The importance of low power sensing for the Internet of Things - Embedded.com

The importance of low power sensing for the Internet of Things

Low power wireless technology is enabling substantial cost reductions for traditional wired sensing systems, as well as opening up new possibilities for sensor networking that simply could not be done with wires.

Low power wireless sensor networking (WSN) standards, particularly mesh architectures that utilize time-synchronized channel hopping (TSCH) enable every node in the network to run on batteries or harvested energy without sacrificing reliability or data throughput. This frees application developers to put sensors anywhere–not just where power is available, but wherever the application requires sensor data.

These technologies go hand in hand to increase opportunities for application developers to deploy systems that require few, if any, battery changes, further reducing the lifetime cost of deploying wireless sensors and spurring the progress of the Internet of Things (IoT).

A 2012 study by ON World shows that the two attributes of a WSN that matter most to industrial customers are reliability and low power (Figure 1 ). Cost is third on the list: without solving the reliability and power issues, cost is not yet a customer priority.

Figure 1: Perceived Importance of WSN Attributes

Based on Dust Networks’ years of research and development of TSCH, it is clear that the combination of precisely synchronized time slotting, channel hopping, and an ultralow power radio enables the lowest power, most reliable WSNs. This focus on low power enables all nodes to run for many years on low -cost batteries, and also opens possibilities for a variety of energy sources, including energy harvesting supplies.

Low power radios
The introduction of the IEEE 802.15.4 standard created an excellent radio platform for WSNs. IEEE 802.15.4 defines a 2.4GHz, 16-channel spread spectrum low-power physical (PHY) layer upon which many IoT technologies have been built, including ZigBee and WirelessHART. It also defines a medium access control (MAC) layer, which has been the foundation of ZigBee. However, the single-channel nature of this MAC makes its reliability unpredictable.

To improve reliability, the WirelessHART protocol, also known as IEC62591, defined a multichannel link layer based on the 15.4 MAC to achieve high reliability (>99.9%), which is required for industrial WSN applications. In early 2012, a new version of the 802.15.4 MAC called 802.15.4e was ratified, and this MAC embodies multichannel mesh and time slotting. The typical power output for 802.15.4 compliant radios is around 0dBm, with transmit and receive currents in the 15-30mA range. Best in class transmit current at 0dBm is 5.4mA, and best-in-class receive current is 4.5mA (based on Linear’s LTC5800).

Time synching for low power and channel hopping
The original 802.15.4 MAC requires that the nodes in the mesh network that route information from neighboring nodes are always on, while nodes that only send/receive their own data, often called ”reduced function devices,” can sleep between transmissions. In order for every node in the network to be low power, communications between nodes must be scheduled, and it is necessary to have a shared sense of time in the network.

The tighter the synchronization, the less time the routing node radios must be in an ‘on’ state, which minimizes power consumption. Best-in-class TSCH systems synchronize all nodes in a multi-hop mesh network to within a few 10s of microseconds. Once there is a shared sense of accurate time in the network, and a schedule of time slots for pairwise transmission between nodes in the network, channel assignment can be incorporated into that schedule, thereby enabling channel hopping.Channel hopping to reduce interference and multipath fading
Thewireless channel is unreliable in nature, and a number of phenomena canprevent a transmitted packet from reaching a receiver; these can beexacerbated as radio power decreases. Interference occurs when multipletransmitters send simultaneously over the same frequency. This isparticularly problematic if they cannot hear each other, yet thereceiver can hear all the transmitters – the “hidden terminal problem”.

Backoff,retransmission, and acknowledgment mechanisms are required to resolvecollisions. Interference can come from within the network, anothersimilar network operating in the same radio space, or from a differentradio technology operating in-band–a common occurrence in the 2.4GHzband shared by Wi-Fi, Bluetooth, and 802.15.4.

A second,unpredictable phenomenon called multipath fading can prevent successfultransmission even when the line-of-sight link margin is expected to besufficient. This occurs when multiple copies of the transmission bounceoff objects in the environment (ceilings, doors, people), with eachreflected copy traveling a different distance. When interferingdestructively, fades of 20-30dB are common. Multipath fading depends onthe transmission frequency, device position, and on every nearby object;predicting it is practically impossible.

