5G and Edge AI: Tackling Traffic Management
Guest WriterGuest Writer
The way we commute may have transformed over time, but the way traffic is managed has not changed. The INRIX Global Traffic Scorecard reports that the world’s 20 most congested cities lost between 164 and 210 hours in congestion per capita through 2018. Exponential rise of vehicles in urban cities is the core reason behind congestion. Better public transport is the solution, but along with this, we need to also look at how improving the efficiency of the traffic management can better the scene. Traffic authorities have tried initiatives to transform reactive management into proactive traffic management but have been constrained by network speeds and processing capabilities at the edge. 5G along with AI provides a huge opportunity here for traffic management.
II #IoT #IoTForAll" quote="'With 5G, edge devices will become even more powerful in transmitting and processing larger amounts of data by AI analytics servers which can only work to benefit traffic management.' - Vinod Bijlani" theme="]
With denser and more complex road networks, newer technologies, and larger data, 5G will provide greater visibility and control over traffic. This will, in turn, help clear transit networks faster, minimize blockages further, eliminate cascading effects, and make the roads safer for all users. With 5G, edge devices will become even more powerful in transmitting and processing larger amounts of data by AI analytics servers which can only work to benefit traffic management.
Pegged at being 70 times faster than 4G, it will give combined visibility into the movement of all road users – people as well as traffic – enabling better overall planning. With multitudes of sensors, cameras, and even drones, 5G will convert road networks into a fleet of mini-clouds, each communicating with the other, including self-driven vehicles. The huge amount of data generated by sensors in self-driven or autonomous vehicles can effortlessly be accommodated by 5G to allow for inter-vehicle and inter-sensor communication.
Crucial information will be picked by the sensors in these vehicles to make decisions and change course, based on recorded observations. Martti, the autonomous vehicle from VTT Technical Research Centre in Finland, has been tested for detecting icy road conditions ahead of time as well as for inter-vehicle transmission of 3D views.
The power of Artificial Intelligence (AI) and Big Data combined with the supremacy of 5G technology will provide a robust solution coupling high reliability and ubiquitous network access. Low latency offered by 5G is key here, with AI models using real-time network information and historical data to detect the possibility of incidents and instantaneously devising optimized response plans to be delivered at high speeds. Traffic metadata from the entire road network can be captured in real-time using a combination of traditional and edge-based AI systems. This combination of 5g and AI will hold the answer to transforming traffic management in the coming decade. It could also signal the much-needed boost for autonomous vehicles within a collaboratively connected system. Let's look at two specific AI-based solutions and their effect on vehicle activity.
AI-based traffic light controls will have a high impact on vehicle activity, significantly reducing conflict in vehicular movement and increasing road network capacity. The integrated set-up for effectual traffic management would involve a self-adaptive traffic light system, an edge system, and a backend monitoring system. The video captured using IP cameras is relayed to the edge-based AI system that analyses the data before sending it for backend monitoring. Pre-trained deep learning models send back the processed information to the self-adaptive traffic lights in real-time to create traffic fluency.
With the adaption of traffic lights to changing traffic in real-time, on-road movement can be controlled by traffic light timing that adjusts itself by the second. The changing traffic scenario and the timing at intersections can be shared through interoperable communication so that all intersections are prepared to optimize the flow of approaching traffic. A pilot system deployed at Pittsburgh, Pennsylvania, has reportedly reduced travel time by 26 percent, idling time by 41 percent, and emissions by 21 percent. Interestingly, the adaptive traffic light system also reduced total and fatal incidents by 13-36 percent.
With incidents being unanticipated and sometimes catastrophic, incorporating AI into building an integrative sustainable traffic incident management system with intelligent traffic lights can transform traffic monitoring. This is where the confederation of blended technologies comes in. Big data from the IP cameras, GPS, mobile phone tracking, probe vehicles, and loop detectors is merged to arrive at more precise inferences than if the enormous information was studied independently. AI algorithms then continually and instantaneously analyze the data thus fused to detect potential incidents.
Traffic simulators can study both archived and real-time data at the time and location of the incident to analyze the impact. AI models predicting incident duration can also indicate specific points requiring attention as well as the overall effect on the road sub-networks. Further, deep learning models can explore the correlation between the intensity and the overall impact, helping prioritize the incident and its response. The integration of data analytics helps in testing various traffic scenarios from which an effective, real-time, automated traffic incident response plan can be drawn.
In Delhi, sensors from over 7,500 CCTV cameras, programmed traffic lights, and one thousand LED signs collect real-time data that AI processes into instantaneous insights, which authorities use to improve traffic management. Data collected from city-wide intelligent cameras installed in Milton Keynes in England is run on a deep learning model to predict traffic conditions 15 minutes ahead with 89 percent accuracy.
To match the promise of 5G, road and transport network management systems also need to evolve over time. There are bound to be more complexities with data from vastly varied sources. The process of all systems working together to be pervasive and instantly reactive all at once will require precision implementation. Amid tech adaptability, it is important that the smart network decisions be autonomous, as well as understandable. This will provide scope for human decisions and interventions alongside technology when the need arises. While we may have turned a century since the world’s first highway was constructed, it’s only now that the world is revving up its engine for the drive.
New Podcast Episode
Recent Articles