Keywords

1 Introduction

A wireless low-power network is composed of tiny sensor nodes powered by a battery. The key component of sensor networks [1] is processing module, sensing module, transceiver module, and power unit as shown in Fig. 1. Energy management in low power nodes is achieved through various protocols to manage power supply units and efficient utilization of available energy in a sensor node. To tackle the energy scarcity in sensor nodes balanced management between use and supply is required. The power consumption modules in a sensor node and its rate of energy usage are shown in Fig. 2. The transceiver unit is the most energy-hungry module in a sensor node. So that the software-based solutions are mainly concentrated on the communication part.

Fig. 1
figure 1

Sensor node—architecture

Full size image
Fig. 2
figure 2

Energy consumption modules in the sensor node

Full size image

The common power source for a sensor node is a battery. Consumer battery lithium-ion is suitable for common applications such as smart homes, parking lots, etc. But extreme environments, like a cold chain which is used to monitor frozen foods, pharmaceuticals, etc. demand bobbin-type batteries. Common batteries are prone to capacity losses (e.g., 30% loss after 1000 cycles). The degradation is directly proportional to environmental conditions such as temperature, humidity, etc. Another problem that reduces the life of a battery is self-discharge. The loss rate depends on the temperature and chemical reactions inside the cell. The constraints of battery supply led the researchers to propose alternate power provisioning techniques by using ambient energy [2]. These energy harvesting techniques have some limitations since there are some situations where the harvesting chance is less than the required power [3].

In the present work, the energy management solutions are divided into two namely, energy harvesting and energy conserving. The top-tier taxonomy described in this literature is shown in Fig. 3. Most power management schemes considered that data collection consumes less power than data transmission. Hence, many research works happened in the area of transmit power control and routing-based solutions [4]. The challenges in network power management are as follows.

Fig. 3
figure 3

Top tier division

Full size image
  1. (1)

    Limited power source such as battery powered

  2. (2)

    Sensor network deployment areas such as dense forest, mining area, and smart furnace.

  3. (3)

    Dynamic network topology due to adhoc nature

  4. (4)

    Node mobility in applications like cattle monitoring.

2 Energy Harvesting Techniques

One of the solutions to overcome the energy constraint issue of low power networks is energy harvesting techniques. Different sort of energy-producing techniques such as solar, wind, thermal and mechanical energy converts different sources of energy to electrical power. The general modules of a harvesting system are shown in Fig. 4. The photoelectric cells are used to convert light energy to electric energy. It will not work efficiently during cloudy and night time. The wind-based system uses turbines to produce electric power. The thermal system uses mechanical resources with an electrostatic generator. The majority of the systems use rechargeable batteries to store generated power.

Fig. 4
figure 4

Components of energy harvesting system

Full size image

The most commonly used resource is light energy. It can be artificial or natural light. This resource is a cheap, pollution less, inexhaustible and clean source of energy applicable for outdoor IoT applications. The amount of power harvesting will depend on the light intensity, atmospheric conditions and the cell area. Another parameter is the incidence of light. For full efficiency, the light source and the cell array should be perpendicular. The disadvantage of this system is (i) not suitable for indoor (ii) Depends on light and incident angle. Kansal et al. [5] conducted a study of voltage properties of different solar cell and associated storage devices in different environments. Similar works based on solar energy is shown in Table 1.

Table 1 Energy harvesting solutions: Light energy-based
Full size table

In most of the solar power-based works hybrid schemes are used, that is to manage the harvested energy efficiently with hardware or software-based modules. The energy generating module in cooperation with the energy management module is presented in [8,9,10,11]. A better choice of power management is based on prediction-based approaches. This can be achieved in two different ways namely, predicts the future energy needs of the communication nodes and predicts the future energy production from the harvesting sources. Table 2 summarises light energy-based hybrid schemes. Wind energy is another popular approach in the field of low power networks. It requires bulky hardware, which lowers the feasibility of sensor network implementations. The proposals made in [12,13,14,15,16,17] is based on wind power. Most of the proposals are based on a prediction approach since the availability of wind will depend on the weather and the previous wind history.

Table 2 Energy harvesting solutions: Light energy-based hybrid schemes
Full size table

Vibration-based works are widely used in WSNs, especially in body area networks [20] and aquatic sensor networks [21]. The power source used in these works is piezoelectric materials. Mechanical stress is converted into electric energy from motion or vibration. The paper proposes the conversion of vehicle vibration into electric energy in smart roadways [22]. Electrostatic based energy generation module uses the principle of distance change between two capacitive plates kept at a constant charge. During vibration, the distance between the plates changes and the energy is produced from the model.

The research works in [22,23,24,25,26] proposed mechanical energy based harvesting system for low power networks. Knight et al. proposed a thermal energy-based scheme, the thermostatic device is used as a power module. It is also applicable in the sensor networks deployed on water surfaces. The works based on power resources other than light is listed in Table 3.

Table 3 Energy harvesting solutions: Non-solar energy resources
Full size table

3 Energy Conserving Techniques

The innovative sensor applications in the area of technology-assisted sensor networks are found to be the most demanding innovation in the area of technology-assisted living.

