Sage Journals: Discover world-class research

Abstract

In wireless sensor networks, a node that increases its sensing rate to gather more data consumes more energy. Consequently, it will be exhausted earlier than other nodes. In this article, we propose a sensing rate control scheme to increase data resolution without increasing the number of blackout nodes in energy-harvesting wireless sensor networks. In the proposed scheme, each node selects a compression algorithm and adjusts the sensing rate to gather more data while the node operates properly. When a node is estimated to have extra energy, it gathers additional data. In contrast, when it is estimated to be exhausted, it gathers less data to save energy. It also selects an appropriate compression algorithm according to the compression ratio of the selected algorithm to reduce the workload of intermediate nodes. An experimental result verifies that the proposed scheme gathers more data with a lower number of blackout nodes than other schemes.

Keywords

Wireless sensor network energy-harvesting compression sensing rate compression rate

Introduction

Wireless sensor networks (WSNs) have been used to gather environmental information such as temperature, humidity, or pressure from inaccessible areas such as virgin forest, volcanoes, or deep water. A WSN consists of many wireless sensor nodes, which have a limited lifetime because they run on batteries. Therefore, it is important to prolong the lifetime of the nodes in WSNs, and many studies have been researched.^1,2

One of the methods of improving the lifetime of a WSN is to use energy-harvesting nodes. Such a node can potentially operate permanently by recharging its battery with environmental energy. Unlike typical WSNs, therefore, the primary research goal of energy-harvesting WSNs is the efficient use of the harvested energy during permanent operation of the nodes. In particular, if a node harvests extra energy, it can gather more data to improve the resolution of data using that energy.

A wireless sensor node in a WSN usually involves multi-hop transmission, passing data from one node to the next, because the radio of a node cannot reach a sink node directly. As a result, the nodes closer to the sink node consume more energy because they have to transmit more data than the outer nodes.^3,4 One of the methods of solving this problem is data compression, which can reduce the burden of intermediate nodes with the compressed data. However, this type of node should be applied carefully because the energy consumed during the performance of the data compression algorithm is not negligible.

In this article, we propose a selective data compression scheme for energy-harvesting WSNs, in which a node adjusts its sensing period and chooses a compression algorithm depending on its energy level to improve the resolution of gathered data. In order to achieve the goal, when a node has excessive energy, it gathers data more frequently and chooses a compression algorithm that has a high compression ratio but consumes more energy. In contrast, when a node is expected to be defunct, it gathers data less frequently and uses a compression algorithm that consumes less energy.

The remainder of this article is organized as follows. In section “Related work,” we review related work. Section “Adaptive sensing and compression rate selection scheme” introduces a sensing rate and data compression selection scheme and describes how nodes choose their sensing rate and compression algorithm. In section “Performance evaluation,” we present experimental results and assess the performance of our scheme. Section “Conclusion” concludes the article.

Related work

Data compression schemes for WSNs

A wireless sensor node is a tiny embedded system with many constraints incurred by the limited hardware capacity, in particular processing speed and memory size. For that reason, a lightweight compression scheme has been needed for WSNs.^5–7

Sadler and Martonosi⁸ proposed two lossless compression schemes. The first one was the Sensor Lempel–Ziv–Welch (S-LZW) that was designed for WSNs based on the Lempel–Ziv–Welch (LZW) algorithm.⁹ The other was a Sensor Lempel–Ziv–Welch with Burrows–Wheeler Transform (S-LZW-BWT), which rearranges a character string into runs of similar characters using the Burrows–Wheeler Transform (BWT)¹⁰ before compression. Marcelloni and Vecchio¹¹ proposed an entropy-based lossless compression scheme that used the difference of gathered data. It is similar to the JPEG (Joint Photographic Experts Group) format commonly used for image compression. In each compression scheme described above, the nodes compress only their own data. Meanwhile, some researchers proposed in-network compression schemes, in which relay nodes compressed data received from other nodes.^12,13 Compressive sensing is the representative scheme of in-network compression. Compressive sensing can significantly reduce the number of sampling points that corresponds to the amount of obtained data.^14–16 Zheng et al.¹⁷ introduced a compressive sensing scheme, which used a random walk algorithm for data gathering in WSNs unlike the conventional approaches that required uniform sampling. Liu et al.¹⁸ proposed a compressive sensing scheme for WSNs, which adopted a power-law decaying data model verified by real data sets. They also proposed a random projection-based estimation algorithm for the data model. Li et al.¹⁴ described how compressive sensing could provide new insights into data sampling and acquisition in WSNs, and an IoT (Internet of Things) framework using a compressive sensing algorithm. Nguyen and Teague¹⁹ proposed a compressive sensing algorithm for clustering schemes in WSNs.

