How Cloud Services Can Help Optimize Public Transportation Systems in Cities

Transportation and traffic management are hot topics when city planners and administrations think about ways to make a city smarter and more livable. Recent statistics tell us that drivers in the U.S. spend an average of 42 hours per year in traffic in cities and lose $1,400 on gas, while idling. In Europe, cities like London and Paris show an average of 74 and 69 hours spent idling per year respectively. Researchers in England found adding an additional 20 minutes of commuting per day has the same negative effect on job satisfaction as receiving a 19% pay cut. These statistics and an increasing desire to be more environment friendly are reasons why city leaders are looking to tackle this problem.

Adaptive Traffic Flow Systems – dawn of real-time analytics, modeling and forecasting

The California Department of Transportation (Caltrans), in partnership with Partners for Advanced Transportation Technology (PATH) at the University of California, Berkeley, developed a project called Connected Corridors that seeks to reduce congestion and improve mobility, travel-time reliability, safety, and system efficiency in California’s most congested corridors. The project includes a Data Hub to consolidate real-time Internet of Things (IoT) sensor datasets; scalable, real-time traffic modeling and forecasting; an incident response Decision Support System informed by data and models; and coordinated, real-time control of signals, ramp meters, and detour signs.

A similar concept is being built by Louisville, Kentucky. By using machine learning, real-time traffic data, IoT infrastructure, and interconnected systems, a next-generation adaptive traffic-flow management system can sense detrimental systemic changes to the circulatory nature of traffic and automatically adjust city infrastructure to mitigate the impact. With such a large amount of data collected and analyzed, they needed a scalable and reliable platform for testing their applications and model. They went all-in on AWS to support data storage, analytics, and scaling for all of their testing environments.

Dev/Test Environments for Smart Cities

Testing and developing new models and applications are key enablers for developing a smart city. Technology is evolving rapidly. If planners focus on pre-packaged solutions that fail to adapt to changes in user requirements, it may hinder or slow down city administrators in meeting the objective of transforming their community into a smart city. The concept of ‘smart’ itself is relative in this context –  a city needs to be smarter than it was before.One way administrators can overcome this challenge is by adopting solutions and technologies that can support evolving demands.

Going back to the traffic management example, a common solution is to deploy traffic sensors on the ground to sense the traffic flow. After you collect enough data and understand traffic flow patterns, you may want to deploy additional functionality or machine learning models directly onto the sensor and extend its capabilities to meet real-time response needs.

AWS Greengrass lets you run local compute, messaging, data caching, sync, and machine learning (ML) inference capabilities for connected devices in a secure way. With AWS Greengrass, connected devices can run AWS Lambda functions, keep device data in sync, and communicate with other devices securely – even when not connected to the Internet. If the connected IoT device/sensor you deployed to monitor traffic is AWS Greengrass compatible, then you can easily extend its capability over time with new Lambda functions or with new ML inference models.

With fast emerging concepts such as connected cars, cities have exciting new opportunities to address traffic challenges. A recent study suggests that a single connected car will send around 25GB of data per hour from its sensors. While most of this data may be sent to private users or vehicle manufacturers, in some contexts and scenarios, cities can derive benefits by leveraging this data, especially when the city owns and/or is responsible for a vehicle fleet.

For example, connected police cars or ambulances may send or receive data as they patrol the city and this data can be used to understand vehicle performance, maintenance needs, and operational readiness. Central Command Center and operational teams may get real-time insights and situational awareness based on data transmitted directly from the vehicle. This data can, in turn, be used to adjust operation of traffic lights in real-time, to reduce the time it takes for an ambulance or police car to reach their destination.

Storage services like Amazon S3 give customers the ability to store virtually an unlimited amount of data cost-effectively, while services like Amazon SageMaker make it easy to build, train, and deploy machine learning models at scale.

Using cloud services, administrators have the opportunity to deliver on the promise of making our cities smarter and more responsive, fulfilling the smart city promise: making lives of citizens more enjoyable and safer. To learn more about how other cities around the world are becoming smarter, increasingly connected, and more sustainable, visit the AWS Smart Cities site.

Post authored by Giulio Soro, Senior Solutions Architect, AWS