改进基于FMCW雷达的微多普勒特征的人类活动分类,以噪声的影响
NgocBinh Nguyen1, MinhNghia Pham1, Van-Sang Doan2
1Faculty of Radio Electronics Engineering, Le Quy Don Technical University, Hanoi, Vietnam.
PloS one
|August 1, 2024
概括
这项研究引入了一种新的无声化算法和一个基于雷达的人类活动分类的交叉残余卷积神经网络 (CRCNN). 该方法通过从雷达信号中去除白色高斯噪声来显著提高分类准确性.
科学领域:
- 信号处理 信号处理
- 机器学习 机器学习
- 人类活动识别 人类活动识别
背景情况:
- 使用雷达传感器进行人类活动分类对于医疗保健,安全和救援行动至关重要.
- 原始雷达信号通常含有白色高斯噪声,降低特征提取和分类可靠性.
- 与基于雷达的方法相比,现有的视觉感知和可穿戴设备等方法具有局限性.
研究的目的:
- 开发一种强大的方法来消除被白色高斯噪声破坏的原始雷达信号.
- 提出一个新的深卷积神经网络 (DCNN) 准确的人类活动分类后denoising.
- 评估拟议的无声化算法和DCNN模型的有效性.
主要方法:
- 作为一个预处理步骤,应用了一种新的无声化算法,以从原始雷达信号中去除白色高斯噪声.
- 一个轻量级的交叉残留卷积神经网络 (CRCNN) 具有可适应的交叉残留连接被设计用于分类.
- 取消算法参数 (范围-bin 间隔=3,切割值=3) 已被优化以获得最佳性能.
主要成果:
- 无声化算法有效地消除了白色高斯噪声,并确定了最佳参数.
- 在无声化后应用的CRCNN模型,与使用噪音数据相比,提高了高达10%的人类活动分类准确性.
- 拟议的CRCNN模型与其他六个最先进的DCNN相比,表现优越.
结论:
- 拟议的消噪算法有效地提高了原始雷达信号质量.
- 该CRCNN模型在基于雷达的人类活动分类准确度方面取得了重大进展.
- 这种综合方法为需要识别人类活动的基本应用提供了更可靠的解决方案.
相关概念视频
01:05Doppler Effect - II
3.4K
The Doppler effect has several practical, real-world applications. For instance, meteorologists use Doppler radars to interpret weather events based on the Doppler effect. Typically, a transmitter emits radio waves at a specific frequency toward the sky from a weather station. The radio waves bounce off the clouds and precipitation and travel back to the weather station. The radio frequency of the waves reflected back to the station appears to decrease if the clouds or precipitation are moving...
3.4K
00:56Doppler Effect - I
3.5K
The Doppler effect and Doppler shift were named after the Austrian physicist and mathematician Christian Johann Doppler in 1842, who conducted experiments with both moving sources and moving observers. Consider an observer standing on a street corner, observing an ambulance with a siren sound passing by at a constant speed. The observer experiences two characteristic changes in the sound of the siren. Initially, the sound increases in loudness as the ambulance approaches and decreases in...
3.5K