Energy-harvesting WSNs

Traditional WSNs usually have a limited lifetime because they consist of wireless nodes that run on batteries. Recently, in order to overcome this limitation, energy-harvesting nodes have been used²⁰ that harvest energy from the sun,^21–26 wind,^26,27 heat,²⁸ and vibration.²⁹ In particular, solar-powered wireless sensor nodes are usually preferred^30–32 because of the high areal density of solar power (about $15 mW / c m^{2}$ ). Unlike traditional WSNs, many studies^32–35, have been proposed for energy-harvesting WSNs to maximize energy utilization while the nodes survive indefinitely.

Adaptive sensing and compression rate selection scheme

In this article, we propose an adaptive sensing and compression rate selection scheme for energy-harvesting WSNs to improve the resolution of gathered data. The scheme is designed for delay-tolerant WSNs³⁶ where each node stores its own sensed data for a while and occasionally transmits it to the sink node. In this scheme, a node periodically selects a data compression algorithm and adjusts its sensing rate according to its estimated residual energy to improve its longevity. When it is estimated that a node has harvested excess energy, it gathers additional data and compresses it with a high-compression algorithm using the extra energy. In contrast, when the node is estimated to be exhausted, it gathers less data and similarly selects an appropriate compression algorithm. In order to achieve this, a node selects one of the following three modes according to its estimated residual energy.

Low-compression mode

A node in a normal state transmits data compressed by an algorithm that has a low compression ratio and consumes relatively less energy in order to reduce energy consumption. In this article, the S-LZW algorithm⁸ is used for this mode. S-LZW is designed to reduce the computation overhead so as to use it for WSNs based on the LZW algorithm.⁹Figure 1(b) shows a simple operation in this mode. As shown in Figure 1(b), this mode is selected by a node, which has no enough energy to select high-compression mode and is estimated not to be exhausted. The node in this mode gathers the larger amount of data than transmission size, and it transmits compressed data packets whose size is reduced to the same size as the transmission size.

Figure 1.

Overview of the sensing and compression rate selection scheme: (a) high-compression mode, (b) low-compression mode, and (c) energy-saving mode.

High-compression mode

If the harvested energy of a node is about to run over the capacity of its energy storage, the node collects sensing data more frequently in order to increase the resolution of data. In general WSNs, however, nodes that are closer to the sink node spend more energy than the outside nodes due to the relay data. Outer nodes may have more energy, and they will be able to sense more frequently. As a result, the additional sensed data impose a heavy burden on the intermediate nodes. Therefore, in a case where a node gathers additional data, it should compress the data using a high-compression algorithm to reduce the overhead of the intermediate nodes.

We use the S-LZW-BWT algorithm for this mode. The S-LZW-BWT algorithm rearranges a character string into runs of similar characters using the BWT¹⁰ before compression. BWT does not shrink the size of the data by itself, but it makes the data easier to compress. It helps to improve the compression ratio, but it consumes much more energy than S-LZW because of the additional operations. Table 1 describes the compression ratio and energy consumption of the compression algorithms used in each mode. The compression ratio in Table 1 is defined as follows

$c o m p_r a t i o = 100 \cdot (1 - \frac{c o m p_s i z e}{o r i g_s i z e})$ (1)

Table 1.

Comparison between two compression algorithms.

Algorithm	S-LZW	S-LZW-BWT
Compression ratio	27.25%	32.975%
Energy consumption	0.045 mJ/B	0.116 mJ/B

S-LZW: Sensor Lempel–Ziv–Welch; S-LZW-BWT: Sensor Lempel–Ziv–Welch with Burrows–Wheeler Transform.

Figure 1(a) represents the operation of a node in this mode. As shown in Figure 1(a), this mode is selected by a node whose energy is expected to overflow in low-compression mode. The node in this mode gathers the larger amount of data than low-compression mode, and it transmits compressed data packets whose size is reduced to the same size as the transmission size similar to low-compression mode.

