As well as providing a fascinating insight into the journey the project has undertaken over the past two years, there were a number of actionable insights and learnings that other local and transport authorities - and indeed anyone considering deploying Open RAN networks - can benefit from.
Introducing the Smart Junctions Project
The project focused on three key areas:
- Smart city use case - demonstrating the benefits to real time traffic control and cost that a 5G connection can bring
- Vendor diversification - exploring the ability to deliver a 5G private Open RAN network that seamlessly integrates between assets whilst prioritising security
- Business model: demonstrating long term commercial benefits using a Network as a Service (NaaS) commercial model with a local authority owned network. Specifically they were interested in the ability to both save costs and potentially rent the network out for further use cases
Alexander Yeomans, Vivacity, introduced the smart city use case and set out the project’s intention to replace existing tech - much of which hasn’t changed since the 1990s and is no longer well suited to today’s needs - with more affordable, agile solutions that can revolutionise a city’s traffic signal control. With automation and AI, Vivacity aim to not just measure congestion but enable optimisation of traffic flow, promoting different and more sustainable travel modes, enacting new citywide policies easily, and measuring the subsequent impact on surrounding air quality. Put simply, their work is enabling dynamic and adaptive control to a road junction throughout the day, with process automation helping to reduce the costs and time to recalibrate.
They deployed nine smart junctions (monitoring not just traffic, but also pedestrian and cycling zones) along the A6, the main corridor in and out of the city centre. An extremely congested area, it also encompasses multiple train stations and Salford University. In total 57 sensors were deployed as part of the project and 49 external 5G routers were retrofitted.
A user interface then enabled Transport for Greater Manchester (TfGM) to analyse what was happening across their network in real time, providing them with live occupancy data and also validating any actions being taken. Using data from the sensors and traffic signals then enables TfGM to make decisions around controlling the traffic lights i.e. extending a green signal, using an AI algorithm.
Real-time deployments require a low latency, reliable connection and Yeomans explained that wired connectivity is too expensive and difficult to scale, while public 4G networks have proven to be unreliable and deliver too high a latency in particularly around highly congested areas. The very points where there are peaks on congestion - such as rush hour - and more dynamic traffic management is needed, are also the points where capacity on public 4G networks becomes limited
Maria Lema, Weaver Labs, then talked to the Open RAN element of the project, outlining the edge-based private 5G network that had been built on street side cabinets and traffic lights. Lema outlined that despite Open RAN being challenging to deploy, they felt it was the right solution to match the customer needs:
- A cost effective solution to enable subsequent scalability
- A small cell network no larger than the size of a shoebox and less than 10kg that could be hung on a lamp post
- A strong environment to test zero trust architectures around the supply chain
She also outlined Weaver Labs’ own stance and belief in open networks and opening up the supply chain.
Sam Li, TfGM, provided insight into the broader strategic backdrop to the project, explaining that their interest was in understanding what exploring these three areas would mean for the citizens of Greater Manchester and how this could help TfGM to do their job better.
At their core, they are interested as a transport authority in thinking beyond individual transport modes and encouraging behaviour change as a whole to drive sustainable and economic growth, with improved quality of life at its core. “A cycling journey is not only better for the environment”, Li stated, “but could improve that individual’s personal wellbeing”. In line with other project events, there was a strong sense that 5G trials and deployments do need to be considered in the round; it’s less about the technology or in this case even the transport mode, and more about how the connectivity - and the technologies it enables - can be harnessed to overcome broader challenges and support strategic goals.
Replicability was also a key driver for all the partners; TfGM are interested in how they “transition from R&D to Business As Usual”, with Li clearly articulating that the connectivity corridor they had established was a blueprint for the rest of Greater Manchester. Similarly, Vivacity and Weaver Labs are keen to take their learnings from the project and explore how they can be scaled and deployed across other areas and with other customers.
Li also touched on the importance of Open RAN networks to TfGM in that scaling discussion, explaining that “we can no longer procure a system that lasts 10 years, so we needed to understand how we can build a more modular network that can be easily upgraded and doesn’t limit ourselves in terms of opportunities”.
Despite significant challenges - primarily concerning access to backhaul, space in the cabinets and the availability of equipment - the project was able to successfully deploy a close-based network running on commercial off-the-shelf equipment and using software tools that cut operational costs. Significantly, this has created a platform for innovation that is now available to other smart city applications, helping to facilitate the creation of stackable use cases.
The conclusion of the project is that at this time, Open RAN is not ready to create seamless automated use case integration, however it was widely acknowledged that the market is maturing rapidly. When Ed King, one of the members of the Technical Design Authority in DCMS’ Future Networks Programme, was asked what the project could have done differently he drily observed “start the project two years later”.
Despite the challenges encountered therefore, the potential is clear. Critically, they were able to demonstrate a 75% cost saving over a single vendor approach. Comparing costs to wired connectivity, they identified that the cost of delivering one smart junction on ethernet is c. £6.5k, whereas it would be £6.2k to deliver eight smart junctions connected to one small cell on 5G (feasible in an urban environment).
