mavlink2rest
is a tool that offers a RESTful API over the MAVLink protocol, facilitating seamless communication between unmanned systems and web applications. The tool supports the ArduPilotMega dialect, iCAROUS, and UAVionix, making it an ideal solution for developers who want to build custom interfaces for unmanned systems.
The current version supports the ardupilotmega dialect, that includes common, icarous and uavionix.
- 💻 Windows
- 🍎 MacOS
- 🐧 Linux
- 🍓 Raspberry
- ARMv6 binary, ARMv7 is also available under the project releases.
For others or different releases, check the releases menu.
If you prefer, you can install via cargo, if you don't know what it is, use the download section.
- ⚙️ Cargo Install:
cargo install mavlink2rest
Capabilities via the command line:
USAGE:
mavlink2rest [FLAGS] [OPTIONS]
FLAGS:
-h, --help Prints help information
-V, --version Prints version information
-v, --verbose Be verbose
OPTIONS:
--component-id <COMPONENT_ID>
Sets the component ID for this service, for more information, check:
https://mavlink.io/en/messages/common.html#MAV_COMPONENT [default: 0]
-c, --connect <TYPE:<IP/SERIAL>:<PORT/BAUDRATE>>
Sets the mavlink connection string [default: udpin:0.0.0.0:14550]
--mavlink <VERSION>
Sets the mavlink version used to communicate [default: 2] [possible values: 1, 2]
-s, --server <IP:PORT>
Sets the IP and port that the rest server will be provided [default: 0.0.0.0:8088]
--system-id <SYSTEM_ID> Sets system ID for this service. [default: 255]
You can also use the mavlink2rest
with docker, the following command will start the service with the default settings:
docker run --rm --init -p 8088:8088 -p 14550:14550/udp --name mavlink2rest mavlink/mavlink2rest
The Dockerfile defines several environment variables that you can override at runtime:
MAVLINK_SRC: The MAVLink source connection string. Default is udpin:127.0.0.1:14550. SERVER_PORT: The IP and port for the REST server. Default is 0.0.0.0:8088. EXTRA_ARGS: Any additional command line arguments you want to pass to mavlink2rest. To customize these settings, use the -e flag with docker run:
docker run --rm --init\
-p 8088:8088 \
-p 14551:14551/udp \
-e MAVLINK_SRC="udpin:0.0.0.0:14551" \
-e SERVER_PORT="0.0.0.0:8088" \
--name mavlink2rest mavlink/mavlink2rest
to build the docker image locally, you can use the following command:
docker build --build-arg TARGET_ARCH=x86_64-unknown-linux-musl -t mavlink/mavlink2rest .
- Main webpage:
GET /
- Provides information about mavlink2rest and available messages.
- Swagger:
GET /docs
- Provides information about mavlink2rest endpoints for the REST API.
- MAVLink JSON:
-
GET /v1/mavlink|/v1/mavlink/*
. The output is a JSON that you get each nested key individually, E.g: -
Any MAVLink message will contain a normal message definition, as described in
GET /v1/helper/mavlink?name=<MESSAGE_NAME>
..- http://0.0.0.0:8088/v1/helper/mavlink?name=HEARTBEAT
{ "header": { "system_id": 255, "component_id": 0, "sequence": 0 }, "message": { "type": "HEARTBEAT", "custom_mode": 0, "mavtype": { "type": "MAV_TYPE_GENERIC" }, "autopilot": { "type": "MAV_AUTOPILOT_GENERIC" }, "base_mode": { "bits": 128 }, "system_status": { "type": "MAV_STATE_UNINIT" }, "mavlink_version": 0 } }
😎 This is really hand when creating messages.
- http://0.0.0.0:8088/v1/helper/mavlink?name=HEARTBEAT
-
... and a status structure defined as:
"status": { "time": { "counter": 2981, "first_update": "2024-05-31T11:59:44.313941926-03:00", "frequency": 10.037036895751953, "last_update": "2024-05-31T12:04:42.214212884-03:00" } }
-
POST /mavlink
. Sends the message to a specific vehicle.- For more information about the MAVLink message definition: https://mavlink.io/en/guide/serialization.html
- header: Is the mavlink header definition with
system_id
,component_id
andsequence
. - message: A valid mavlink message, for more information check
GET /v1/helper/mavlink?name=<MESSAGE_NAME>
.- Check ARM/DISARM example.