Energy-saving mode

If a node is expected to be exhausted, it selects an energy-saving mode to reduce energy consumption. The node in this mode adjusts its sensing period so as not to exhaust its energy, and it transmits uncompressed data. Figure 1(c) shows the operation in energy-saving mode. As shown in Figure 1(c), this mode is selected by a node whose energy is expected to be exhausted in other modes. The node in this mode gathers less data than other modes and transmits uncompressed data whose size is the same as the transmission size.

Operation of a node

A sensor node gathers sensing data every $p_{sense}$ , and it transmits the compressed data every transmission period $p_{tx}$ . At the beginning of its $p_{tx}$ , it determines its sensing period $p_{sense}$ and the compression algorithm, both of those are used during $p_{tx}$ . In this article, the transmission period $p_{tx}$ is described as a round. Figure 2 represents the operation of a node over time during two rounds.

Figure 2.

Operation of a node over time during two rounds.

Operation mode selection

In the proposed scheme, a node determines its sensing period and compression algorithm dynamically in order to improve data resolution while it continues to survive. Therefore, the node determines its operation mode that has a different sensing period and compression ratio according to its estimated residual energy, calculated based on the amount of harvested and consumed energy. A node is initialized as low-compression mode and then selects its mode for the next round at the beginning of each transmission period $p_{tx}$ .

High-compression mode

If harvested energy exceeds the capacity of the energy storage, it gets lost. In order to prevent this waste, if the estimated residual energy is likely to exceed the limit, the node selects a high-compression mode, in which the node gathers additional data to improve data resolution.

In order to examine whether the residual energy exceeds the capacity of an energy storage, a node should estimate its residual energy during the next round. Assuming the current time is t, after the transmission period $p_{tx}$ , the residual energy of a node $E'_{remain} (t, p_{tx})$ is as follows

$E_{remain}^{'} (t, p_{tx}) = E_{remain} (t) + E_{h}^{'} (t, p_{tx}) - E_{c}^{'} (t, p_{tx})$ (2)

where $E'_{h} (t, p_{tx})$ and $E'_{c} (t, p_{tx})$ are the estimated harvested and consumed energy from t during $p_{tx}$ , respectively. By substituting the harvesting power $P_{h} (t)$ and the consuming power $P_{sys} (t)$ into equation (2), $E'_{remain} (t, p_{tx})$ can be expressed as follows

$E'_{remain} (t, p_{tx}) = E_{remain} (t) + \int_{t}^{t + p_{tx}} (P_{h} (τ) - P_{sys} (τ)) d τ$ (3)

$P_{h} (t)$ can be estimated using several energy-harvesting estimation schemes.^21,37–39 $P_{sys} (t)$ is estimated by the moving average of historical data.

If $E'_{remain} (t, p_{tx})$ in equation (3) is larger than the capacity of the energy storage, the harvested energy may be wasted during the next $p_{tx}$ . Therefore, if $E'_{remain} (t, p_{tx})$ satisfies the following condition, then that node selects high-compression mode

$E'_{remain} (t, p_{tx}) > C$ (4)

where C is the capacity of the energy storage.

Low-compression mode

If a node in high-compression mode is expected to be exhausted due to a large consumption of energy, it chooses low-compression mode to save its energy. In order to know whether the energy may be exhausted, we should confirm whether its estimated residual energy is less than the minimum energy required to operate properly during $p_{tx}$ .

By substituting consuming power in high-compression mode $P_{sys}^{h} (t)$ into equation (2), the estimated residual energy $E'_{remain} (t, p_{tx})$ ) becomes

$E'_{remain}^{h} (t, p_{tx}) = E_{remain} (t) + \int_{t}^{t + p_{tx}} (P_{h} (τ) - P_{sys}^{h} (τ)) d τ .$ (5)

$E'_{remain}^{h} (t, p_{tx})$ can be calculated by estimating $P_{h} (t)$ and $P_{sys}^{h} (t)$ in the same way in equation (3). If $E'_{remain}^{h} (t, p_{tx})$ in equation (5) becomes smaller than $E_{\min}$ , which is the minimum energy to survive and determined by the specification of the node, that node may be exhausted. Therefore, if a node satisfies the following condition during $p_{tx}$ , it chooses low-compression mode in order to survive