Figure 2 showsthe packet delivery ratio on a single wireless path between twoindustrial sensors over the course of 26 days, and for each of thesixteen channels used by the system. At any given time, some channelsare good (high delivery), others bad, and still others highlyvarying. Importantly, there was no period observed where a channel wasgood on all paths everywhere in the network.

Figure 2: Packet delivery across 16 channels over 26 days

Forthese reasons, it is critical that WSNs employ multiple channels. Bytime- synchronizing and scheduling the network into slots, transmissionscan be precisely scheduled on specific known channels, and the choiceof channel can change with every transmission. Furthermore, schedulingnetwork transmissions solves the ”hidden terminal problem” and virtuallyeliminates in-network collisions. Such a mechanism is field proven inthe more than 10,000 WirelessHART networks in the field, which routinelyachieve multiyear battery life and > 99.9% reliability.

Energy harvesting considerations
Oncethe power requirements of the WSN are suitably minimized, the choice ofpower source broadens. Ambient energy is everywhere: light, vibration,and heat are but a few examples of energy that can be freely sourced andconverted to sufficient electrical energy to run a low power TSCH WSN.The following examples illustrate some practical energy harvestingtechnologies that generate more than 150µW of power–more than enough torun a typical IPv6 routing node in a 802.15.4e network (for example,the Dust Networks’ SmartMesh IP).

Lighting. Mostareas of a typical office building have sufficient indoor light to run alow power TSCH WSN. According to the United States General ServicesAdministration, which sets the guidelines for U.S. public buildings, themore brightly lit areas, such as workstation areas and readingsurfaces, have 500 lux of lighting. Even in areas considered “normallylit”, such as lobbies, stairwells, and mechanical and communicationsclosets, there are at least 200 lux of light and 300 lux is common formost conference rooms. With 200-300 lux of light, there are a number ofindoor small photovoltaic cells available (e.g., G24i 4100 low lightsolar panel, or Sanyo AM-1815 indoor cell) that can supply sufficientpower to operate an IPv6 router in a 802.15.4e TSCH network.

Thermal Energy. ThermalElectric Generators (TEGs) produce power from the heat dissipation fromhot surfaces, such as waste heat from common devices normally thoughtof as very warm such as computer monitors or high-current motors. Aswireless solutions become more power efficient, the energy produced fromcommonplace temperature differences of as little as 10ºC becomes usableas an energy source. For reference, the typical difference betweeninternal body temperature and room temperature is about 15º C.

Manyenergy harvesting transducers produce only a few hundred millivolts ofoutput, so step-up voltage DC/DC converters are often required toconvert to a usable supply voltage range. ICs such as LinearTechnology’s LTC3105 offer maximum power point control so that the transducers operate atpeak efficiency and enable the addition of a battery backup to thecircuit. Since the batteries in these circuits are only used when theambient energy source is insufficient or absent, battery life can beextended dramatically, thus reducing the costs associated with changingbatteries. The inclusion of battery backups to an energy harvestedcircuit can provide added assurance and power continuity if the energysource happens to be intermittent, for example if lights or machineryare turned off for the weekend.

Summary
The realizationof the Internet of Things is accelerated by making it practical andeasy to deploy sensors ubiquitously. Low power, reliable wireless sensornetworks translate to no wires/no worries for customers and developersalike. Time synchronized, slotted multichannel systems confercustomer-critical benefits to WSNs: reliability and network-wide lowpower operation.

The WirelessHART and 802.15.4e standards areexcellent embodiments of this networking approach. Low power operationensures great flexibility in the choice of power source, and offers thepotential for perpetual power. These factors all add up to making iteasier and more practical to put sensors anywhere.

Joy Weiss is president, Ross Yu is product marketing manager, and Jonathan Simon is Systems Engineering Director of the Dust Networks Product Group at Linear Technology .

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.