Skip to main content

Pioneering AVs in public transportation using AWS with Swiss Transit Lab

Swiss Transit Lab and UMB use AWS infrastructure to securely deploy autonomous vehicles for future public transportation.

Benefits

increase in operational efficiency

TB of data uploaded daily for each vehicle

lower monthly costs

Overview

Swiss Transit Lab (STL), a center of expertise for intelligent mobility in Switzerland, is pioneering autonomous vehicles (AVs) for future public transportation. Because safety and security are of the utmost importance in the project, the organization is using Amazon Web Services (AWS) for the core infrastructure. Collaborating with UMB, an AWS Partner, STL created a scalable, compliant cloud environment on AWS to run autonomous-driving software and remotely monitor the vehicles. The organization successfully launched the AV training phase in Furttal, Switzerland, and will open the project for public use in the coming months.

Missing alt text value

About Swiss Transit Lab

Nonprofit organization and competence network Swiss Transit Lab develops and tests intelligent mobility solutions under real-world conditions to enhance transportation efficiency, safety, and sustainability in Switzerland.

Opportunity | Using AWS to secure cloud infrastructure for STL

Specializing in intelligent transportation systems, STL brought together public sector transportation stakeholders to launch the Intelligent Automated Mobility (iamo) project. “We are testing the mobility of the future in the Furttal area by deploying self-driving vehicles on public roads,” says Matthias Rödter, president of STL. “We aim to figure out how these vehicles can further develop public transportation.”

The organization chose AWS to meet the project’s security requirements (under the AWS Shared Responsibility Model). “We needed a large provider with a strong reputation to host data in Switzerland,” says Christine Mauelshagen, iamo project lead at STL. STL also relied on UMB to implement AWS services and monitor the cloud environment throughout the project.

Solution | Implementing data protection and security controls on AWS

Together with UMB and other collaborators, STL established a cloud foundation on AWS, adopting governance and security practices to meet regional and national requirements for data protection. To validate the system’s architecture and security posture, the organization worked alongside the UMB and AWS teams in addition to cybersecurity specialists.

On AWS, STL deployed AVs that were equipped with software from WeRide, which provides AI-powered autonomous-driving solutions. To run the application backend, the organization uses Amazon Elastic Kubernetes Service (Amazon EKS), a service for building, running, and scaling production-ready Kubernetes applications. By moving from self-managed clusters to Amazon EKS, STL reduced operational burden and strengthened security by keeping clusters patched and up to date.

For secure, near real-time ingestion of vehicle telemetry data, STL uses AWS IoT Core, a service for easily and securely connecting devices to the cloud. Additionally, the system live streams data over secure transfer routes from vehicles to local teleoperation stations in Switzerland so that human operators can intervene remotely if something goes wrong.

To prepare for autonomous operations, the organization mapped the driving area and trained algorithms and AI models by using on-demand GPU instances in the AWS Europe (Zurich) Region. This achieved cost-effective compute for intensive workloads.

Using Amazon Simple Storage Service (Amazon S3), an object storage service, STL uploads the data that the iamo vehicles collect while out on the road. To protect sensitive data, UMB implemented security policies to lock down the Amazon S3 buckets for review. The partner also created a web application to help STL verify that on-vehicle data anonymization was successful before using the data in WeRide’s software.

The system exports the anonymized data to Amazon OpenSearch Service, an AWS managed service for running and scaling OpenSearch clusters. This way, STL can use telemetry data for operational analytics. To monitor data access, STL uses AWS CloudTrail—a secure, standardized API logging service. “Security and trust are so relevant for an autonomous-driving project, which is why we chose AWS,” says Rödter.

Outcome | Improving public transportation coverage with AVs

Using AWS, STL can now upload about 1 TB of data per vehicle per day and has reduced monthly costs by 40 percent. UMB estimates that it achieved a 40–50 percent increase in operational efficiency by using AWS rather than an on-premises environment. “We implemented the web application in a few days, and now we have good visibility to monitor and act on specific events,” says Roberto Aliano, product manager for cloud services at UMB.

The project is currently in the testing phase, operating three vehicles with safety drivers aboard. The next step will be to seek approval for fully autonomous deployment to cover a 110 km network in Furttal. STL also plans to add autonomous shuttles to the fleet.

The collaboration among multiple stakeholders was a highlight of this complex project. “There’s no blueprint for this kind of project,” says Mauelshagen. “We need to develop everything from scratch, and we’re excited to work alongside AWS, UMB, and WeRide. We have a strong drive to succeed and the mindset to break new ground.”

Missing alt text value
Security and trust are so relevant for an autonomous-driving project, which is why we chose AWS.

Matthias Rödter

President, Swiss Transit Lab

Did you find what you were looking for today?

Let us know so we can improve the quality of the content on our pages