Garudadri Harinath, Baheti Pawan K
Qualcomm Incorporated, San Diego, CA 92121, USA.
Annu Int Conf IEEE Eng Med Biol Soc. 2009;2009:2450-3. doi: 10.1109/IEMBS.2009.5333969.
In this work, we propose an effective application layer solution for packet loss mitigation in the context of Body Sensor Networks (BSN) and healthcare telemetry. Packet losses occur due to many reasons including excessive path loss, interference from other wireless systems, handoffs, congestion, system loading, etc. A call for action is in order, as packet losses can have extremely adverse impact on many healthcare applications relying on BAN and WAN technologies. Our approach for packet loss mitigation is based on Compressed Sensing (CS), an emerging signal processing concept, wherein significantly fewer sensor measurements than that suggested by Shannon/Nyquist sampling theorem can be used to recover signals with arbitrarily fine resolution. We present simulation results demonstrating graceful degradation of performance with increasing packet loss rate. We also compare the proposed approach with retransmissions. The CS based packet loss mitigation approach was found to maintain up to 99% beat-detection accuracy at packet loss rates of 20%, with a constant latency of less than 2.5 seconds.
在这项工作中，我们针对体域网（BSN）和医疗遥测环境下的丢包缓解问题提出了一种有效的应用层解决方案。丢包的发生有多种原因，包括过度的路径损耗、来自其他无线系统的干扰、切换、拥塞、系统负载等。由于丢包可能会对许多依赖人体区域网络（BAN）和广域网（WAN）技术的医疗应用产生极其不利的影响，因此需要采取行动。我们缓解丢包的方法基于压缩感知（CS），这是一种新兴的信号处理概念，其中可以使用比香农/奈奎斯特采样定理建议的数量少得多的传感器测量值来以任意精细的分辨率恢复信号。我们给出的仿真结果表明，随着丢包率的增加，性能会逐渐下降。我们还将所提出的方法与重传方法进行了比较。结果发现，基于CS的丢包缓解方法在丢包率为20%时能保持高达99%的心跳检测准确率，且延迟恒定小于2.5秒。
