Enabling Monitoring and Preventative Maintenance for Railway Stations

A provider of transportation solutions turned to Altoros to develop a high-load monitoring system for railway stations.Develop similar
TRANSPORTATION
CLOUD-NATIVE
JAVA
KUBERNETES
MACHINE LEARNING
PYTHON
TENSORFLOW

About the project

Brief results of the collaboration:

  • The company got an additional revenue channel.
  • The delivered solution allows for monitoring and aggregating petabytes of data per day.
  • The system recognizes potentially dangerous situations and notifies about critical accidents in real time.

The Customer

Being a division of a German-based conglomerate, the company focuses on delivering solutions for rail transportation. The product portfolio of the customer includes commuter, regional, high-speed, and intercity trains, locomotives and metros, as well as automation and power systems.

The Need

The customer wanted to develop a system for monitoring railway crossroads and notifying responsible parties about accidents or any malfunctions with railway equipment. The solution should also be able to recognize potentially dangerous situations and send notifications in real time. When the customer turned to Altoros, it already had a legacy system. However, the application did not scale and was unable to deliver critical messages in time.

The Challenge

Under the project, the team at Altoros had to address the following issues:

  • Railway stations include multiple equipment and vehicles that generate petabytes of data to monitor per day, so the system needed to be scalable enough to sustain high loads.
  • As the monitored assets came from different vendors and used different protocols to aggregate data, there was a need for standardization and unification.
  • Critical notifications about potential or happening accidents needed to be delivered in real time.

The Solution

By delivering a microservices-based architecture, our developers made it possible to scale the system as required to sustain petabytes of data and throughput of megabytes per second daily. To enable standardization across numerous assets, engineers at Altoros created a customized application-level protocol, which supported multiple devices and protocols from different vendors.

By implementing the HiveMQ protocol, our experts ensured the system sent critical notifications in real time. Furthermore, developers at Altoros enabled the system to distinguish between three types of message importance— critical, warning, and for consideration. Our team also created the corresponding policies responsible for aggregating notifications based on their type. By compressing messages, HiveMQ also allowed for cutting on expenses related to the Internet traffic used to send notifications.

To abide by security means, developers at Altoros delivered an Apache Kafka-based policy protocol.

Using TensorFlow, our engineers built a module responsible for analyzing data from video cameras at railway stations to detect potentially dangerous situations and prevent accidents through notifying responsible parties.

The Outcome

Partnering with Altoros, the customer developed a solution that allows railway stations to monitor and aggregate multiple assets through a unified dashboard. The solution also enables preventative maintenance of the stations’ equipment and notifies about potential or critical accidents in real time. With this system, the customer got a revenue channel in addition to the products/services it delivers.

Technology stack

Server Platform
Kubernetes
Programming Language
Java, Python
Technology
Node.js, HiveMQ, Apache Kafka, Apache Beam, Hadoop, Spark,TensorFlow, scikit-learn, GBoost
Database
Couchbase Server, HBase, MongoDB, PostgeSQL

Contact us

Let's see what we can do together

Siarhei Sukhadolski

Artificial Intelligence practice head

Headquarters

830 Stewart Dr., Suite 119Sunnyvale, CA 94085
First Name*
Last Name*
Email*
Phone*
Your company name*
How can we help you?