5G Radio Access for Ultra-Reliable and Low-Latency Communications
Industrial process automation and remote surgery are examples of use cases that will be possible in future 5G systems. We call them mission-critical machine type communication use cases and they require communication with very high reliability and availability, as well as very low end-to-end latency. We can show that it is feasible to design a 5G radio interface which is capable of providing this.
Machine-to-machine (M2M) communication will make up a large part of the new types of services and use cases that 5G systems will address. From a communication technology perspective, M2M can be divided into two main categories: Massive machine-type communication (MTC) is about connectivity for large numbers of low-cost and low-energy devices in the context of the Internet of things; mission-critical MTC is envisioned to enable real-time control and automation of dynamic processes in various fields, such as industrial process automation and manufacturing, energy distribution, intelligent transport systems – and requires communication with very high reliability and availability, as well as very low end-to-end latency going down to millisecond level.
Massive MTC enhancements and enablers are already in the scope of current standardization, both in 3GPP and IEEE. Mission-critical MTC, however, is still in the early-development phase, and there are a lot of challenging research problems to solve. In Ericsson Research, we have accepted this challenge, investigated the feasibility requirements and proposed enabling solutions for mission-critical MTC in line with all the 5G technology components and the architecture we are working on.
The requirements of mission-critical MTC have to be fulfilled in three dimensions: latency, reliability and availability.
Latency refers to the time delay between data being generated – e.g., at a sensor – and the same data being correctly received – e.g., by the actuator – as illustrated in the figure below. The most stringent requirement on the end-to-end latency may be 1 ms, as explained for example in ITU-T Technology Watch Report on the Tactile Internet.
Figure 1: Latency distribution based on Tactile Internet ITU-T Watch report numbers. 100 µs can be assumed as the air interface delay per direction.
Reliability refers to the capability of guaranteeing successful message transmissions within a defined latency budget – or delay. The reliability requirements vary among different mission-critical MTC services, but may go down to one per billion messages as shown in an ETSI report – TR 102 889-2. As an example, in industrial automation, only one message in one billion data transfers may be lost or delayed within the given latency budget.
System availability has to ensure that critical applications are not in outage when they are needed. To equate availability of wireless and wired solutions, an availability of 99.999% can be sought.
All in all, if we want to address all the potential use cases of mission-critical MTC, the radio technology that we design should be scalable for these stringent requirements. Then, the questions are expectedly: “How can these requirements be fulfilled with a minimal complexity introduced to the radio access design?” and “What would be the practical limits of a mission-critical MTC system for a typical application?”.
To respond to the first question, let’s take 3GPP Long Term Evolution (LTE) as the baseline. Like in LTE, we have chosen the air interface based on orthogonal frequency-division multiplexing (OFDM) – so as to avoid inter-symbol interference, provided that the cyclic prefix of an OFDM symbol is longer than the delay spread of the channel. To enable the lower delays and higher reliability discussed above, we have identified a number of modifications to LTE that would be needed:
- Reduced transmission time intervals, e.g., down to 100 μs, and shorter OFDM symbol durations enabling fast and efficient data transmission
- Redesign of physical channels allowing early channel estimation
- Use of convolutional codes (e.g., for data channels) and block codes (e.g., for control channels) providing fast and reliable decoding
- Implementation of high diversity levels improving the reliability of signal detection and decoding, as well as availability
Then to the second essential question: “What would be the practical limits of a mission-critical MTC system for a typical application?” To answer this, we have modeled a simulation scenario based on a factory floor layout with the dimensions 100 m x 100 m. We assume that the industrial application requirement dictates only one message in one billion data transfers may be lost or delayed within a 100 µs air interface delay. The simulation results, depicted below, show that the mission-critical MTC system, with aforementioned enablers, is able to support thousands of industrial automation machines spread out in a factory floor (e.g., for uplink data transmissions). We also found that the increased diversity improves the robustness of the system against unexpected or unmitigated interference sources. This is also shown in the figure below, by the interference-plus-noise ratio (INR) metric which refers to the relative interference-to-noise power in the channel. Therefore, we highlight the diversity as an essential tool for the availability and reliability in mission-critical radio access.
Figure 2: Mission-critical MTC capacity per access point at 5.2 GHz carrier frequency with 100 MHz bandwidth. We assume that the packet size is 100 bits and the sensor updating period is 1 ms. Two device antennas and eight base station antennas are used to enable diversity. The robustness against interference is shown in terms of interference-plus-noise ratio (INR).
So we can conclude that it is feasible to design a 5G radio interface capable of providing sub-millisecond radio transmission with a failure rate down to 10-9, with minimal additional complexity. Our teams in Ericsson Research continue to seek further enablers for mission-critical MTC. Recently, we investigated the details of practical challenges of the factory automation use case discussed here. The results of that will be published soon. Make sure to follow us and stay tuned!
Osman N. C. Yilmaz, Niklas A. Johansson Ericsson Research