-
GET /v1/helper/mavlink?name=MAVLINK_MESSAGE_NAME
: Helper endpoint to create JSON compatible MAVLink messages, whereMAVLINK_MESSAGE_NAME
is the mavlink message name. E.g:- http://0.0.0.0:8088/v1/helper/mavlink?name=COMMAND_LONG
{ "header": { "system_id": 255, "component_id": 0, "sequence": 0 }, "message": { "type": "COMMAND_LONG", "param1": 0.0, "param2": 0.0, "param3": 0.0, "param4": 0.0, "param5": 0.0, "param6": 0.0, "param7": 0.0, "command": { "type": "MAV_CMD_NAV_WAYPOINT" // Random value }, "target_system": 0, "target_component": 0, "confirmation": 0 } }
- http://0.0.0.0:8088/v1/helper/mavlink?name=COMMAND_LONG
-
- Information:
GET /info
, provides information about the service version.- http://0.0.0.0:8088/info
{ "version": 0, "service": { "name": "mavlink2rest", "version": "0.10.0", "sha": "bd7667d", "build_date": "2021-03-03", "authors": "Author <email>" } }
- http://0.0.0.0:8088/info
curl --request GET http://0.0.0.0:8088/v1/mavlink
# The output is huge, you can get it here: [https://gist.github.com/patrickelectric/26a407c4e7749cdaa58d06b52212cb1e](https://gist.github.com/patrickelectric/26a407c4e7749cdaa58d06b52212cb1e)
curl --request GET http://0.0.0.0:8088/v1/mavlink/vehicles/1/components/1/messages/ATTITUDE
{
"message": {
"pitch": 0.11506611853837967,
"pitchspeed": 0.00003909762017428875,
"roll": 0.02339238114655018,
"rollspeed": 0.00035849903360940516,
"time_boot_ms": 87110407,
"type": "ATTITUDE",
"yaw": -2.4364013671875,
"yawspeed": 0.000020137056708335876
},
"status": {
"time": {
"counter": 22750,
"first_update": "2024-05-31T11:59:44.313941926-03:00",
"frequency": 5.495169162750244,
"last_update": "2024-05-31T13:08:45.196118069-03:00"
}
}
}
curl --request GET http://0.0.0.0:8088/v1/mavlink/ATTITUDE/status/time/last_update
"2020-03-28T14:28:51.577853-03:00"
curl --request GET http://0.0.0.0:8088/v1/helper/mavlink?name=ATTITUDE
{
"header": {
"system_id": 255,
"component_id": 0,
"sequence": 0
},
"message": {
"type": "ATTITUDE",
"time_boot_ms": 0,
"roll": 0.0,
"pitch": 0.0,
"yaw": 0.0,
"rollspeed": 0.0,
"pitchspeed": 0.0,
"yawspeed": 0.0
}
}
Request vehicle to be armed:
# ARM: param1 is 1.0
curl --request POST http://0.0.0.0:8088/v1/mavlink -H "Content-Type: application/json" --data \
'{
"header": {
"system_id": 255,
"component_id": 240,
"sequence": 0
},
"message": {
"type":"COMMAND_LONG",
"param1": 1.0,
"param2": 0.0,"param3":0.0,"param4":0.0,"param5":0.0,"param6":0.0,"param7":0.0,
"command": {
"type": "MAV_CMD_COMPONENT_ARM_DISARM"
},
"target_system": 1,
"target_component": 1,
"confirmation": 1
}
}'
Request vehicle to be disarmed:
# ARM: param1 is 0.0
curl --request POST http://0.0.0.0:8088/v1/mavlink -H "Content-Type: application/json" --data \
'{
"header": {
"system_id": 255,
"component_id": 240,
"sequence": 0
},
"message": {
"type":"COMMAND_LONG",
"param1": 0.0,
"param2": 0.0,"param3":0.0,"param4":0.0,"param5":0.0,"param6":0.0,"param7":0.0,
"command": {
"type": "MAV_CMD_COMPONENT_ARM_DISARM"
},
"target_system": 1,
"target_component": 1,
"confirmation": 1
}
}'
Note: For any invalid
GET
, you'll receive a 404 response with the error message. Note: The endpoints that allowGET
and provides a JSON output, also allow the usage of the query parameterpretty
with a boolean valuetrue
orfalse
