By Varadaraj Yatirajula
There has been interest in the motorsports industry to provide 5G connectivity for the cars on track. Typically, the cars use legacy technologies like Wi-Fi or DVB (Digital Video Broadcast) to provide the connectivity to backhaul telemetry, video, and audio data from the cars back to the pit and media centre for broadcasting. However, the current solutions provide a limited bandwidth, and throughput is in the range of 10 to 12 Mbps. This throughput is not sufficient for advanced use cases and capabilities like 4K streaming, 360-degree video and telemetry for the racing cars.
The automotive industry is also moving towards a Software Defined Vehicle (SDV) strategy, where the cars will need a highly reliable and low latency connectivity for frequent updates to the software, and to push gigabytes of data from the hundreds of different sensors on the cars to the pits, where the data analysis is used to derive valuable insights into the performance of the cars. The move to electrification and softwarization also needs a connectivity solution which is future proof ready for these requirements and remains relevant for the next decade.
Using private 5G networks to provide connectivity for the motorsports industry cars is an attractive proposition. They can help overcome and mitigate some of the issues involved in using wireless communications technologies for the use cases in motorsports.
In this article, we will explore the feasibility and design considerations to provide 5G connectivity for Motorsports.
The following figure shows the proposed ORAN Disaggregated Architecture for the motorsports use case. The solution uses the 7.2 split architecture which is modular and extensible.
In this architecture, the Distributed Units (DUs) run some of the physical layer algorithms besides the Radio Link Control (RLC) and Media Access Control (MAC) layers. The DUs are housed in the Track Side Units (TSU) which also provide the power for the DUs.
The Centralized Unit (CU) runs the Radio Resource Control (RRC) and Packet Data Convergence Protocol (PDCP) layers.
There is a redundant 25G fibre ring to backhaul the traffic from the DU to the Centralized Unit (CU). The CU and 5G core are centralized at the event technical centre.
This split architecture makes the solution scalable as the number of radios will depend on the track layout and dimensions. The CU can control multiple DUs and run the required algorithms specific to the motorsports use cases.
Some of the design considerations and proposed solutions when using 5G to provide the throughput requirements for racing cars are described below.
The cars move at a very high speed on track of up to 350 kilo metres an hour. This high speed introduces unique design challenges in the design of a 5G solution. The Doppler shift or Doppler spread is a concern because of the there is a shift in the frequency of the signals sent and received to the fixed radios from the moving cars. The Doppler shift is directly proportional to the speed of the car. This also results in inter carrier interference.
Solution: Algorithms need to be designed to compensate for the Doppler shift introduced due to the speed at which the cars are travelling. The Doppler shift depends on the speed and direction of movement of the car relative to the trackside radio. Thus, the Doppler results in an increase or a decrease in the frequency, depending on the direction and speed of movement of the car relative to the radio. The transmit frequency is adjusted in the RF front end to compensate for the Doppler shift.
Several racing tracks for example in Formula One are in cities, where there is infrastructure like buildings, fences, trees, grandstands, and street furniture. This infrastructure causes interference and scattering of the 5G radio signals to and from the cars, degrading the signal quality.
Solution: Antenna diversity & Multipath reception needs to be implemented to maintain the quality of the signal received at the radio and the UE. The Channel State Information (CSI) can also be computed & predicted more frequently to determine the channel condition. ML algorithms can be used to make a more accurate estimation of the channel conditions. This will require a higher compute power at the base station to accurately calculate the CSI and compensate accordingly.
A unique challenge faced due to the high speed at which the car travels is that the time for handover between the neighbouring cells is small, of the order of a few tens of milliseconds to prevent drops or glitches in the telemetry audio and video data that is sent from the car. Hence the seamless handover needs to be implemented for these racing cars.
Solution: Multi-TRxP features like Co-ordinated Multi Point (CoMP) and multi-panel beamforming can be used to cancel out the interference from neighbouring cars. CoMP utilises different techniques to dynamically coordinate the transmission and reception for a particular UE to and from multiple base stations. The UE is connected to more than one base station and transmits/receives from both at the same time. The CoMP algorithm resolves duplicate packets received. At the cell edge, when moving from one cell to another, the handover time is very low, and the packet loss is minimised. This improves the robustness of the transmission, hence the connection is more reliable. This result is that there are no dropped video frames in the transmission from the cars.
The cars typically have several sensors that take measurements like acceleration, speed, temperature, vibration, and tire pressure. These sensors can create electromagnetic interference which can hinder the reception and transmission of the 5G radio signals.
Solution: The antennas need to be mounted strategically on the racing cars so that they are not prone to the interference from the other components in the car. The Radio Frequency circuit needs to be designed with a high level of tolerance from spurious signals.
The racing cars are travelling at a very high speed hence a lower Modulation and Coding Scheme (MCS) is required to be used to provide a robust transmission and reception of signals. Similarly, a robust code rate with more redundant bits is needed for the Forward Error Correction (FEC) to recover from any transmission errors.
