5G Internet of Things and Standardization

5G goes beyond traditional cellular services for personal use. A large chunk of traffic will come from human-to-machine and machine-to-machine communication. The internet-of-things (IoT) requires new services, standards and innovations. 

Industry Trends and Requirements

The number of devices that need to be wirelessly connected in industries such as retail, healthcare, manufacturing, transportation and agriculture will be much greater than the human population. The number of machine-to-machine connections will easily be in the hundreds of billions. However, requirements for machine-to-machine communication in each industry will be drastically different: some industries will require a very high data rate, some will require very short latency, and others will require extreme reliability. This creates great challenges for 5G networks. Considering service ubiquity and a massive number of connections, a 5G network infrastructure has to be very densely deployed. If efficiency remains at current levels, more energy will be consumed.
The cellular industry is a major contributor to global carbon dioxide emissions. Researchers need to devise smarter ways to reduce harmful interference, shrink circuits, and reduce power consumption. Increasing energy efficiency involves more cost-effective site planning, construction, and maintenance. These account for a large proportion of energy consumption in a cellular system. 5G terminals also need to be energy-efficient. To achieve a 10- to 100-fold increase in the density and number of connections, system designers and device manufacturers have to cooperate closely.

Some IoT applications, such as meter reading, involve collecting information at low speeds. Other IoT applications, such as video monitoring, involve collecting information at high speeds. Meter reading is characterized by a massive number of connections, low-cost terminals, low power consumption, and a large number of small packets transmitted. With video monitoring, there is a bottleneck in the uplink, where high data rates are required, and traffic density is high. Control-type machine-to-machine communication may be delay-sensitive or not delay-sensitive. Vehicle-to-vehicle communication is a typical delay-sensitive service with millisecond-level latency constraint. It also requires extreme reliability; e.g., nearly 100% success rate for decoding when the speed of passing vehicles is about 300 km/h each. There are many IoT applications in daily life that are not delay-sensitive. The challenge with these applications lies in the huge number of devices and connections.

Standardization

IoT network solutions have been specified by several 3GPP working groups over the past five years (Fig. 1). In the 3GPP-SA group, machine-type communication (MTC) devices can be grouped by MTC features. The serving time for delay-tolerant or infrequent-transmission features can be tightly controlled. For example, the time that a device is not permitted to access the network can be specified. This length of time is an MTC feature that depends on an operator's strategy. Overload can also be controlled by rejecting connection requests. A RAN node can be permitted to differentiate its low priority MTC traffic/signaling from other RAN nodes. A "delay tolerant" indication can be added to the RRC connection procedure for MTC devices with delay-tolerant access. An eNB can reject an RRC connection according to whether a core node is overloaded, and the eNB can add "extended wait time" in the RRC connection reject/release information to request MTC devices to wait before sending a new connection request. In the overload start information, a new parameter called "barring factor" is introduced for different types of M2M services. This enables the RAN to select appropriate parameters.

In Rel-12 RAN2, there are severe collisions in a random access channel (RACH) when numerous MTC devices try to access the RAN. Given that the RACH is the most vulnerable channel at the MAC layer during congestion, RACH overload control is a major area earmarked for improvement. Several schemes have been proposed: 1) MTC specific back-off, 2) dynamic allocation of RACH resources, 3) separating RACH resources for MTC, 4) slotted access, and 5) pull-based scheme. In the RAN2 specification, another objective is to identify and evaluate the mechanism to increase the ability of the RAN to handle traffic profiles comprising small data transmission. Data can be generated by both M2M and non-M2M devices and applications. Signaling overhead reduction and UE power consumption are two main areas in this study.