The power harvesting assisted sensor nodes are always not a good solution for all the applications because of extra hardware requirements along with the actual nodes. In this section, the various approaches proposed by researchers to manage battery dependant node's efficiency for keeping the network live are discussed. Here, the paper presents a broad idea of different approaches done in network protocols. Such a scheme can be categorized into three categories based on its characteristics as given in Fig. 5.

Fig. 5
figure 5

Energy conserving approaches

Full size image

Data-driven approaches—energy management through data-driven concentrates on collected samples. This is achieved through two basic schemes (i) sensor data transmission based and (ii) sample data acquisition based. Data compression is one of the techniques that can be adopted to reduce the size of sending data.

Many compression algorithms are proposed for minimal centering power sensor nodes [27,28,29]. An alternative approach in the data reduction is, the part of sample data can be predicted so that the transmission rate towards the sink node can be lowered. Constant prediction [30], Exponential smoothing [31] and ARIMA [32] are the proposed prominent works based on this scheme.

Routing based approaches—one of the promising network-level energy management technique is energy-aware routing schemes. Variant transmitting modes and cluster management are the auspicious approaches in this area [33,34,35]. The sensor nodes are divided into different clusters with single or multiple CH to reduce the communication distance. This cluster level communication management can reduce energy depletion between sink node and data collection points. Most of the schemes try to select cluster head intelligently for the efficient management of energy [36,37,38,39]. One such approach is LEACH [40]. Many extensions to this protocol are proposed in [41,42,43]. Recently some variants of hierarchical routing such as tree-based, location-based and chain based are proposed [44, 45] for managing energy efficiency. The routing schemes such as energy-efficient routing [46], Ad-hoc on-demand distance vector routing (AODV) [47], routing protocol for low power network (RPL) [48] etc. ensures low energy consumption at communication nodes. Geographical random forwarding is a location-aware approach for selecting relay nodes [49].

Duty cycling based approaches—the network life can be increased by switching the node state between sleep and active mode. Most of the research works in this area focused on variant sleep node selection criteria. Adaptive self-configuring follows such an approach [50]. One of the efficient approaches proposed in [51, 52] is making some number of redundant mobile nodes to sleep and keeping others in wake-up mode. In research work, [53] proposes a sleep scheduling scheme-based linear distance approach. The sleep decision is taken by a node based on the probability which is proportional to the distance from the sink node. In the basic energy-conserving scheme [54] three states are defined namely active, sleep and idle for each node. Based on the routing or application layer information, nodes switch among these states. Dynamic sensor MAC [55] proposed a dynamic sleeping cycle based on power availability at nodes and network latency. The major problem with these schemes is latency unless the scheduling scheme chosen is not effective.

Mobile node-based approaches—the mobility in a low-power network is termed micro-mobility since the network contains few mobile nodes and the mobility environment is a limited area. In the present work, Greedy maximum residual energy [56], proposed a mobile sink node to collect sensor information from nodes. Another approach is made in [57], which is based on a mobile relay, and data collection is done through message ferries. These nodes move around the fields, collect data and send it to the destination node. In the literature [58] and [59] has done an experiment based study to show the effect of transmit power and energy consumption in adhoc networks. The results are shown in the Figs. 6 and 7. Various power conservation-based schemes discussed are summarized in Table 4 (Fig. 6).

Fig. 6
figure 6

Remaining energy for different transmit power in mobile network

Full size image
Fig. 7
figure 7

Remaining energy for different transmit power in static network

Full size image
Table 4 Summary of energy-conserving solutions
Full size table

Low power network protocols—as the popularity and applications of low power networks are increasing day by day, different standard protocols are exclusively defined for low power networks. The characteristics of the network protocols are given in Table 5. Conventional communication protocols are defined for sending a large amount of data (Fig. 7).

Table 5 Power consumption and properties of low power network protocols
Full size table

Traditionally the sensor nodes are dealing with small scalar values like pressure, temperature, humidity, etc. It becomes wastage of energy and bandwidth if traditional protocols are used to communicate with these small-sized data. So, the protocols like low power Wi-Fi, Bluetooth Low Energy, Zigbee, Z-Wave, and LoRaWAN [65,66,67,68,69] are introduced for low power communication networks.

The inferences made from the investigations are as follows. (1) The power management schemes in low power networks can be categorized into two: energy harvesting and energy conserving. (2) The energy harvesting schemes are efficient in terms of long-life networks. (3) The efficiency of energy harvesting schemes depends on (i) deployment environments and (ii) the proper management and storage of harvested energy, since the availability of common resources like light, wind, etc. depends on environmental factors. (4) The network-based energy conserving protocol demands prior knowledge of network power distribution and its consumption. (5) Duty cycling-based schemes are prone to delay because the sleep node selection and synchronization is a hurdle. (5) The power level management needs proper power level selection for efficient communication.

4 Conclusions

This investigation work presented different energy-aware schemes in IoT network communication. The review is done in two directions: energy harvesting approaches and energy-conserving techniques. A tabular-based summary with performance metric comparison is presented for all energy management kinds of literature discussed in this work. It is recommended that a hybrid scheme that combines energy harvesting and energy-conserving techniques, which will be the best choice for increasing network lifetime. However, energy management is still an open challenge to researchers, and lots of studies are required towards the efficient functioning of low power networks.