Yeomans explained that Vivacity has deployed other smart junction itrials in Peterborough and Cambridge with wireless 4G and that they experienced high latency variability, which was especially problematic in peak periods. They have identified that an acceptable latency for real-time junction optimisation is 400 ms. For Vivacity therefore, 5G is the ideal solution as it promises reliability and replicability at scale, no latency peaks during congested periods and cost benefits particularly at scale when connecting a corridor or region.
He acknowledged that their 5G network had struggled with stability issues due to frequency interference, resulting in down time on sensors, and this was something they wanted to further improve. It was recognised that difficulty in sourcing kit for an Open RAN network had reduced the amount of time for lab testing and live trials; such is the nature of innovation projects. However, results observed when the 5G network was stable and the radio was fully under control has provided some key indications as to the value of a 5G Smart Junction versus a 4G connection during a peak period (4-7pm).
In this scenario they recorded the 5G network delivered an acceptable latency 99.4% of the time (compared to 94.5% on 4G) and the jitter (the variation in latency) was only 83 ms compared to 347 ms in 4G. Impressive stats but what does this mean in the real world?
Using webtag analysis the project has calculated this equated to a 1% journey time improvement over the junctions they controlled. With 16m journeys across those junctions happening each year with an average journey time across the junction time 100 seconds, this equates to a 1 second improvement per journey and for one urban four-arm junctions, this delivers a £45k yearly total road user cost saving.
There was open and honest discussion throughout the event about both the challenges and successes of the project, providing significant learnings and insights to the partners, with Li stating that “the Smart Junction project has enabled us to become a smarter customer”. Critically, many of these learnings can also benefit others and we have summarised the key insights below:
- There is no one killer use case: echoing the sentiment of many other projects, the importance of stacking use cases was highlighted, with Greater Manchester Combined Authority considering how they can scale the connected corridor deployed to use not just for smart junctions, but a range of other use cases from CCTV to remote apps for NHS Trusts.
- Communication and collaboration: all partners clearly articulated this as a vital component of their success, especially for those operating at this early stage of market maturity. All speakers recognised that innovation is complex and unpredictable, but that open communication and transparency plays a critical role in unblocking barriers.
- Skills of contractors: TfGM observed that their typical contractors are used to installing traffic signals and had not worked with 5G radios before. This also led to lots of subcontracting of services which made it difficult for them to identify who they needed to contact when faults arose. It was recommended therefore to carefully select one contractor from the start who has the necessary skill set.
- Identifying Assets: Lema flagged that it took them nine months to gather all the relevant data around the assets that could be used for the project due to there being no single repository for this data, no up-to-date diagrams and multiple layers of data ownership. It was identified that the Digital Connectivity Infrastructure Accelerator project will play a key role in plugging this gap. More details on DCIA can be found here.
- Data privacy: there was an interesting debate around privacy especially when it comes to unfiltered video feeds in control rooms. It was suggested by Transport for West Midlands, who sat on one of the panels, that transport and local authorities may think about blurring identifying features of individuals on feeds to protect an individual’s right to privacy. There was also much discussion about the fact that it’s not enough to have the best intentions and safeguards, you also need to ensure the public understand and trust the measures you put in place.
- Cybersecurity: the project focused on building a zero trust network but acknowledged this was more difficult than expected as securing an organisation needs to go beyond the product or service offering. The project developed an integrated cyber framework and ranking tool which was shown for the first time at the showcase event. This aims to map an organisation's current position, their desired end point and opportunities to work to improve their cyber score.
- Physical security: it was identified that with all the focus on cybersecurity, it’s important to not lose sight of your physical security policies and for instance the amount of people who have access to roadside cabinets.The value of
- Systems Integrators: the project encountered significant challenges around integrating an Open RAN network. It was recognised therefore that a traditional customer / supply relationship simply doesn’t work based on where the equipment is currently. Instead, a more collaborative approach is needed and Lema highlighted the increased role of the SI in this new environment “for both installing kit and getting the software done”
- Maturity and availability of equipment: this was undeniably a challenge for the project but there was a sense of optimism across the day that the market has matured significantly over the past two years. Lema summarised: “We’re still at an early stage but it’s this next generation of kit that draws less power, weighs less and occupies less space which is where this now starts to make sense”. Others looking to deploy 5G connected smart junctions should be cognisant of the fact this is a growing and maturing market and be ready to encounter delays and complexities inevitably associated with that.
- The Value of Bench Testing: Yeomans advised attendees to not underestimate the importance of this stage, especially with a multi vendor approach, emphasising the need “to stretch and test in a controlled environment” and create documentation that can be referred to in the field.
TfGM were optimistic about both the results from the project and what it would lead to, stating that through their participation in the project they could now see the roadmap for the development of kit and the deployment for scale up. Hannah Tune, TfGM stated that “the project has opened peoples’ eyes to smart junctions, applications and how we can adapt signal control multi-modally based on real-time data”.
One of the final thoughts of the day came from John Steward, Greater Manchester Combined Authority: “It feels like we've reached a tipping point in the maturity of tech – we now need to think about how we build applications on top of that to deliver the return on investment that we need”.