In Rel-11 RAN1, low-cost MTC has a very clear business goal: LTE M2M devices are designed cheaper than GSM M2M devices so that M2M traffic can be shifted from GSM to LTE. To be more competitive, LTE M2M devices also need to be energy efficient and have enhanced coverage. Moreover, low-data rate channels of LTE M2M devices, such as channels with 118 kbps in the downlink and 59 kbps in the uplink, have to coexist with other LTE channels. The reference for the study is LTE category 1 UE for MTC devices. Assuming a single-band operation, six areas are studied:
●  reduced maximum bandwidth supported by a UE. Both RF and baseband bandwidth can be reduced, for example, to 1.4 MHz at the cost of reduced frequency diversity. ePDCCH can be used when the network operating bandwidth is wider than that of M2M low-cost devices.
●  single RF chain receiver. This reduces coverage for various downlink physical channels by 3 to 5 dBs.
●  reduced peak rate. This can be achieved by limiting the maximum block size, i.e., < 1000 bits, or the number of physical resource blocks (PRBs) allocated each time, i.e., <= 6, or the modulation order, i.e, QPSK only.
●  reduced UE maximum transmit power. This directly reduces the coverage.
●  half-duplex operation. This affects scheduler implementation at eNB.
●  reduced downlink transmission mode (TM), i.e., limited to TM1 and TM2.
In Rel-13, two physical-layer solutions are being standardized: eMTC and narrow-band IoT (NB-IoT). eMTC is an enhancement on top of low-cost MTC in Rel-12. Physical control channels and broadcast channels are all confined within 6 PRBs, so a building penetration loss greater than 20 dB has to be supported. Since most control channels are redesigned, much effort has been put into the specifications. NB-IoT was originally studied in GERAN with the purpose of using refarmed GSM bands for low-cost, good-coverage MTC devices. Narrow band similar to a 200 kHz GSM channel is used for extreme coverage. The two strongest candidates in the study were NB-CIoT and NB-LTE. Shortly after the standardization work was moved to the RAN, NB-LTE was accepted and NB-CIoT was abandoned.

 
Potential Technologies for 5G IoT

The most challenging requirement in this scenario is support for a massive number of devices. This means that the cost of a terminal should be significantly lower than that of a mobile device. The power consumption has to be low enough so that devices can be battery-powered for years without recharging. Also, coverage should be robust enough so that even devices inside basements can connect to the network. New enabling technologies for IoT are non-orthogonal access and control channel optimization, advanced channel coding, and short radio frames.

Non-Orthogonal Access and Control Channel Optimization

Non-orthogonal access enables multiple users to simultaneously access the network using the same frequency resources. The access can be contention-based, which significantly reduces control overheads for granting the resource and indicating the transmission format. Control channel optimization can reduce signaling overhead when many devices connect to the network.
Multi-user shared access (MUSA) is a non-orthogonal multiple-access scheme in the code domain. Conceptually, each user’s modulated data symbols are spread by a specially designed sequence that enables robust successive interference cancellation compared with the sequences employed by traditional direct-sequence CDMA (DS-CDMA). These spread symbols are then transmitted concurrently on the same radio resource by means of shared access. Finally, each user data from superimposed signals can be decoded at the base station side using SIC technologies. The design of spreading sequence is crucial to MUSA because it determines system performance as well as the interference between users. Short-length spread sequence with multi-level complex values can have low cross-correlation, which is helpful to MUSA.
Grant-free transmission can be used to minimize control signaling overheads where users generate the spreading sequence locally without coordination by base stations. To keep signaling overhead low, no closed-loop power control is implemented. Such control is usually implemented to compensate for fast fading.

Advanced Channel Coding

Small-packet transmission is a characteristic of M2M communication. However, during the standardization of 3G and 4G cellular systems, much effort is spent on designing codes that approach the Shannon limit and perform well when the block is large. When the block is short, CRC bit reduction is an effective way of reducing overhead and can be used for both turbo and polar codes. A small block is often associated with low SNR, and advanced coding schemes can provide more powerful error protection in extreme coverage scenarios. Those coding schemes not only optimize code structure but also enhance hybrid acknowledgment repeat requests (HARQ) to better combat channel fading.

Short Radio Frames

5G-enabling technologies are new physical channel design that can reduce a single transmission delay. Such new design focuses on short radio frames that require a whole new set of subframe structure including reference signals, resource mapping, and HARQ.