, E.g: http://0.0.0.0:8088/helper/mavlink?name=COMMAND_LONG&pretty=true
It's also possible to connect multiple websockets with the following path /ws/mavlink
, the endpoint also accepts the query parameter filter
, the filter value should be a regex that matches MAVLink message names, E.g: /ws/mavlink?filter=.*
for all messages, /ws/mavlink?filter=RC_.*
will match RC_CHANNELS_RAW and RC_CHANNELS, resulting in the following output:
{ // First message
"header": {
"component_id": 1,
"sequence": 98,
"system_id": 1
},
"message": {
"chan10_raw": 0,
"chan11_raw": 0,
"chan12_raw": 0,
"chan13_raw": 0,
"chan14_raw": 0,
"chan15_raw": 0,
"chan16_raw": 0,
"chan17_raw": 0,
"chan18_raw": 0,
"chan1_raw": 1500,
"chan2_raw": 1500,
"chan3_raw": 1500,
"chan4_raw": 1500,
"chan5_raw": 1500,
"chan6_raw": 1500,
"chan7_raw": 1500,
"chan8_raw": 1500,
"chan9_raw": 0,
"chancount": 16,
"message_information": {
"counter": 3732,
"frequency": 4.0,
"time": {
"first_message": "2020-09-01T20:36:24.088099-03:00",
"last_message": "2020-09-01T20:51:57.278901-03:00"
}
},
"rssi": 0,
"time_boot_ms": 3122812,
"type": "RC_CHANNELS"
}
}
{ // Second message
"header": {
"component_id": 1,
"sequence": 98,
"system_id": 1
},
"message": {
"chan1_raw": 1500,
"chan2_raw": 1500,
"chan3_raw": 1500,
"chan4_raw": 1500,
"chan5_raw": 1500,
"chan6_raw": 1500,
"chan7_raw": 1500,
"chan8_raw": 1500,
"message_information": {
"counter": 3732,
"frequency": 4.0,
"time": {
"first_message": "2020-09-01T20:36:24.088310-03:00",
"last_message": "2020-09-01T20:51:57.279438-03:00"
}
},
"port": 0,
"rssi": 0,
"time_boot_ms": 3122812,
"type": "RC_CHANNELS_RAW"
}
}
For a demonstration, please check the example under the examples filder: websocket_client.py
The following benchmarks were extracted from a raspberry pi 3 connected to a pixhawk running ArduSub.
-
In idle.
6% CPU usage
-
1 client requesting all mavlink messages at 10Hz
9% CPU usage
-
1 client requesting all mavlink messages at 100Hz
20% CPU usage (~5% each core)
-
1 websocket with no filters
11% CPU usage
-
5 websockets with no filters
24% CPU usage (14% @ 1 core, ~3% @ 3 cores)
-
20 websockets with filter only for ATTITUDE message (receiving at 10Hz)
9% CPU usage
-
20 websockets with filter only for NAMED_VALUE_FLOAT message (receiving at 70Hz)
17% CPU usage (9% @ 1 core, ~2% @ 3 cores)
-
20 websockets with no filters
48% CPU usage (20% @ 1 core, ~9% @ 3 cores)
-
1 client requesting all mavlink messages 1000 times
60% CPU usage (~15% each core) Time taken for tests 3.7 seconds Total requests 1000 Successful requests 1000 Failed requests 0 Requests per second 273.60 [#/sec] Median time per request 3ms Average time per request 4ms
-
10 clients requesting all mavlink messages, 100 requests for each client.
140% CPU usage (~46% each core) Time taken for tests 1.4 seconds Total requests 1000 Successful requests 1000 Failed requests 0 Requests per second 733.14 [#/sec] Median time per request 13ms Average time per request 13ms Sample standard deviation 3ms
-
100 clients requesting all mavlink messages, 1000 requests for each client.
140% CPU usage (~46% each core) Time taken for tests 13.8 seconds Total requests 10000 Successful requests 10000 Failed requests 0 Requests per second 725.83 [#/sec] Median time per request 132ms Average time per request 137ms Sample standard deviation 54ms