Last two years, “5G” is mostly heard slogan on the mobile market. Vendors, governments, journalists and finally users are fascinated about the promising benefits of 5G technology coming to the mass market.
One of the myths which are widely heard is, that 5G mobile network will allow the cars/vehicles to be remotely driven from home/office.
Let’s verify this thesis using Dr Watson rather than Mr Sherlock Holmes methodology, and have a look at the facts and plans of that 5G industry.
Before we get to remotely control car over 5G network, we can already seat comfortably in a car and enjoy its self-driving.
The meaning of “Autonomous car” by first glance, suggests that it can drive and control the way it drives autonomously as well as independently from human, which means can be driven by the system. Let’s check what does it mean and how it is defined. The definition was delivered a few years ago by SAE (Society of Automotive Engineers).
Table 1, presents six levels, where the first three define Human driver and three define Automated driving by systems. We can draw a few conclusions. First, personal, our cars are already at SAE level 1, as Cruise control is commonly delivered system by car manufacturers.
Second, more important for our analysis, is that we are considering here vehicles on SAE level 3-5, which means partially or fully self-steering, monitoring driving environment, performing Dynamic Driving Tasks as well controlling all of Driving Modes. The system can be e.g. the computer mounted inside the car (AI based), but it can be also remote control of vehicle via mobile network, by system or human. In the first case, we still talk about the level of SAE 5 automation, but in the second case, human remotely driven can be SAE level of 1-4.
If we look at the today’s car producer market, we can find at least one car producer delivering solution on SAE level 5, Tesla cars (it is at least Tesla’s opinion about the level 5).
Tesla cars are already driving itself by car’s internal system using signals from sensors (radars, lidars, cameras, near-field sensors, etc.), onboard computer (software supported by AI technology) as well as communication with servers (downloading Maps, checking weather condition and traffic information).
Does it mean that future of 5G mobile network on the car market is already lost for the benefit of self-driving cars only? on the other hand, do we really need to wait for 5G mobile network to get remotely driving vehicles?
Let’s have a look at standardization bodies, which are intensively working on vehicle connectivity based on V2X standard, as it may help us to understand where is the future of 5G mobile network in the car industry. Let’s check also how much the standards are going to be used for support of driving by human and how much going to change the role of driving by a human.
Standard bodies engineering systems/protocols for V2X communication.
The idea of communication between vehicles and road environment was formed more than 20 years ago, however, the real works on standards, which also define use cases like (avoiding collision, verification of pedestrian, etc.) were documented since 2014.
Two standard organisation DSRC (dedicated short-range communications) in the US and ITS-G5 (intelligent transportation system) in Europe were formed to cooperate on short-range radio solution. The communication between vehicle is based on dedicated for that task 5.9GHz band. In Europe, the standard is developed by ETSI organisation. Both standard suites are based on IEEE 802.11p (WiFi) technology, however, it implies many limitations (range, quality of communication). This technology due to its short range of communication was intended mainly for delivering infrastructure to cities.
In that case, there was no doubt that the standard should be also developed based on a long-range communication technology, like 4G (LTE) as well as 5G in the future. The leading role in the standardization of that topic was taken by 3GPP and 5GAA organisation.
3GPP, however, did not limited standardization only to long-range communication (LTE-Uu), but decided to prepare standard directly competing with IEEE 802.11 on the short-range radio communication – so-called Direct PC5 interface (PC5 or alternatively PC5 and LTE-Uu both, in hybrid mode). PC5 does not require mobile network assistance in Mode 4 (Distributed radio resources management). PC5 Mode 3, works with the assistance of long-range network communication (Network-assisted resources management).
The obvious benefits of the mobile network-based solution are the support of QoS (also QoE – Quality of Experience, especially in case of 5G), larger coverage, lower infrastructure costs (especially when eNB /NR will be playing a role of RSU), high authentication security (SIM) and higher data rate for moving vehicles. It gives the possibility to develop more use cases than just short-range communication between near vehicles/infrastructure. It open also the possibility for business cases like multimedia (MBMS over LTE-Uu standard), as well as the possibility of remote driving.
The battlefield between IEEE and 3GPP is currently under development, however, more and more researches predict that the 3GPP standards are more suitable for the future as it allows to cover much more use cases. It is visible that PC5 Mode 3 will be used in the LTE/5G coverage zones and PC5 Mode 4 out of coverage where the vehicles can relay only on direct communication.
Looking at today’s ongoing projects in the world, mostly in Europe (Germany leading), we can see hybrid solutions used. The future will shows which technology is gonna be a winner. In my opinion, the 3GPP standard, supported by Qualcomm/Huawei/Samsung and its chipsets, will show the way forward. Qualcomm already delivers the first type of chipset for trials and initial commercialization – Qualcomm® 9150 C-V2X. It supports V2V, V2I and V2P, based on 3GPP Release-14.
The same processor Qualcomm® 9150 is going to be used by the RSU vendors to deliver V2I infrastructure devices. First devices of V2I, size of bigger WiFi box are already delivered to the market by few manufactures: CalAmp, Commsignia, Kapsch, Leonardo, Savari, Wieson, etc.
How V2X communication works.
V2X standard assumes direct (or indirect) communication between another entity on the street. It can be vehicles (V), pedestrians/bikes/motorbikes (P), infrastructure – street lights/cameras (I) as well as network (N).
The V2X entities are exchanging many parameters with near or long-range traffic participants: speed, location, warning signals, and video and lidars parameters, in the future. Based on this information, all participants can cooperate in one-to-one or one-to-many templates. Basic use cases are presented in Table 4 (3GPP standard].
