CM-S03 Vehicle data management and insights
Connected vehicles have hundreds of controllers and sensors producing thousands of individual data elements for operating, and conveying the state of a vehicle. Vehicle data management helps vehicle manufacturers to harness data as an asset, to drive sustained innovation and create actionable insights and improve their customer experience. Vehicle manufacturers are seeking cost-effective ways to simplify the process of collecting data from vehicles that are connected to the cloud help power insights and improve vehicle performance while maintaining the highest levels of confidentiality and security. There are regulations (such as CCPA and GDPR) related to data privacy and data sharing that require customers to provide data lineage and data governance capabilities.
User stories
CM-S03-UC01 Telemetry: Automakers collect a variety and large volume of data that is configurable based on rules and filters to support business needs, such as predictive maintenance, location-based services, and insurance claims. In the next generation of vehicles, with expansive amounts of data continuously ingested at high rates from the vehicle, there are multiple verticals that would want to access the vehicle's data, including insurance, municipalities, and content providers. With automakers monetizing the vehicle data, managing the balance between cost, customer experience, data privacy, and revenue realized important.
When determining the backend architecture for the vehicle's payload, there are a few key concepts to keep in mind, namely size and frequency. Automakers could operate a hybrid approach to ingestion, with support for both high-frequency as well as low-frequency data ingestion. This hybrid approach allows automakers to optimize their cost while maximizing the data monetization potential for the most valuable data attributes from the vehicle.
The telemetry collection system should be able to withstand a sudden connection loss and maintain data integrity. The majority of telemetry data is time series data. To maximize the value of this data, processing and visualization of this data is a key capability needed to help create valuable insights and build new solutions to support those insights. Time series data differs from traditional vehicle sensor data in that it's used to perform queries in time windows across differing time frames.
CM-S03-UC02 Data management and governance: Vehicle manufacturers and vehicle owners need to establish data governance standards for collection, storing and dissemination of data and adhere to local rules and regulations (such as CCPA and GDPR). Vehicle owners and manufacturers need a consent management system to control, share, and delete data that supports meeting local, national, and international data privacy regulations and protects privacy of consumers.
CM-S03-UC03 Data sharing and neutral servers: Neutral servers are servers operated and financed by operators not connected with vehicle manufacturers. Such servers can be used by vehicle manufacturers to make vehicle data readily discoverable and accessible to interested vendors with necessary security and privacy controls.
CM-S03-UC04 Insights and generating value from connected vehicle data: Vehicle manufacturers would like to generate value for their customers from the connected vehicle data acquired throughout the lifetime of the vehicle, for example:
-
Provide Usage Based Insurance (UBI) dependent upon driver behavior, distance driven, and so on.
-
Provide subscription-based value-added service, such as Feature on Demand through over-the-air (OTA) update.
-
Increasing advertising reach and targeted advertising.
-
Independent service stations would like to access vehicle data and fault codes to repair and maintain vehicles.
-
Smart maintenance scheduling to reduce maintenance cost, increase fleet availability, and deliver a differentiating customer service.
-
Drive vehicle optimization, efficiency, and early detection of deviations from normal operating condition using ML algorithms and advanced simulations on a digital twin of the vehicle in the cloud.
Reference architecture
Vehicle data management and insights reference architecture
Figure 4: CM-03: Vehicle data management and insights reference architecture
-
The connected vehicle, with a unique identity principal (X.509 certificate), has hundreds of sensors to collect data. AWS IoT FleetWise Edge Agent collects, stores, and organizes data from vehicle. Based on the campaign defined in AWS IoT FleetWise, the agent decodes signals from the vehicle and sends data payloads through AWS IoT Core.
-
The vehicle can communicate using OEM chosen protocols, such as MQTT and HTTPs, and use AWS services, such as AWS IoT Core and Amazon API Gateway. Vehicles can send video streams to the cloud using Amazon Kinesis Video Streams.
-
Improve data relevance by creating time- and event-based data collection campaigns that send the exact data you need to Amazon Timestream or Amazon S3.
-
By using purpose-built data processing components, vehicle manufacturers can generate data ready for consumption, organized by subject areas, segments, and profiles.
-
AWS Lake Formation makes it easier to centrally govern, secure, and globally share data. A data lake can be used to store and analyze multiple data types from wide variety of sources. Use AWS Glue Data Quality to measure and monitor the data quality and take corrective actions.
-
Enable different user personas from fleet aggregators to data scientists, analysts, and vehicle owners.
-
You can use Amazon SageMaker AI improve ADAS/AV models to optimize vehicle design for performance and efficiency. Insights from structured and semistructured data can be gathered by using Amazon Redshift. Utilizing Quick and other analytics platforms to continually improve vehicle quality, safety, and autonomy using near real time data from AWS IoT FleetWise.