Hence the bandwidth required can be quite large, approx. 200MHz to provide the required throughput for the racing cars. This raises the issue of which spectrum to use where a sufficiently large bandwidth is available
Solution: The unlicensed 5Ghz or 6GHz bands are an option where a large bandwidth can potentially be available for use. However, this band is currently also shared for Wi-Fi and hence the reliability and guaranteed availability of this spectrum may be an issue. Another option is to use the custom band like the 10GHz band which is currently used by the military for radar stations. The requirement is to keep the transmit power low so that there is no interference to incumbent users in this band. Custom radios and trackside infrastructure would need to be developed when using these bands to provide the required QoS and reliability for the racing cars.
System Bandwidth Calculations
The following table shows the system bandwidth required for a typical racing circuit.
If using a higher order 64QAM modulation, the system bandwidth required is approx. 100MHz.
One of the key decisions for the solution is the choice of spectrum to be used for the 5G system, so that the system can operate in all the countries while providing consistent KPIs like throughput, latency etc. This section describes the options for the operating frequency of the solution.
The 3.5 GHz mid-band spectrum is crowded, and the spectrum is allocated to the operators in a most countries, with the operators having about 20 to 50 MHz bandwidth. There are a few countries for example in Spain, where the individual operators like Telefonica, Vodafone and Orange have 90 to 100 MHz of bandwidth. In Great Britain the operators have about 20 to 60 MHz of bandwidth each.
In countries like Bahrain, Canada, Russia, Netherlands and Turkey where the 3.5 GHz spectrum has not yet bean auctioned and is planned.
Hence to operate in the 3.5 GHz band, partnerships with two or more operators would be needed to use 100 MHz spectrum.
Unlicensed/Shared Spectrum: A few countries have reserved part of spectrum in the mid-band for unlicensed and shared use (not available universally) — this is shown in the table below. The available spectrum is primarily for private indoor use and should not interfere with incumbents operating in the same frequency band.
Pros of 3.5 GHz band:
- Better transmission characteristics and wider coverage compared to mmWave
- Availability of commercial off the shelf hardware and software
- Mature ecosystem of vendors and partners
- Spectrum in most countries has been licensed, so the ownership is clear
Cons of 3.5 GHz band:
- Crowded and fragmented band
- Required bandwidth is not available with a single operator, hence need to work with multiple operators to get the bandwidth
- Operators have to spend significant amount to get the license, hence may be unwilling to share the spectrum
- Private spectrum is not available in all countries
The 5G NR FR2 frequency bands are given in the table below:
The licensed mmWave spectrum is available in a few different bands: 26 GHz, 28 GHz, 37GHz and 39GHz & 47 GHz. Typically, the amount of frequency allocated to each operator in the mmWave bands is higher — from 200 MHz to 400 MHz or more. This is ideal in terms of a large chunk of bandwidth required for the motorsport solution.
- USA: 37 to 37.6 GHz (3x200MHz) has been allocated by FCC for shared/unlicensed use
- Germany, UK, Australia: 24.25 27.5 GHz for local licenses
- Japan: 28.3 to 29.1 GHz (150 MHz outdoor use) for local license
Pros of mmWave:
- Availability of commercial off the shelf hardware and software
- Mature ecosystem of vendors and partners
- Larger bandwidth (200 to 400MHz) allocated to operators
- Nearly 40+ commercial network in various stages of rollout worldwide. More networks will be rolled in 2022
- Spectrum is cheaper to license per Hz of bandwidth compared to 3.5 GHz band. Price per MHz — POP in mmWave bands ranges from just over $0.001 to $0.01. In the 3.5 GHz band, prices are $0.20 per MHz — POP. Hence operators would be more likely to share the mmWave spectrum bands
- Radio Units with up to 800MHz bandwidth are currently available
Cons of mmWave:
- Spectrum is not yet auctioned in all the countries
- Need to work with operators for access to the licensed spectrum
- Comparatively poor transmission characteristics — higher path loss & attenuation. Hence needs a denser radio network for full track coverage
- Needs line of sight from UE to Radio Unit
- Performance may degrade in rain
Disaggregated RAN architecture consisting of the Radios, Distributed Units and Centralized Unit can address help address emerging connectivity requirements in motor sports. This modular architecture makes the network flexible and easily upgradeable. For instance, going forward radios supporting multiple bands (tri and quad band radios) will be available in the market. In this case only the radios need to be swapped out, while the rest of the network elements remain unchanged. Additionally new features introduced in subsequent releases of 3GPP can be added to the network via software upgrades over the air.
The modular architecture makes it easy to move the 5G network between the race venues with a short turnaround time.
We have also analyzed the 3.5GHz mid-band versus the mmWave band for motorsports. It is recommended to use the 3.5GHz band (or the larger n77 band from 3300 to 4200 MHz) due to the better transmission & propagation characteristics. Within the n77 band, unlicensed sub-bands or sub-licensed bands from operators can be used to build and operate the 5G RAN