3GPP defines 4 basic V2X interfaces: V2I, V2V, V2P and V2N. Details are described in Table 2 [TR 22.185 Release 14]. There are also indirect variants presented in the same Table 2.
We can see e.g. the group of V2X safety applications: Forward Collision Warning, Control Loss Warning, V2V Use Emergency Vehicle and Stop Warning, Warnings about pedestrian/cyclist collision.
There are also V2X non-safety applications: Cooperative Adaptive Cruise Control, V2N Traffic Flow control, which will help to steer the traffic flow of the streets as well as impact saving of oil and limit of pollution (Co2) and street heat.
Looking at the 3GPP Release 14, we can see plenty of use cases, which shall help the participants of the streets, to feel saver thanks to preventive communication. We shall recognize an extensive list of features preventing Pedestrian/cycles safety.
We can expect first V2X products by 2019-2020 in vehicles supporting V2V, V2I as well as user terminals/mobile phones supporting V2I, V2P interfaces.
Does it mean, that there is no, plan for remote driving communication? Let’s check than what 3GPP engineers are cooking in a frame of 5G mobile network world [TR 22.186].
Future of V2X
When we open freshly cooked 3GPP Release 16 standard documents, or read 5GAA (5G Automotive Association) road-map, we can find a very interesting list of extensions to Release 14 (3GPP Release 15, delivers only PC5 improvements, traffic aggregation, 64QAM, diversity & short TTI).
3GPP Release 16 delivers interface NR-V2X which is defined for support of 5G NR (New Radio) radio systems but including complete backward compatibility to LTE-V2X (standard approval is planned for December 2019).
Qualcomm, the main player of the chipset required for that revolution, predicts commercial implementations of 3GPP Release 16 in the middle of the year 2021.
For our analysis, the most important is the use cases defined for a future of autonomous driving: Vehicle Platooning (a synchronized convoy of cars), Cooperative Operation, Sensor sharing, Remote Driving and Advanced Driving. A detailed list of planned use cases is presented in Table 5.
Table 3, specify planned technical requirements for all 4 groups of future use cases. There is a plan for the remote driving!
Does it mean we will soon, seat at home and drive remotely vehicles escorting children to school and eat breakfast at the same time?
In the body of the standard document we can read what were the intentions of engineers for that application [TR 22.886]:
Remote driving is a concept in which a vehicle is controlled remotely by either a human operator or cloud computing.
While autonomous driving needs a lot of sensors and sophisticated algorithm like object identification, remote driving with human operators can be realized using less of them.For example, if the on-board camera of the vehicle feeds the live video to a remote human operator, a human operator can easily understand the potential hazard of the vehicle without the assistance of any sophisticated computing. Based on this video, the remote operator sends commands to the vehicle.
Remote driving can differently use case than autonomous driving. Buses follow pre-defined static routes and a specific lane and stop at pre-defined bus-stops. Thus, the characteristic of operating these buses are somewhat different from what is required for operating autonomous vehicles. For these buses, live video stream includes not only the outside-bus image but also inside-bus image, so remote operators additionally need to react to a more diverse scenario such as passengers getting on/off the bus.
The intention of engineers is clear, to use it in any case. Is it practically possible using 5G mobile network?
Looking at Table 3, and defined Data rate for Remote Driving application, we can see demand for 25 Mbps UL (the assumption is that H.265/ HEVC HD stream up to 10 Mb/s on two video streams are delivered to a remote driver, for an absolute speed of up to 250 km/h) with latency on level of 5ms. Now, imagine (numbering case is taken from the same 3GPP standard) 9 840 cars are considered per 1 km in the scenario with high vehicle density related to congested traffic road on US Freeway with 5 lanes in each direction (or 10 lanes total per highway), and up to 3 highways intersecting.
Let’s assume 50% of those vehicles uses Remote Driving feature consuming demanded data rates. A simple calculation shows it requires about 123 Gbps UL throughput capacity on a 1 km long Freeway with 5ms latency (not included Mobility Management services like handover etc.). How many eNB/NR shall we deploy on 1 km Freeway? I do not even try to do calculate it, definitely a lot. However, if we assume only 1% of vehicles with RD feature, we need about 2,46 Gbps, what is much more practical.
It is not realistic, from planning and deployment perspective to cover the roads in the city or in the urban area with so high capacity, due to costs (mobile towers, radio systems, transmission to each location), interference levels as well limitation of electromagnetic field secure levels.
Today, we can hazard the answer to the initial thesis: Remote driving of vehicles over 5G mobile network will be possible, but only for limited use cases. It can be an emergency car, police, cars delivering some hazardous items.
Hard works ongoing on the standardization 3GPP of Release 16, is in my opinion very interesting engineering work, which may imply changes to the future behaviour of all street users.
Definitely we shall be happy that the 3GPP Release 14, V2X standards are soon coming to our environment. I personally mostly like the idea of equipping the user with terminals with V2P and V2I applications. The set of features promise an increase in street safety for pedestrians and bike/motorbike users. It will also help in street traffic control once the vehicles with V2V and V2N features will become a significant part of the traffic.
If you want to learn more about these topics, I suggest to visit first 5GAA Automotive Association (http://5gaa.org) and look at the 3GPP standards (https://www.3gpp.org/) mainly TR22.185, TR22.186, TR22.885, TR22.886, TR23.285.
Detailed use cases:
- If you like the text – share it with your friends.
- Please leave a comment. It is a source of guidance for me.
- You can follow me on LinkedIn and Facebook.