With the growth of technologies such as artificial intelligence (AI), it may be tempting to believe the human-staffed traffic management control room to be well on its way to obsolescence. But while it’s certainly true that technology is driving some profound changes in the way that control rooms and their supporting infrastructure are built and operated, it’s equally true to say – and with a high degree of certainty – that the most powerful piece of control room technology is the alert and observant human operator.
Traffic management, and in particular the management of traffic incidents, requires effective real-time intelligence gathering from multiple sources and the balancing of many, often conflicting, priorities. One day, there may be computers capable of managing such feats but that day is still a long way off, placing the emphasis of effective contemporary control room design still very firmly in the human realm.
Visual acuity is one of humankind’s greatest assets, with the largest parts of our brains devoted to the processing and derivation of meaning from the visual signals received through our eyes. It is not surprising, then, that control rooms tend to be designed around the easy availability of visual information.
For many decades, the main direction of technological advance in control room design has been towards enhancing the ability of human operators to ‘see’ situations on the ground in ever-greater detail. As in computers, it’s often necessary to split the task of processing across many agents, which in this case are human operators. This parallel processing requires all operators to be able to access the same intelligence at the same time, hence the central role of the big screen video wall that is still the most effective method of achieving this synchronicity.
Control room display technology is therefore crucial and advancing in two main areas: the clarity and resolution of the display hardware itself and the ability to effectively manage ever-greater volumes of data so human operators are empowered – not over-powered – by it.
It is not now uncommon for control rooms to be at the center of a network on many thousands of cameras and other devices such as flow monitors. Managing these manually would be an almost impossible task. Consequently, systems are increasingly employing camera networks based around a mixture of conventional and machine vision technologies, using AI to recognize potential problems such as slowing traffic, vehicles moving in unexpected ways or the suspicious behavior of individuals.
If a situation is detected, not only does the system alert the operators, but the video wall screen controller automatically changes the video wall format to focus on the relevant camera feeds and any related data sources. In some applications, AI-assisted visualization has enabled dispatchers to respond on average 10 times faster to incidents.
Increasingly complex camera and sensor networks present another problem – that of system scalability. How to ensure the visualisation system can cope not only with today’s challenges but tomorrow’s as well? The traditional approach of a hardware-based screen controller and downstream switchers has become increasingly outpaced by the requirement for open-ended versatility. Control room installations have a typical life expectancy of 15-20 years, and it’s certain that within that timeframe the pace of technological development will have rendered components of that system obsolete or inadequate for the task.
To mitigate that risk, a new generation of screen content systems is becoming more prevalent. Instead of a single hardware-based controller, the system borrows techniques from the world of cloud computing to create an amorphous virtual processor that can change and adapt to the demands placed upon it.
Mitsubishi Electric’s S-SF suite is one such example. The suite consists of five software platforms, designed to run on PCs or dedicated microcomputers called NUCs. The individual components of the S-SF network work together to share and deploy their collective processing power as needed. Should greater processing be required – for example if a detailed vector graphic needs to be scaled-up to fill the video wall screen – extra processing power is dynamically assigned. With the number of inputs not now defined by hardware, it has become possible to accommodate the addition of IP data sources. Processing power can be increased as required by the addition of more NUCs to the network.
The final link in the interface between network and operator is the display technology itself. For many years now, digital light processing (DLP) technology has been the dominant force in large-scale installations, with smaller or less demanding applications delivered using tiled LCD displays. DLP has proved itself to be a robust and reliable technology, and with the advent of LED-powered projection, one with a long operational lifespan. Some Mitsubishi Electric DLP projectors, for example, are now rated to run continuously for 130,000 hours, which is around 15 years of 24/7, maintenance-free operation. However, DLP has its limitations, particularly in applications requiring very large screens where it becomes increasingly expensive and difficult to achieve acceptable brightness levels.
Direct-view LED screens are becoming an increasingly popular alternative in these applications. These employ a similar technology to the large outdoor screens seen in sports stadiums, but with vastly miniaturized pixel pitches to achieve the required resolution and viewing distances required for indoor use. As an emissive display technology, LED is able to achieve a constant level of light output no matter how large the screen. Mitsubishi Electric recently installed a very large 1.5mm pixel pitch direct-view LED screen in the new Metropolitan Police headquarters in Tokyo.
Direct-view LED screen installations such as these are still comparatively rare, however. Cost is one factor, and another is the relative infancy of the technology. There are only a handful of professional-grade direct-view LED screens available. Even Mitsubishi Electric, which has over 30 years experience in outdoor LED displays, took many years to overcome the technical challenges of ensuring the screen stability and longevity required by 24/7 applications such as control rooms.
A new twist
LCD is another technology being repurposed for the demands of 21st century control room applications. LCD has for many years played a supporting role to DLP technology, being reserved mainly for smaller rooms or breakout suites due to its limited ability to operate round-the-clock in some applications and the obtrusive mullions between the individual tiles in an LCD video wall that could prove distracting for operators.
While the limitations on 24/7 operation are still present, a new generation of LCD displays with extremely thin bezels is providing an attractive alternative to smaller DLP systems. With the gap between screens greatly reduced – down to 0.8mm in the case of the Mitsubishi 4Diamond Grid Razor display – it’s possible to achieve an almost seamless image in a very much smaller footprint and at a greatly reduced cost.
LCD displays are now readily available in larger formats, and are increasingly being used not for display purposes but as mini video walls or workstations. These mirror the content of the main video wall to allow operators to manipulate content directly using touchscreen technology. In conclusion, while the topic of AI is certainly engaging a lot of interest in the many areas of technological development, within the world of the control room, the focus remains very much on supporting the role of the human operator. For now at least, organic rather than artificial intelligence remains the core technology behind traffic management networks.