An interactive multi-platform Connected Car simulator for generating and streaming realistic vehicle telemetry.
knowgo-vehicle-simulator has been developed to aid in the
development and validation of data-driven Connected Car services and
models that require easy access to realistic synthetic driving data,
both for static and streaming applications. It was originally designed
for generating event records for the KnowGo Car platform, but has
been generalized so that it may be useful both to Connected Car service
developers and researchers.
The vehicle simulator generates a single unique vehicle, which can be controlled either directly through the UI or through an optional REST API. This may be further interfaced with OEM-specific external data sources and models in order to permit the simulation state to act as an automotive digital twin. For fleet simulation workloads, multiple instances of the simulator may be run in parallel, with each generated vehicle being manually joined to a specified fleet.
A live demonstration of the Simulator is available here.
Installation from a binary release is recommended. Regular releases are made to various app stores, please refer to the one appropriate for your platform:
Releases can also be obtained directly from GitHub.
For deployment of a self-contained web-based instance of the simulator, a number of deployment options have been provided:
Multi-arch images are provided under knowgo/knowgo-vehicle-simulator. The image can be run directly as:
$ docker run -p 8086:8086 knowgo/knowgo-vehicle-simulator
To create a Kubernetes Deployment including a single instance of the
simulator:
$ kubectl apply -f https://raw.githubusercontent.com/knowgoio/knowgo-vehicle-simulator/simulator-deployment.yaml
An optional Service exposing the simulator port on the cluster can
also be applied:
$ kubectl apply -f https://raw.githubusercontent.com/knowgoio/knowgo-vehicle-simulator/simulator-service.yaml
For additional documentation and tutorials, please refer to the documentation.
A Postman Collection and pre-configured environment for interacting with the Simulator REST API in a local simulation environment is available here.
The Simulator itself consists of several different components:
- The Vehicle Simulation model
- An
Event loopfor generating vehicle events, run as either an Isolate or Web Worker depending upon the target platform. - An optional
HTTP Server isolatefor exposing a REST API with basic vehicle controls - starting/stopping the vehicle, updating the vehicle state, handling vehicle notifications, and querying vehicle events.
As the simulation state can not be shared directly across the isolates, the simulation model in the main isolate acts as the source of truth across the system:
- Updates from the
Event loopare applied to the simulation model periodically, in line with the event generation frequency: once per second by default. - The
HTTP Server isolatemaintains its own cached copy of the simulation state, which is updated with changes from the Event isolate, UI interaction, and the REST API. Changes received through the REST API are cached in theHTTP Server isolateand proxied back to the simulation model directly. - The UI in the
main isolateis redrawn based on changes to the simulation model, triggered by UI interaction and updates from theEvent looporHTTP Server isolate.
An overview of the overall interactivity patterns for the different target platforms is provided in the table below:
| Flutter Web | Other Target Platforms |
|---|---|
 |
 |
- Linux desktop
- Windows desktop
- macOS desktop
- Web
- Android
- iOS
By default, generated events are only logged in the console. Events can be published to a custom notification endpoint, a KnowGo API backend, as well as MQTT and Kafka brokers (as well as any combination thereof). The specific configuration for each is outlined below.
Configuration of the simulator can be tuned through a config.yaml
file, which will be parsed and updated by configuration changes within
the UI. The format of the file is:
sessionLogging: true
eventLogging: true
# Allow unauthenticated requests to REST API
allowUnauthenticated: true
# Optional endpoint to post generated events to
notificationUrl: http://myserver.com/endpoint
# Optional KnowGo Backend Configuration
knowgo:
server: <knowgo-API-server>
apiKey: <knowgo-API-Key>
# Optional Kafka Broker Configuration
kafka:
broker: <kafka-broker-address>
topic: <kafka-topic>
# Optional MQTT Broker Configuration
mqtt:
broker: <MQTT-broker-address>
topic: <MQTT-topic>A number of environment variables can also be set:
| Environment Variable | Description | Default value |
|---|---|---|
| KNOWGO_VEHICLE_SIMULATOR_CONFIG | Path to config file | <appDocDir>/knowgo_vehicle_simulator/config.yaml |
| KNOWGO_VEHICLE_SIMULATOR_LOGS | Path to log directory | <appDocDir>/knowgo_vehicle_simulator/logs |
| KNOWGO_VEHICLE_SIMULATOR_IP | IP address to bind for REST API | 0.0.0.0 |
| KNOWGO_VEHICLE_SIMULATOR_PORT | HTTP port to bind for REST API | 8086 |
| KNOWGO_SIGNING_KEY | Signing and validation secret for API keys | secret-key |
Please file feature requests and bugs at the issue tracker.
Licensed under the terms of the MIT license, the full version of which can be found in the LICENSE file included in the distribution.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent