This page provides an overview of SDV Telemetry and addresses key aspects such as architecture, functionality, and integration.
SDV Telemetry provides a secure, efficient, and scalable way for OEMs to collect, organize, and process vehicle data. You can use the data in these ways:
- Vehicle health monitoring: Track system performance, identify potential issues, and enable proactive maintenance.
- Driving-based insurance: Provide data-driven insights for insurance providers.
- Crash data collection: Gather data during accidents to improve vehicle safety.
- Feature development: Understand usage patterns to guide feature development.
- Performance optimization: Identify bottlenecks and optimize vehicle performance.
- Fleet management: Collect vehicle information to manage and optimize the efficiency of commercial vehicle fleets.
Key features
SDV Telemetry provides the following key features:
Scalability
Multiple telemetry instances can be run on individual SDV (or IVI) instances, allowing for data collection from different zones within the vehicle. This flexible design accommodates different vehicle architectures where communication is restricted across vehicle zones and enables independent data collection from different ECUs.
Security
The system is implemented in Rust to prevent memory exploits and uses Android's built-in security features, such as SELinux and process isolation. Structured input using protobufs is validated before processing to avoid errors.
Updatability
All components of SDV Telemetry can be updated independently, ensuring that the system can be maintained and improved. Data collection is fully configurable using telemetry campaigns.
Configurability
Telemetry campaigns are defined by metrics configurations, which can be created and updated in the cloud. These configurations define what data to collect, how to process it, and when to report it.
Edge processing
To reduce data transfer to the cloud, the system includes an edge processing engine that processes data locally in the vehicle and sends only the relevant information to the cloud.
Flexibility
The system is designed to work on both SDV and IVI instances. Core data collection components are reusable, and target-specific components provide access to target-specific data sources.
Testability
A simulation framework lets you validate the metrics configurations before deploying them to the vehicle.
Telemetry SDK
Use the SDK to interact with SDV Telemetry. Doing so facilitates the integration for telemetry clients and telemetry data sources that use the Configurable Publisher Registry. The SDK is provided for Rust. An experimental version is provided for Java.
Components
This diagram shows the key telemetry components:
Telemetry service
The Telemetry service is an onboard SDV agent responsible for collecting data from data sources in the vehicle. It reads metrics configurations (definitions for data collections and transformations) and generates metrics reports containing the collected data. This service includes an edge processing engine.
Telemetry SDK: Telemetry client library
This library provides convenient access to the Telemetry service and helps OEM client apps manage metrics configurations, collect metrics reports, and receive relevant event notifications.
For more information, see Rust telemetry client library.
Telemetry SDK: Configurable Publisher Registry library
The Configurable Publisher Registry library simplifies the process of creating custom telemetry publishers and registering them with the Configurable Publisher Registry. We provide implementations for Java and Rust.
For more information, see Configurable Publisher Registry library.
Telemetry simulator
A CLI tool that lets you simulate metrics configurations based on prerecorded or artificial publisher data.
Metrics config generator (MCG)
This cloud service is targeted to generate the highly optimized protobuf-based metrics configurations from a user-friendly JSON format. MCG also performs validation of metrics configurations and can automatically infer message types based on the observed and processed vehicle signals.
Cloud-based telemetry simulation
A backend-system that lets you manage and run telemetry simulations on Google cloud at scale. You can deploy the cloud-based simulation system by using Terraform on any Google Cloud tenant.