Real-Time Vehicle Diagnostics and Health Monitoring

Background

With the evolution and expansion of Internet of Things(IoT) since the early 2000s, there was an inevitable  need for the auto manufacturers to put the vehicle diagnostics and health assessment data in the hands of the owners and drivers of their vehicles. The industry’s initial focus was predominantly on sourcing and sharing of the vehicle status data, more as nice to have reminders for the customers to get the vehicles to their dealers and service centers for repairs, scheduled maintenance, or service campaigns/recalls.

Since, the market demand for vehicle health related information has not only grown at a great speed, but also changed significantly.  Today’s customers demand a frictionless and ubiquitously “connected” experience with their vehicles, almost as an extension of themselves in mobility. From their mobile or at-home devices and in-vehicle technologies, to their preferred the dealership/service center, the owners and drivers want to be remain aware and gain better control over managing risks related to their safety and transport, as well as, their overall driving experience.

With the changing customer needs and wants, the automobile companies had to play catch up to a certain extent, but also build an actionable plan for the future of their diagnostic and health monitoring data, technologies, and capabilities.

 

Business Need

Dimiour got the opportunity to work with one of the top global automobile manufacturing companies to assess their legacy system for a major uplift in these areas:

  • Scalability – scale to growing types and number of vehicles out in the market, across their global regions
  • Timeliness and Speed of Information Delivery – provide meaningful and actionable information and communication when it is relevant and needed; in real-time and/or near real-time depending on the severity and impact of the issues being diagnosed
  • Availability and Reliability – “always-on” monitoring and services design with customer’s safety in mind
  • Cost Management – reduce and manage on-going hosting and operational costs for the expanding set of services and systems

Solution

With an extensive experience on the cloud and scalable systems, Dimiour experts put together a data lake based architecture collecting information from varying sources, in real-time. The sources include the IoT streams from numerous sensors which are actively channeled through an event driven streaming analytics platform. The aggregated data then help the system gain learnings and insights during its data ingestion and organization processes. The data outputs from the lake are leveraged in a number of different business scenarios:

  • reporting, modeling, and analysis needs for business planning and decisions
  • detection of patterns and anomalies for timely alerts and notifications to the impacted parties, including the customer, manufacturer, and customer’s preferred dealers/services centers
  • holistic vehicle health checks run against maintenance schedule and thresholds/warranties for the vehicle’s overall upkeep and preventive care measures

 

The implemented solution can be broken down into following major stages of data processing.

 

Collect

Transactional datasets get pumped into the streams from various IoT sensors, making up majority of the data volume. To complete the master dataset, additional data to be associated with the vehicle, such as parts, owners, drivers, dealers/service centers, are pulled from various other sources on a timely manner.

 

Organize

To optimize accessibility and correlations, the data is cleansed, standardized, enriched and organized into multiple logical and reasonable facets within the lake.

 

Analyze

A set of rules, conditions, and thresholds are defined to continuously monitor some data segments, while periodically analyzing other datasets.

 

Infuse

The results of the analysis are then used to trigger automated updates, processes and notifications.

 

Following considerations helped to shape the technical architecture of the data processing platform:

 

  • A pure serverless approach used to setup the data analytics pipelines with AWS Kinesis streams
  • Analytics workflow using Apache Flink to write to AWS DynamoDB Accelerator (DAX) for the low latency transactions
  • Data storage and archiving leveraging S3 and S3 Glacier, respectively
  • Step functions, lambda functions, and SQS queues implemented to support complex batch processing

 

API management using AWS API Gateway with serverless compute components to handle the service layer

Benefits

The solution implemented by Dimiour resulted in a number of benefits for our client, including:

    • The launch and operational support for multiple countries and continents to provide health and diagnostics based services to customers around the globe
    • Real-time vehicle health analysis and diagnostics alerts to customers globally
    • Notable cost savings in hosting, operations, and support
    • Significantly reduced instances of customer service requests and incidents
    • Increased availability and resilience