$E'_{remain}^{h} (t, p_{tx}) < E_{\min}$ (6)

Energy-saving mode

If a node may be exhausted in both high- and low-compression modes, that node should choose energy-saving mode to save its energy. Therefore, a node operates in energy-saving mode if it satisfies the following condition

$E'_{remain} (t, p_{tx}) < E_{\min}$ (7)

If it is estimated that a node in energy-saving mode would not be exhausted due to the harvesting energy, the node switches to low-compression mode. Thus, a node should switch to low-compression mode if the node satisfies the following conditions

$E'_{remain} (t, p_{tx}) > E_{\min}$ (8)

$P_{h} (τ) - P_{sys} (τ) > 0$ (9)

Determination of sensing period

As mentioned in subsection “Operation mode selection,” in the proposed scheme, a node varies its sensing period according to its operation mode. However, dynamically varying the amount of transmitting data makes it difficult to estimate the consumption of energy because the load of intermediate nodes cannot be estimated correctly. To avoid that, the amount of transmitting data has to be stable. For that reason, in order that the amount of transmission data remains stable, the sensing period of a node should be adjusted according to the compression ratio of algorithms used in each mode.

Assuming that the amount of data gathered by a node at each sensing period $p_{sense}$ is $S_{sense}$ , the amount of compressed data that gathered during the transmission period $p_{tx}$ in order to transmit is expressed as follows

$S_{tx} = S_{sense} \frac{p_{tx}}{p_{sense}} (1 - R_{compress})$ (10)

where $R_{compress}$ is the compression ratio. This ratio is determined depending on the application type and compression algorithm used in each mode; $S_{sense}$ , $p_{tx}$ , and $p_{sense}$ are predetermined according to the applications. However, note that $S_{tx}$ should be the same in all modes. Thus, the sensing period in high-compression mode $p_{sense}^{h}$ can be calculated as follows

$p_{sense}^{h} = \frac{S_{sense}}{S_{tx}} p_{tx} (1 - R_{compress}^{h})$ (11)

Entire process of the proposed scheme

The proposed scheme can be used for an energy-harvesting WSN, which consists of many tiny sensor nodes that monitor their environments. Every sensor node initializes its operating mode as low-compression mode and reselects its mode every $p_{tx}$ . (1) If a node selects low-compression mode, it gathers sensing data every $p_{sense}$ , aggregates them for a transmission period $p_{tx}$ , and transmits compressed data with a low-energy compression algorithm. After $p_{tx}$ , it selects its mode again by calculating equations (4) and (7). (2) If a node selects high-compression mode, it gathers sensing data every $p_{sense}^{h}$ , aggregates them, and transmits compressed data with a high-energy compression algorithm. After $p_{tx}$ , it selects its mode again by calculating equation (6). (3) If a node selects energy-saving mode, it gathers sensing data every $p_{sense}^{\max}$ , aggregates them, and transmits uncompressed data. After $p_{tx}$ , it selects its mode again by calculating equations (8) and (9).

By repeating the above process, the proposed scheme can reduce the number of blackout nodes and increase the amount of obtained data at the sink node effectively. Algorithm 1 shows the mode selection algorithm of the proposed scheme as described above.

Algorithm 1 Mode selection algorithm.
Require: $p_{tx}$ of a node expires
Ensure: Determine the mode of the node
1: Calculate Equation (3) and (5)
2: if $m = LOW_COMPRESSION_MODE$ then
3: if $E'_{remain} (t, p_{tx}) > C$ then
4: $m \leftarrow$ SelectHighCompressionMode( )
5: else if $E'_{remain} (t, p_{tx}) < E_{\min}$ then
6: $m \leftarrow$ SelectEnergySavingMode( )
7: end if
8: else if $m = HIGH_COMPRESSION_MODE$ then
9: if $E_{remain}^{' h} (t, p_{tx}) < E_{\min}$ then
10: if $E'_{remain} (t, p_{tx}) < E_{\min}$ then
11: $m \leftarrow$ SelectEnergySavingMode( )
12: else
13: $m \leftarrow$ SelectLowCompressionMode( )
14: end if
15: end if
16: else if $m = ENERGY_SAVING_MODE$ then
17: if $E'_{remain} (t, p_{tx}) > E_{\min} \land P_{h} (τ) - P_{sys} (τ) > 0$ then
18: $m \leftarrow$ SelectLowCompressionMode( )
19: end if
20: end if
21: functionSelectHighCompressionMode
22: $p \leftarrow p_{sense}^{h}$
23: $a \leftarrow$ S-LZW-BWT
24: return $HIGH_COMPRESSION_MODE$
25: end function
26: functionSelectLowCompressionMode
27: $p \leftarrow p_{sense}$
28: $a \leftarrow$ S-LZW
29: return $LOW_COMPRESSION_MODE$
30: end function
31: functionSelectEnergySavingMode
32: $p \leftarrow p_{sense}^{\max}$
33: $a \leftarrow$ NONE
34: return $ENERGY_SAVING_MODE$
35: end function

