成熟的去耦合措施来降低压缩感知雷达成像的计算量;同时,通过引入近似的消息传递,新算法获得了不错的收敛速度。理论分析和仿真实验表明,新算法能够实现非完全观测数据的压缩感知雷达成像,与现有的同类压缩感知雷达成像方法相比,其在保证不增加单步迭代计算量的同时,具有更高的收敛速度。
針對大觀測矩陣引起的壓縮感知雷達成像高計算量的問題,提齣瞭一種基于近似觀測模型的近似消息傳遞算法,通過引入逆線調頻變標算子,壓縮感知雷達成像中的大觀測矩陣能夠得到很好的近似,從而能夠有效地運用現有的成熟的去耦閤措施來降低壓縮感知雷達成像的計算量;同時,通過引入近似的消息傳遞,新算法穫得瞭不錯的收斂速度。理論分析和倣真實驗錶明,新算法能夠實現非完全觀測數據的壓縮感知雷達成像,與現有的同類壓縮感知雷達成像方法相比,其在保證不增加單步迭代計算量的同時,具有更高的收斂速度。
침대대관측구진인기적압축감지뢰체성상고계산량적문제,제출료일충기우근사관측모형적근사소식전체산법,통과인입역선조빈변표산자,압축감지뢰체성상중적대관측구진능구득도흔호적근사,종이능구유효지운용현유적성숙적거우합조시래강저압축감지뢰체성상적계산량;동시,통과인입근사적소식전체,신산법획득료불착적수렴속도。이론분석화방진실험표명,신산법능구실현비완전관측수거적압축감지뢰체성상,여현유적동류압축감지뢰체성상방법상비,기재보증불증가단보질대계산량적동시,구유경고적수렴속도。
In order to solve the great computational complexity in compressed sensing radar imaging due to a large observation matrix, an approximate observation model based approximate message passing algorithm is proposed in this paper.It uses inverse chirp scaling operator to approximate the large observation matrix in compressed sensing radar imaging, thereby effectively reducing the computational complexity by the current mature decoupling technolo-gy, and in the meantime, the approximate message passing is used to improve convergence rate.The theoretical a-nalysis and simulation show that compared to the currently used compressed sensing radar imaging methods, the proposed method exhibits higher convergence rate while suffers same computational complexity in each iteration.It realizes compressed sensing radar imaging of non-complete observation data.