With the advancements in sensing-material technology, the possibilities of engineering smaller-sized sensors have increased with the ability to deploy sensors in a group and wireless capabilities to form a wireless sensor network. These sensor networks are revolutionizing many applications, including environmental and structural monitoring, military, health and fitness, agriculture, sports and entertainment, security, robotics, smart grid, and many more.
The wireless sensor network technology has been the trending research focus to make the wireless sensor nodes more efficient with improved design and architecture. Most of the individual sensor nodes are battery-driven, giving them the freedom to be deployed anywhere in the field but at the cost of a limited amount of power availability. The dominant constraint for such systems is the power consumption of the onboard sensor and electronic components, and the key to minimizing the consumption is to use an effective power-management system.
Analyzing the Sensor Nodes and Pinpointing the Power-Hungry Subsystems
Looking at the system architecture of a generic sensor node, it consists of four subsystems: a computing unit consisting of a microcontroller, a wireless communication system, a sensing system to measure the physical parameters, and a power generation and management system comprising battery, DC-to-DC converters, and an energy-harvesting system.
Among these subsystems, significant power consumption is spent during the wireless communication phase, which can be either transmitting or receiving data. Several factors affect the power consumption characteristics of the communication subsystem, including the type of modulation scheme, data-transfer rate, transmission rate of power, and the operational duty cycle. The onboard rf module’s average power consumption stands at 200 μW during sleep mode and up to 40 mW during active transmission mode.
Developments in sensor node architectures have also led to the shift from communication-centric to information-centric wireless sensor networks. The data which was earlier transported in its raw form is now processed on the sensor node itself, which is also known as edge computing and requires a processing unit with more powerful microcontrollers (MCU). Earlier, 4-bit MCUs were used mainly to acquire on-off signals like light and movement detection; now the modern 16- to 32-bit ARM-based MCUs are capable of handling even multimedia data like voice or image. These 32-bit MCUs come with a larger number of power-down modes, allowing the MCU to enter the power-saving mode for 99.9% of the time while running 0.1% of the time using the wake-up timer.
The sensing subsystem also contains various sources of power consumption, including the MEMS-based sensor itself, an analog-to-digital converter, and a signal-conditioning device. Due to a diversity of sensor types, calculating an average power consumption is difficult. Passive sensors such as touch, temperature, light sensor, etc. consume negligible power. Whereas active sensors like proximity, humidity, and flow control might sometimes consume more energy than the wireless radio during the data-acquisition period.
Power-Management Techniques in a Wireless Sensor Network
In general, a sensor node operates in at least two different modes: active mode, wherein the sensor-data acquisition and wireless communication take place, and an idle mode, when everything is inactive. The power consumed during active mode is acceptable as it is required to perform the basic function of the sensor node. On the other hand, power consumed during idle power, even if it is less, is a complete waste and the unnecessary subsystems should be completely turned off. Still, there might be some power leakage in the sensor node, which slowly drains the battery. Even with mature battery technology, the power-hungry sensors, when deployed in huge numbers (50-100), require frequent replacement of batteries. Therefore, power-management techniques are adopted in these systems, and a few of them are discussed further.
Power Gating
The wake-up strategy is also among the most energy-efficient solutions for idle mode, as the node receiver wakes up only when there’s a message for it. The key feature of such a receiver is that it consumes a fraction of the power that a regular transceiver consumes. The data rate is usually lower, the sensitivity is lower, and the latency is usually higher. Overhearing is also reduced with this technique as only the destined nodes will wake up and listen. Wake-up receivers, when active, also act as power-consumption sources and therefore should be ultra-low power to acquire the energy gain.
Dynamic Voltage Frequency Scaling
This is a method of selectively reducing both voltage and frequency on the fly so that the block just meets its deadlines for the current task. Slower blocks are run at a lower frequency and supply voltage; this technique is somewhat like power gating, but the supply isn’t cut off completely. This solution requires adequate knowledge of the application’s power requirements to compute the required reduction while the system is in operation. It also involves the control of voltage and frequency through software and hardware to maximize power savings.
Energy Harvesting from Surrounding
The sensor nodes source their power requirements from a battery that can store only a limited amount of energy, placing an upper bound on the network lifetime. Energy-harvesting devices together with rechargeable batteries comes up as an attractive alternative to battery-only systems. These systems can harvest energy from their surroundings to substantially increase the lifespan and capability of the sensor nodes. Energy is harvested using technologies like thermoelectric, solar, electrostatic, electromagnetic, and Piezoelectric. However, the nature of the surrounding is unpredictable, and one type of harvester might not guarantee the safe operation of the node. Therefore, hybrid harvesters that host two or more environmental harvesters in it are being used.
Prospects of Wireless Sensor Networks
In recent years, the use of wireless sensor networks has been witnessing rapid growth due to their portability and ease with which they can be deployed without having to worry about wired communication mediums. By integrating efficient power-management systems, the lifetime of the sensor nodes can be increased from a few days to a few weeks or maybe even months, making it suitable for monitoring large geographical areas in remote locations. Wireless sensor networks have the capacity to be installed everywhere and are one of the most pressing needs today due to their ubiquitous nature.