Performance evaluation

In order to evaluate our scheme, we performed a simulation. In this section, we explain the simulation and its results.

Simulation environment

We wrote a simulation in C++ to compare our scheme with other schemes such as noncompression, S-LZW (low compression), and S-LZW-BWT (high compression). In this simulation, we used TelosB mote model^40,41 with energy-harvesting module, which has been widely used in WSNs. We measured the average residual energy, the number of blackout nodes, the total amount of sensed data, and the total amount of data arriving at the sink node. We deployed 100 energy-harvesting nodes and one sink node at random positions, and each test set ran 10 times for 1000 rounds to obtain the average values. At the beginning of each round, all nodes found their routes to the sink node using the minimum depth tree (MDT) algorithm, and they transmitted data to the sink node. After that, the nodes determined their next mode and sensing period at the end of the round and started the next round. Table 2 summarizes the important parameters used in this simulation. The hardware specification in the table is derived from TelosB mote specification⁴¹ and the power consumption model of TelosB.⁴⁰

Table 2.

Important parameters used in the simulation.

Parameter	Value
Radio chipset	CC2420
Radio range	20–30 m
TX power	–5 dBm at 3 V
RX power	0 dBm at 3 V
Baud rate	250 kbps
Super capacitor	10,400 mAh
$S_{sense}$	104 byte
$p_{tx}$	20 min
$P_{h}$	1–3 mW
$ϵ$	15%
Rounds	1000 rounds

Simulation results

Figure 3 shows the average residual energy with the passage of time. In the noncompression scheme, nodes were always in overcharge condition because they did not consume energy to compress data. Conversely, the average residual energy of nodes in other schemes was lower than that in the noncompression because they consume more energy to collect extra data and compress it. In particular, because S-LZW-BWT consumed more energy to compress and collect additional data, its average residual energy was the smallest among the schemes. However, the proposed scheme exploited the S-LZW and S-LZW-BWT algorithms selectively. Consequently, its average residual energy was intermediate between the S-LZW and S-LZW-BWT algorithms. In Figure 4, however, the proportions of blackout nodes of S-LZW, S-LZW-BWT, and the proposed scheme were 15.9%, 23.1%, and 10.6%, respectively. The results, apart from the proposed scheme, were similar to the average residual energy in Figure 3, but the proportion of blackout nodes of the proposed scheme was as small as S-LZW. This was because the adaptive mode selection of the proposed scheme effectively suppressed the occurrence of blackout nodes.

Figure 3.

Average residual energy over time.

Figure 4.

Number of blackout nodes over time.

Figure 5 shows the total amount of data sensed by each node with the passage of time. The scheme using the higher compression ratio algorithm collects the more data because all schemes transmit the same amount of data. In Figure 5, therefore, S-LZW and S-LZW-BWT schemes, respectively, gathered 37.5% and 49.2% more data than noncompression scheme. However, the result of the proposed scheme was similar to S-LZW-BWT in spite of higher energy level of the proposed scheme than S-LZW-BWT as shown in Figure 3.

Figure 5.

Change in the amount of sensed data.

Figure 6 represents the total amount of data obtained at the sink over time. Comparing it with the result in Figure 5, the proposed scheme only lost 16.7% sensed data during transmission, while S-LZW and S-LZW-BWT lost 17.2% and 25.6%, respectively. Note that the proportion of lost data in S-LZW-BWT was large. This is because many intermediate nodes using S-LZW-BWT would be exhausted by the significant amount of energy needed to compress the data, as shown in Figure 4, and the data which they should have relayed could not be transmitted. However, the proposed scheme adaptively selects sensing periods and compression algorithms to gather more data within range to operate properly. As a result, the proposed scheme could effectively reduce the number of blackout nodes and then collect more data than other schemes.

