by Martha Zemede, Keysight Technologies
Proliferating wireless standards put heavy demands on instrumentation. Here are some factors to consider when gearing up for IoT development work.
You can say one thing about the Internet of Things scenarios being trumpeted in the media. All the applications and services being envisioned rely on networks of sensors and actuators, often linked by radio. And because IoT applications are so diverse, no single radio technology can effectively address all the needs of this evolving industry.
The range of wireless technologies available for use in the IoT is diverse and growing more so every day. Currently, there are more than 60 legacy and new RF formats in use for M2M- and IoT–related applications. Near-field communication (NFC) will handle mobile payments; geosynchronous satellites will handle communication with unattended remote weather stations. Bluetooth, wireless LAN (WLAN), ZigBee, point-to-point radio, cellular, and other technologies will all have IoT roles.
An IoT network will need to cope with all kinds of special devices having different communication requirements. At one end will be simple wireless devices, such as battery-powered sensors and actuators that will transmit minimal data while operating unattended for several years. At the other end of the spectrum will be mission-critical services and devices that require constant, reliable and super-secure connections.
Key to uniquely identifying each device is a vast IP address space. One problem is the current Internet Protocol version 4 (IPv4) addressing space is too limited, so it requires the use of concentrators (for example, routers and gateways). The most recent version of the Internet Protocol is IPv6, which will be a key enabler for IoT devices. IPv6 uses a 128-bit address, theoretically allowing 2128, or about 3.4×1038 addresses. The total number of possible IPv6 addresses is more than 7.9×1028 times as many as with IPv4, which uses 32-bit addresses and provides approximately 4.3 billion addresses. The two protocols are not designed to interoperate, but several IPv6 transition mechanisms have been devised to let IPv4 and IPv6 hosts talk to each other.
IPv6 provides other technical benefits in addition to a larger addressing space. Device mobility, security and configuration aspects have been considered in the design of the protocol.
Server/cloud-based big-data analytics and machine learning play a role in the majority of IoT business models. IoT devices at the end nodes connect to the cloud or server for intelligence and analytics. Some connect directly, but often with gateways.
Gateways aggregate traffic from less trafficked networks onto higher capacity LANs and WANs. They typically include greater power supply and computing resources than end-nodes (things). Edge or fog applications running in gateways offload processing from both cloud and end-node sensors and actuators. End-nodes are often designed to have a long battery life, necessitating the efficient use of embedded computers and radio transmission. Intelligent threshold triggers in gateway applications make traffic more efficient by passing actionable information to central cloud servers.
Gateways interface with the cloud and end-nodes through a heterogeneous mix of wireless technologies, both cellular and non-cellular. Radio interfaces address varying application needs depending on coverage, latency, throughput, energy efficiency and cost.
As an example, some home-automation applications use smartphones as a gateway. The wide availability of WiFi makes it the first choice for many IoT applications. When WiFi links are unavailable, cellular protocols are frequently substituted. In wearable applications, Bluetooth is often the choice. NFC is the natural choice when security is aided by proximity. ZigBee, Z-Wave and Thread offer robust, low-power mesh networks for home automation and smart energy devices.
It is useful to consider IoT technologies grouped by operating range. NFC is a short-range system based on ISO 14443 at 13.56MHz. Perhaps best known for use in mobile payment systems, NFC devices can behave as terminals, also called proximity-coupling devices (PCD) or readers. They may also behave as cards, also known as proximity inductive-coupling cards (PICCs) or tags; cards are often powered by the RF field generated by the terminal.
In the IoT space, Bluetooth low energy (BLE) is getting a lot of interest. Designed for lower data throughput, it consumes significantly less power than Bluetooth devices and operates for years using coin-cell batteries. It supports simplified models for device discovery, service discovery and data exchange in ways that use little airtime and consume little power. This lets BLE serve in small devices such as watches, health monitors and battery-powered appliances.
A number of short-range wireless technologies use a standard called IEEE 802.15.4 as the physical (PHY) and media access control (MAC) layers. For protocols that include ZigBee, Thread, WirelessHART and ISA100.11a, the developer of the higher layers specifies the higher-level protocol appropriate for the target application. This low-rate wireless personal area network (LR-WPAN) supports rates that range from 20 to 250 kbps. It is designed for home networking, industrial control and building automation, all of which need low data rates, low complexity and, in many cases, long battery life.
ZigBee devices can connect, exchange information and disconnect quickly before returning to sleep mode. One key attribute is the use of a mesh network topology that can include thousands of nodes. ZigBee radios operate with low duty cycles, so their applications run for years on inexpensive batteries. Target applications include smart energy, home automation, healthcare, retail and lighting control, each of which has a specific ZigBee profile and certification.
