AWS IoT FleetWise Specialized2021年〜
A service for efficiently collecting, transferring, and analyzing vehicle data in the cloud
What It Does
AWS IoT FleetWise is a service for efficiently collecting sensor data from cars and commercial vehicles and transferring it to the cloud. It supports in-vehicle network protocols like CAN bus and OBD-II, filters only the data you need, and sends it to the cloud. You can define vehicle models to standardize sensor signals and manage and analyze data across your entire vehicle fleet in a unified way.
Use Cases
Used for collecting and analyzing driving data from connected cars, predictive maintenance of vehicle fleets (predicting battery degradation and part wear), collecting sensor data for autonomous driving development, and fuel efficiency optimization and route analysis for commercial vehicles.
Everyday Analogy
Think of it like the vehicle management system of a large trucking company. From the many sensors installed in each truck, it selectively sends only the data that headquarters needs. Since sending all data would result in enormous communication costs, you can set smart collection rules like 'only send detailed data when engine temperature is abnormal.'
What Is AWS IoT FleetWise?
AWS IoT FleetWise is a service for leveraging the massive amounts of sensor data generated by vehicles in the cloud. Modern cars have hundreds of sensors that generate enormous volumes of data - engine RPM, vehicle speed, battery voltage, tire pressure, and more. However, sending all this data to the cloud would be prohibitively expensive. FleetWise provides a mechanism to filter data at the edge and efficiently transfer only the necessary data to the cloud.
Vehicle Models and Signal Catalogs
In FleetWise, you first create a signal catalog to define the vehicle's sensor signals (vehicle speed, engine temperature, brake pressure, etc.). Then you create vehicle models to define which signals are available for each vehicle type. This standardization lets you handle vehicle data from different manufacturers and models in a unified format. Individual vehicles are registered based on vehicle models and managed as fleets. For practical tips on working with vehicle models and signal catalogs, technical books on Amazon are also a helpful resource.
Data Collection Campaigns
Data collection is defined as campaigns. In a campaign, you specify the signals to collect, collection conditions (time-based or condition-based), and the destination (S3 or Timestream). With condition-based collection, you can collect data only when specific events occur, such as 'when brake pressure spikes suddenly' or 'when battery voltage drops below a threshold.' This dramatically reduces data transfer volume while ensuring critical data is reliably captured.
Things to Watch Out For
- Billed based on the number of vehicles and messages. Run cost estimates in advance for large fleets
- An edge agent must be deployed on the vehicle side. Check the supported in-vehicle platform requirements beforehand
- CAN bus decode definitions (DBC files) must be obtained from the vehicle manufacturer and are often not publicly available