Figure 6.

Change in the amount of data delivered to the sink.

Figures 7 and 8, respectively, depict the amount of data sensed by all nodes and data arriving at the sink node with the amount of harvested energy. The amount of sensed data in Figure 7 was similar to the result in Figure 5, but the amount of received data in Figure 8 was unlike the result in Figure 7 according to the energy-harvesting ratio. In the case of the higher harvesting ratio, since there were few blackout nodes, almost all sensed data arrived at the sink node, and the result was similar to that in Figure 7. However, when the harvesting ratio decreased, S-LZW-BWT, which consumes more energy, incurred more blackout nodes. As a result, the amount of gathered data of S-LZW-BWT was lower than that of S-LZW or noncompression scheme. In contrast, the proposed scheme showed a superior performance over the other schemes in all cases because it could suppress the occurrence of blackout nodes effectively according to the circumstances.

Figure 7.

Amount of sensed data with the charging rate.

Figure 8.

Amount of gathered data with the charging rate.

Table 3 shows the comparison of the simulation results of all schemes at the 500th round as described above. The results show that the proposed scheme has better performance than other schemes though some nodes have been blackout.

Table 3.

Comparison of the simulation results at the 500th round.

	No compression	S-LZW	S-LZW-BWT	Proposed scheme
Blackout (#)	0	15.9	23.1	10.6
Sensed data (B)	11,940	15,826	16,999	17,070
Received data (B)	11,828	12,464	11,392	16,625

S-LZW: Sensor Lempel–Ziv–Welch; S-LZW-BWT: Sensor Lempel–Ziv–Welch with Burrows–Wheeler Transform.

Conclusion

In this article, we have proposed a sensing and compression rate selection scheme for energy-harvesting WSNs, in which nodes adaptively select a compression algorithm and adjust the sensing rate in order to suppress the occurrence of blackout nodes and increase the resolution of sensed data at the same time.

In cases of expected overcharge, these nodes use the extra harvested energy to collect more data and compress it to reduce transmission energy. As a result, they can collect more data while they consume more energy within range to operate properly. Conversely, in cases of expected lack of energy, they expand the sensing rate and compress the data using a low-compression algorithm that consumes less energy in order to reduce the amount of transmitted data. Simulation results verified that the proposed scheme could reduce the occurrence of blackout nodes compared to a high-compression scheme and could increase the amount of data that arrives at the sink node compared to a low-compression scheme. Consequently, the proposed scheme could lead to better performance using the advantages of other schemes adaptively.

Footnotes

Academic Editor: Chunming Qiao

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This research was supported by the Next-Generation Information Computing Development Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education,Science and Technology (No. 2012M3C4A7032182).

References

Yick

Mukherjee

Ghosal

Wireless sensor network survey. Comput Netw 2008; 52(12): 2292–2330.

Akyildiz

Sankarasubramaniam

. Wireless sensor networks: a survey. Comput Netw 2002; 38(4): 393–422.

Rooshenas

Rabiee

Movaghar

. Reducing the data transmission in wireless sensor networks using the principal component analysis. In: Proceedings of the sixth international conference on intelligent sensors, sensor networks and information processing (ISSNIP), Brisbane, QLD, Australia, 7–10 December 2010, pp.133–138. New York: IEEE.

Olivares

Tirado

Orozco-Barbosa

. Simulation of power-aware wireless sensor network architectures. In: Proceedings of the ACM international workshop on performance monitoring, measurement, and evaluation of heterogeneous wireless and wired networks, Torremolinos, 2 October 2006, pp.32–39. New York: ACM.

Kimura

Latifi

. A survey on data compression in wireless sensor networks. In: Proceedings of the international conference on information technology: coding and computing (ITCC2005), Las Vegas, NV, 4–6 April 2005, vol. 2, pp.8–13. New York: IEEE.

Srisooksai

Keamarungsi

Lamsrichan

. Practical data compression in wireless sensor networks: a survey. J Netw Comput Appl 2012; 35(1): 37–59.