Thread technology is similar to ZigBee in that it is based on the IEEE 802.15.4 PHY and MAC, but it uses the IPv6 over low-power wireless personal area network (6LoWPAN) protocol. It’s an encrypted mesh network designed to connect hundreds of home-automation products and devices. The network is self-healing and is configured such that there is no single point of failure. Its short messaging conserves bandwidth and power, while a streamlined routing protocol reduces network overhead and latency.
WiFi is the most widely used wireless Internet connectivity technology, with 802.11a/b/g/n being most common. Two recent amendments to the PHY layer address the need for high throughput data rates: 802.11ac, which operates below 6 GHz and is becoming the standard in mobile phones, tablets and PCs; and 802.11ad, which operates in the 60-GHz band. An upcoming version called 802.11ah (HaLow) is intended to support low energy for IoT applications. It uses low power and low data rates. It operates in the sub-gigahertz band and has a range of up to 1 km.
802.11p is also called wireless access in vehicular environments (WAVE). It was created specifically for applications such as telematics, roadside assistance, fleet management and young-driver insurance validation. In the future, 802.11p will also enable vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) connectivity for tasks such as vehicle safety, traffic management and toll collection.
The two main wireless neighborhood area network (WNAN) technologies are Wi-SUN (IEEE 802.15.4g) and ZigBee-NAN. The aim of this standard is to provide a framework that facilitates large-scale process-control applications. Wi-SUN provides a low-rate wireless network capable of supporting large, geographically dispersed smart utility networks with minimal network infrastructure, with potentially millions of fixed endpoints.
ZigBee-NAN is a recent extension that has an operating range to 1 km and supports end-to-end IPv6. The ZigBee Alliance is currently working with the Thread Group on interoperability for home and commercial sensor networks.
A lot of innovation is happening in low power wide area (LPWA) networks. For applications with low data rates and low duty cycles, LPWA extends battery life, reduces cost and improves link budgets compared with currently deployed cellular formats. LPWA systems, such as LoRa and SIGFOX, are being rolled out nationally in some countries using lightly licensed or unlicensed spectrum. Anticipating strong growth in low-power M2M applications, 3GPP radio access network (RAN) working groups are developing cellular protocols to support LPWA in licensed spectrum.
3GPP Release 12 (Rel-12) introduced a new low-complexity device category (Cat-0) for LTE machine-type-communication (MTC). Cat-0 improves efficiency for low-data-rate applications as a stepping stone to more significant advances. The newest 3GPP Rel-13 includes enhanced-MTC (eMTC) Cat-M1; a 1.4-MHz bandwidth optimization of LTE. Also included is Extended Coverage General Packet Radio Service (EC-GPRS); using retransmission and other protocol updates to improve link budgets. Finally, it specifies a narrow band IoT (NB-IoT) also referred to as Cat-M2; this is a new radio format optimized for LPWA applications.
Tasks associated with design and test become all the more difficult given the challenges of IoT work. For example, as the IoT becomes more pervasive, design engineers must work harder to maximize power efficiency, manage electro-thermal effects, and deal with the more severe electromagnetic coupling that results when designs become more compact. Additional hurdles will include evaluation and selection of the best technology mix (GaAs, GaN, SiGe/Si/SOI, CMOS), as well as integration of subsystems and verification of performance relative to industry standards.
And as designs become more complex, circuit simulation becomes more difficult. The electronic design automation software for designing and simulating new IoT devices must handle the challenges inherent in new communication formats. The best approach is to simulate new devices early in the development process and give system architects and algorithm developers the freedom to innovate at the PHY layer of wireless communications systems. Instrumentation supporting this approach should also include virtual measurement tools that can attach to nodes in the simulation to provide a view of expected performance.
Exploring Solutions for Design and Test
As the design moves from simulation to hardware, physical device modules can substitute into the simulation, with real measurements or hardware-in-the-loop replacing virtual tools. This practice allows developers to compare simulated and actual performance.
The next step is to measure and analyze the design. When choosing a test instrument for this purpose, typical selection criteria includes performance specifications, measurement speed, physical footprint, configuration scalability and cost (upfront and ongoing). No single instrument setup will be best for all needs.
The IoT will rely on simple battery-powered sensors and actuators that transmit little data while providing years of unattended operation. The design and development of such devices employs tools that can measure battery drain during three main conditions: sleep mode, idle mode and transmit mode. For signal creation and analysis during the design phase, the preference is for benchtop instruments that are general-purpose (for example, swept-tuned spectrum analysis and functions supporting signal analysis and troubleshooting).
Later in the product lifecycle, criteria like test speed, flexibility and footprint are more important. Here, modular and one-box testers are better candidates. Important software includes that which can synthesize and analyze custom and standard-compliant test signals for wireless communications formats, including cellular, IEEE 802.11 variants, Bluetooth, ZigBee and Wi-SUN. In the best of all worlds, the solutions for simulating, designing and testing IoT devices will be integrated to enable product feedback across the entire lifecycle.
All in all, design and test instrumentation will ensure IoT device designs will allow reliable connectivity.