| 摘要: |
| 精密雷达伺服转台传动机构会随着装备不断运行使用逐渐磨损,表现为齿隙随着机构的磨损逐渐增大。传统双电机消隙控制策略能够消除齿隙,但该策略需要基于控制经验及装备初始传动机构齿隙一次性配置完成,这会导致随着机构磨损消隙效果逐渐变差,影响雷达跟踪精度。针对此缺陷本文提出一种基于深度卷积神经网络(DCNN)的精密雷达伺服转台消隙策略,通过采集位置闭环传动轴振动数据,利用连续小波变换(CWT)得到时频图,作为DCNN训练输入,训练后得到识别模型,最后根据模型识别出伺服转台传动机构磨损程度来调整双电机消隙控制的偏置电流和拐点电流,通过对比实验验证了调整后消隙效果优于传统消隙方式,极大提高装备运行的可靠性,降低雷达伺服转台的维护成本。 |
| 关键词: 深度卷积神经网络 精密雷达伺服转台 双电机消隙 可靠性 |
| DOI： |
| 分类号:TN957 |
| 基金项目: |
|
| Radar Servo Turntable with Anti-backlash Method Based on Deep Convolutional Neural Networks |
|
|
|
|
| Abstract: |
| The transmission mechanism of the precision radar servo turntable will gradually undergo wear during continuous equipment operation, resulting in an increase in backlash. While the traditional dual motor backlash elimination control strategy can alleviate this issue, it depends on control experience and initial gear backlash configuration, leading to a gradual decline in the effectiveness of backlash elimination as the mechanism wears down and impacting radar tracking accuracy. To overcome this limitation, this paper proposes a precision radar servo turntable backlash reduction strategy based on deep convolutional neural network (DCNN). By collecting vibration data from the position closed-loop transmission shaft and utilizing continuous wavelet transform (CWT) to generate time-frequency graphs for training a recognition model. Subsequently, using this model to identify the degree of wear in the servo turntable transmission mechanism and adjust bias current and inflection point current for dual motor anti-backlash control. Comparative experiments confirm that post-adjustment anti-backlash effects are superior to those achieved through traditional methods, significantly enhancing equipment reliability and reducing maintenance costs for radar servo turntables. |
| Key words: Deep convolutional neural networks Precision radar servo turntable Dual motor backlash elimination reliability |