Razzaque

Bleakley

Dobson

Compression in wireless sensor networks: a survey and comparative evaluation. ACM T Sensor Network 2013; 10(1): 5.

Sadler

Martonosi

. Data compression algorithms for energy-constrained devices in delay tolerant networks. In: Proceedings of the 4th international conference on embedded networked sensor systems, Boulder, CO, 1–3 November 2006, pp.265–278. New York: ACM.

Welch

TA.

A technique for high-performance data compression. Computer 1984; 6(17): 8–19.

10.

Burrows

Wheeler

DJ.

A block-sorting lossless data compression algorithm. Technical report 124, 1994, http://www.hpl.hp.com/techreports/Compaq-DEC/SRC-RR-124.pdf

11.

Marcelloni

Vecchio

A simple algorithm for data compression in wireless sensor networks. IEEE Commun Lett 2008; 12(6): 411–413.

12.

Petrovic

Shah

Ramchandran

. Data funneling: routing with aggregation and compression for wireless sensor networks. In: Proceedings of the first IEEE international workshop on sensor network protocols and applications, Anchorage, AK, 11 May 2003, pp.156–162. New York: IEEE.

13.

Arici

Gedik

Altunbasak

. PINCO: a pipelined in-network compression scheme for data collection in wireless sensor networks. In: Proceedings of the 12th international conference on computer communications and networks (ICCCN 2003), Dallas, TX, 20–22 October 2003, pp.539–544. New York: IEEE.

14.

Da Xu

Wang

Compressed sensing signal and data acquisition in wireless sensor networks and internet of things. IEEE T Ind Inform 2013; 9(4): 2177–2186.

15.

Luo

Compressive data gathering for large-scale wireless sensor networks. US8280671 B2 Patent, 2012.

16.

Haupt

Bajwa

Rabbat

. Compressed sensing for networked data. IEEE Signal Proc Mag 2008; 25(2): 92–101.

17.

Zheng

Yang

Tian

. Data gathering with compressive sensing in wireless sensor networks: a random walk based approach. IEEE T Parall Distr 2015; 26(1): 35–44.

18.

Liu

Zhu

Kong

. CDC: Compressive Data Collection for wireless sensor networks. IEEE T Parall Distr 2015; 26(8): 2188–2197.

19.

Nguyen

Teague

KA.

Compressive sensing based data gathering in clustered wireless sensor networks. In: Proceedings of the IEEE international conference on distributed computing in sensor systems (DCOSS), Marina Del Rey, CA, 26–28 May 2014, pp.187–192. New York: IEEE.

20.

Sudevalayam

Kulkarni

Energy harvesting sensor nodes: survey and implications. IEEE Commun Surv Tut 2011; 13(3): 443–461.

21.

Kansal

Hsu

Zahedi

. Power management in energy harvesting sensor networks. ACM T Embed Comput S 2007; 6(4): 32.

22.

Jiang

Polastre

Culler

Perpetual environmentally powered sensor networks. In: Proceedings of the 4th international symposium on information processing in sensor networks (IPSN 2005), Los Angeles, CA, 24–27 April 2005, pp.463–468. New York: IEEE.

23.

Raghunathan

Kansal

Hsu

. Design considerations for solar energy harvesting wireless embedded systems. In: Proceedings of the 4th international symposium on information processing in sensor networks, Los Angeles, CA, 24–27 April 2005, p.64. New York: IEEE Press.

24.

Hsu

Zahedi

Kansal

. Adaptive duty cycling for energy harvesting systems. In: Proceedings of the 2006 international symposium on low power electronics and design, Tegernsee, 4–6 October 2006, pp.180–185. New York: ACM.

25.

Simjee

Chou

PH.

Everlast: long-life, supercapacitor-operated wireless sensor node. In: Proceedings of the 2006 international symposium on low power electronics and design (ISLPED ’06), Tegernsee, 4–6 October 2006, pp.197–202. New York: IEEE.

26.

Park

Chou

. AmbiMax: Autonomous energy harvesting platform for multi-supply wireless sensor nodes. In: Proceedings of the 3rd annual IEEE communications society on sensor and ad hoc communications and networks (SECON ’06), Reston, VA, 25–29 September 2006, vol. 1, pp.168–177. New York: IEEE.

