摘要
在系统中引入时滞,相应的动力学系统成为非线性时滞动力系统,它的解空间是无限维的,具有更加丰富的动力学行为。目前时滞系统的研究多为理论和数值仿真。时滞实验方面,国内曾对时滞滤波器抑制起重机载荷残留摆动的理论进行实验验证,国外的学者则将时滞实验的重点放在对于时滞诱发双耦合非线性电路和双耦合热学光学振子的振幅死区的研究方面。文献[4]采用摄动增量法研究了时滞双向联想记忆神经网络由时滞诱发的周期解和完全同步周期解,给出了较为准确的理论分析结果。在这一理论背景下,本文设计了基于MAT-LAB/Simulink的数值仿真模块图,以及基于TMS320F2812芯片的时滞神经网络电路实验,结果表明数值仿真和实验结果与理论分析结果相吻合,从而进一步验证了摄动增量法理论分析的准确性。
In a delayed nonlinear dynamical system,the solution space is infinite dimensions,so the delayed system always has much richer dynamical behaviors.By now,a delayed system is mostly studied theoretically and numerically.For the delayed experiment,the delayed filter technique is applied to the anti-sway control of cranes with load and is shown to significantly eliminate load residual sway interiorly.Meanwhile,most researches abroad are interested in the delay-induced amplitude death experiments of a pair of thermo-optical oscillators and a pair of nonlinear electronic circuits.In this paper,the numerical simulation module diagram based on MATLAB /Simulink software and the delayed neural network circuit experiments based on digital chip TMS320F2812 were introduced to justify the theoretical analysis that delay-induced periodic solution and synchronization periodic solution derived from Hopf bifurcation in delayed BAM neural network were studied by the perturbation incremental scheme(PIS).The numerical simulation and circuit experiment results are in good consistence with theoretical analysis.Further,our work is successfully to verify the validity of the PIS.
出处
《力学季刊》
CSCD
北大核心
2011年第3期376-385,共10页
Chinese Quarterly of Mechanics
基金
国家自然科学基金重点项目(11032009)
上海市重点学科资助项目(B302)
关键词
时滞神经网络电路实验
双向联想记忆神经网络
Simulink数值仿真
time delay neural network circuit experiment
bi-directional associative memory neural network
Simulink numerical simulation