[论文解读] To Compress or Not To Compress: Processing vs Transmission Tradeoffs for Energy Constrained Sensor Networking

Davide Zordan, Borja Martinez+2 more|arXiv (Cornell University)|Jun 11, 2012

Underwater Vehicles and Communication SystemsEngineering参考文献 25被引用 19

一句话总结

本文研究了在能量受限的无线与水下传感器网络中，有损压缩是否能降低能耗，通过比较处理与传输能耗成本。研究发现，在无线网络中轻量级时间压缩（LTC）可节省能量，而在水下网络中基于DCT的方法更优，因传输能耗较高；文中提供了分析公式，用于根据硬件和信号参数评估能耗权衡。

ABSTRACT

In the past few years, lossy compression has been widely applied in the field of wireless sensor networks (WSN), where energy efficiency is a crucial concern due to the constrained nature of the transmission devices. Often, the common thinking among researchers and implementers is that compression is always a good choice, because the major source of energy consumption in a sensor node comes from the transmission of the data. Lossy compression is deemed a viable solution as the imperfect reconstruction of the signal is often acceptable in WSN. In this paper, we thoroughly review a number of lossy compression methods from the literature, and analyze their performance in terms of compression efficiency, computational complexity and energy consumption. We consider two different scenarios, namely, wireless and underwater communications, and show that signal compression may or may not help in the reduction of the overall energy consumption, depending on factors such as the compression algorithm, the signal statistics and the hardware characteristics, i.e., micro-controller and transmission technology. The lesson that we have learned, is that signal compression may in fact provide some energy savings. However, its usage should be carefully evaluated, as in quite a few cases processing and transmission costs are of the same order of magnitude, whereas, in some other cases, the former may even dominate the latter. In this paper, we show quantitative comparisons to assess these tradeoffs in the above mentioned scenarios. Finally, we provide formulas, obtained through numerical fittings, to gauge computational complexity, overall energy consumption and signal representation accuracy for the best performing algorithms as a function of the most relevant system parameters.

研究动机与目标

• 评估有损压缩是否能降低能量受限传感器网络的整体能耗。
• 比较不同压缩算法和网络类型（无线 vs. 水下）下处理与传输能耗成本。
• 识别在不同硬件和信号统计条件下最有效的压缩技术。
• 推导计算复杂度、能耗和重构精度的分析模型，作为关键系统参数的函数。

提出的方法

• 作者使用真实传感器数据集，分析了多种有损压缩技术，包括轻量级时间压缩（LTC）、基于DCT的低通滤波（DCT-LPF）等。
• 将能耗建模为每比特传输能耗（$E_{Tx}^\prime$）、微控制器每周期能耗（$E_0$）及算法特定计算成本（$N_c$）的函数。
• 推导出一个关键的分析不等式：$\frac{E_{Tx}^\prime}{E_0} > \frac{N_c(\eta)}{1-\eta} = \frac{\alpha\eta + \beta}{1-\eta}$，用于判断压缩是否能节省能耗。
• 通过数值拟合获得$N_c(\eta)$的参数（$\alpha$, $\beta$），从而可预测计算成本与能耗节省。
• 在两种场景下验证该框架：无线网络（CC2420无线电）与水下网络（AquaModem声学调制解调器），并使用硬件特定的能耗比。
• 该方法使系统设计者能够评估特定硬件与信号类型下，给定压缩算法是否具有能耗效率。

实验结果

研究问题

• RQ1 在何种条件下，有损压缩能降低传感器网络的总能耗？
• RQ2 在无线与水下网络中，不同压缩算法的处理与传输能耗成本如何比较？
• RQ3 在每种网络场景中，LTC与基于DCT的压缩技术中，哪种能提供更好的能耗节省？
• RQ4 计算复杂度、能耗与信号精度如何定量关联到压缩比与硬件特性等系统参数？

主要发现

• 在使用CC2420无线电的无线传感器网络中，LTC仅在压缩比$\eta \leq 0.6$时提供能耗节省，而DCT-LPF因计算成本过高，始终不具备能耗效率。
• 在传输能耗较高的水下网络中（$E_{Tx}^\prime = 10$ mW），由于$E_{Tx}^\prime / E_0$比值显著更高，LTC与DCT-LPF几乎在所有$\eta$值下均能实现能耗节省。
• 能耗节省的临界点由不等式$\frac{E_{Tx}^\prime}{E_0} > \frac{\alpha\eta + \beta}{1-\eta}$决定，其取决于硬件与算法参数。
• 对于LTC，拟合参数为$\alpha = 0.17$, $\beta = 0.08$，可预测计算成本与能耗节省。
• 对于DCT-LPF，参数为$\alpha = 0.58$, $\beta = 0.32$，表明计算成本显著更高。
• 所推导的公式使设计者能够通过比较$E_{Tx}^\prime / E_0$与算法阈值，评估新型硬件架构下压缩效率。