| 摘要: |
| 基于超声波的手势识别系统中引入手势检测方案一方面可以在无手势的时候不执行高复杂度的识别运算，有效地降低功耗；另一方面还可以分割出有效时间片段，提高手势识别的精度。为提升手势识别系统的性能，提出了一种手势检测方案并且以硬件实现了相应功能。采用动目标与静目标双重检测以及杂波图技术，提供手势识别系统在低功耗的检测模式以及高精度运算的识别模式间切换的条件，实现对环境静态杂波的跟踪和滤除，同时解决了传统单纯动目标检测在连续手势操作时存在的漏检问题。本方案由检测模式向识别模式切换的漏警率为0，识别模式向检测模式切换的虚警率为2 |
| 关键词: 超声波雷达 动目标检测 静目标检测 手势检测 杂波图 |
| DOI：10.20079/j.issn.1001-893x.220401003 |
|
| 基金项目:深圳市科技计划项目（JCYJ20200109143003935） |
|
| Gesture detection scheme and hardware design for dual detection of moving and static targets |
| XU Yunlan,LI Xiangyu |
| (1.School of Integrated Circuits,Tsinghua University,Beijing 100084,China;2.Research Institute of Tsinghua University in Shenzhen,Shenzhen 518071,China) |
| Abstract: |
| Introducing gesture detection scheme into ultrasonic-based gesture recognition system can reduce power consumption effectively without performing high complexity recognition operators when there is gesture.On the other hand,it can also divide the effective time segment to improve the accuracy of gesture recognition.To enhance the performance of gesture recognition system,a gesture detection scheme is proposed and the function is implemented in hardware.It adopts dual detection of moving and static targets detection and clutter map technology to provide the conditions for gesture recognition system to switch between detection mode of low power consumption and the recognition mode of high precision operation,implementing the tracking and filtering of environmental static clutter.At the same time,it solves the missed alarm caused by consecutive gestures in traditional simple moving target detection.The missed alarm rate of switching from detection mode to recognition mode in the proposed scheme is 0.The false alarm rate of switching from recognition mode to detection mode is 2%.The response time for mode switching is 733.47 ns. |
| Key words: ultrasonic radar moving target detection static target detection gesture detection clutter map |
《电讯技术》编辑部 