摘要
针对最小二乘法等传统方法的不足,提出了解决传感器系统非线性校正问题的改进遗传算法。该算法在遗传算法中融入模拟退火算法的思想,将退火温度作为控制参数,通过温度的缓慢的冷却,同时利用模拟退火算法来控制变异和交叉操作,有效地防止了基本遗传算法的早熟收敛。计算机仿真表明,采用这里所提出的改进遗传算法进行传感器非线性校正,其拟合精度远高于传统上应用最为广泛的最小二乘法,并且具有较强的鲁棒性。
Aiming at the flaws of least square algorithm and other conventional methods, the modified genetic annealing-inspired algorithm is proposed to solve the problem of non-linear correction in sensor systems. This modified algorithm, integrated with the idea of annealing-inspired algorithm, with annealing temperature as control parameter and gradual cooling of the temperature, and by using annealing-inspired algorithm to control the variation and interlace operation, effectively avoids the premature convergence of the basic genetic algorithm. Computer simulations indicate that the proposed algorithm, when used in sensor nonlinearity correction, is of much higher fitting precision than the widely-applied traditional least square method, and it is also of comparatively strong robustness.
出处
《通信技术》
2009年第12期156-158,共3页
Communications Technology
关键词
传感器
非线性校正
模拟退火算法
改进遗传算法
sensor
non-linear correction
annealing-inspired algorithm
modified genetic algorithm