27.

Weimer

Zane

. Remote area wind energy harvesting for low-power autonomous sensors. In: Proceedings of the 37th IEEE power electronics specialists conference (PESC ’06), Jeju, South Korea, 18–22 June 2006, pp.1–5. New York: IEEE.

28.

Stordeur

Stark

. Low power thermoelectric generator-self-sufficient energy supply for micro systems. In: Proceedings of the XVI international conference on thermoelectrics (ICT ’97), Dresden, 26–29 August 1997, pp.575–577. New York: IEEE.

29.

Meninger

Mur-Miranda

Amirtharajah

. Vibration-to-electric energy conversion. IEEE T VLSI Syst 2001; 9(1): 64–76.

30.

Roundy

SJ.

Energy scavenging for wireless sensor nodes with a focus on vibration to electricity conversion. PhD Thesis, University of California, Berkeley, CA, 2003.

31.

Roundy

Steingart

Frechette

. Power sources for wireless sensor networks. In: Karl

Wolisz

Willig

(eds) Wireless sensor networks. Berlin: Springer, 2004, pp.1–17.

32.

Stojmenovic

Handbook of sensor networks: algorithms and architectures (vol. 49). Hoboken, NJ: John Wiley & Sons, 2005.

33.

Kansal

Potter

Srivastava

MB.

Performance aware tasking for environmentally powered sensor networks. Perf E R Si 2004; 32(1): 223–234.

34.

Seah

Tan

HP.

Wireless sensor networks powered by ambient energy harvesting (wsn-heap)-survey and challenges. In: Proceedings of the 1st international conference on wireless communication, vehicular technology, information theory and aerospace & electronic systems technology (Wireless VITAE), Aalborg, 17–20 May 2009, pp.1–5. New York: IEEE.

35.

Yang

Wang

Noh

. SolarStore: enhancing data reliability in solar-powered storage-centric sensor networks. In: Proceedings of the 7th international conference on mobile systems, applications, and services, Wroclaw, 22–25 June 2009, pp.333–346. New York: ACM.

36.

Fall

. A delay-tolerant network architecture for challenged internets. In: Proceedings of the 2003 conference on applications, technologies, architectures, and protocols for computer communications, Karlsruhe, 25–29 August 2003, pp.27–34. New York: ACM.

37.

Piorno

Bergonzini

Atienza

. Prediction and management in energy harvested wireless sensor nodes. In: Proceedings of the 1st international conference on wireless communication, vehicular technology, information theory and aerospace & electronic systems technology (Wireless VITAE), Aalborg, 17–20 May 2009, pp.6–10. New York: IEEE.

38.

Moser

Thiele

Brunelli

. Adaptive power management in energy harvesting systems. In: Proceedings of the conference on design, automation and test in Europe, Nice, 16–20 April 2007, pp.773–778. San Jose, CA: EDA Consortium.

39.

Cammarano

Petrioli

Spenza

. Pro-energy: a novel energy prediction model for solar and wind energy-harvesting wireless sensor networks. In: Proceedings of the IEEE 9th international conference on mobile ad-hoc and sensor systems (MASS), Las Vegas, NV, 8–11 October 2012, pp.75–83. New York: IEEE.

40.

Cody-Kenny

Guerin

Ennis

. Performance evaluation of the 6LoWPAN protocol on MICAz and TelosB motes. In: Proceedings of the 4th ACM workshop on performance monitoring and measurement of heterogeneous wireless and wired networks, Tenerife, 26–30 October 2009, pp.25–30. New York: ACM.

41.

MEMSIC. MPR2400CB specification, http://www.memsic.com/userfiles/files/datasheets/wsn/6020-0060-04-b_micaz.pdf

Adaptive sensing and compression rate selection scheme for energy-harvesting wireless sensor networks

Abstract

Keywords

Introduction

Related work

Data compression schemes for WSNs

Energy-harvesting WSNs

Adaptive sensing and compression rate selection scheme

Low-compression mode

High-compression mode

Energy-saving mode

Operation of a node

Operation mode selection

High-compression mode

Low-compression mode

Energy-saving mode

Determination of sensing period

Entire process of the proposed scheme

Performance evaluation

Simulation environment

Simulation results

Conclusion

Footnotes

Declaration of conflicting interests

Funding

References