摘要
测量冠状动脉血管血流储备分数的压力导丝中的超微型压力传感器,具有严重的温度以及非线性漂移问题,且目前只有硬件方法进行补偿。针对硬件补偿方法存在的补偿精度不高及成本昂贵等问题,提出一种改进粒子群优化的BP神经网络算法,对压力导丝进行温度和非线性软件补偿。Matlab仿真结果表明,改进PSO-BP神经网络与现有标准PSO-BP算法、BP神经网络等软件以及硬件补偿方法相比,具有成本低、稳定性高、不易陷入局部最优、泛化能力强等优点。
The subminiature pressure sensor in pressure guidewire used for measuring fractional flow reserve ( FFR ) in coronary artery is provided with serious problems of temperature and nonlinear drifts, and at present, the hardware method is the only way to compensate. Aiming at the demerits existing in hardware compensation method, such as low compensation accuracy and high cost, etc. , the improved PSO-BP algorithm is proposed for realizing temperature and nonlinear compensation with software. The results of Matlab simulation show that comparing with hardware compensation method, and the software compensation methods using PSO -BP algorithm, or using BP neural network, the proposed software method using improved PSO-BP algorithm features many merits, including low cost, high stability, strong generalization capability and avoid getting into local optimum.
出处
《自动化仪表》
CAS
2016年第6期16-20,共5页
Process Automation Instrumentation
基金
广东省科技计划基金资助项目(编号:2012A032200015)
关键词
压力导丝
压力传感器
温度非线性补
Pressure guidewire
Pressure sensor
Temperature
Nonlinear compensation
Nonlinear drift
Particle swarm optimization(PSO)
BP neural network
Matlab
