摘要
提出了用人工神经网络压力传感器阵列实现机器人手触觉功能的方法。对所设计的神经单元电路运用计算机模拟分析，结果表明，神经单元的特性符合稳定性的要求。以１１×１１ 四突触神经网络阵列与六突触神经网络阵列为例，分析其稳定性、联想性、鲁棒性，对其触觉信号的输入进行灰度图像处理。从中可以直观地看出，触觉神经网络阵列比非神经网络阵列具有更明显的柔性和灵活性，更近似于人手的触觉。
Tactual sensation arrays of robots is realized by means of artificial neural networks. Simulations show that the characteristic of the neuron is stable. Taking the four connector array and the six connector array as examples, the stability, association and robustness of neural networks are tested and analyzed. From the ash picture changing results, the artificial neural arrays are found to be more flexible than non neural networks.
出处
《微电子学》
CAS
CSCD
北大核心
1999年第4期258-261,共4页
Microelectronics
