JT808-2013 Live Streaming (JT1078) + AI Events Feature Request & Clarification
Video streaming is definitely not supported yet. As for events, I'm not sure.
Thanks for the quick response.
Since JT1078 video streaming is not supported yet, I would like to understand the recommended direction if we plan to implement it ourselves and contribute it back.
Extending JT1078 Support
Could you please clarify the best way to integrate JT1078 into Traccar?
Should this be added as an extension to the existing JT808ProtocolDecoder?
Or is it better to build a completely separate JT1078 protocol module that runs in parallel?
Does Traccar currently have any internal hooks for handling binary frame streams that are not location messages?
A minimal guideline would help us structure the implementation correctly.AI / Intelligent Event Packets
Regarding the uncertainty about extended AI/ADAS events:
Would it help if I provide sample raw hex packets for:
DMS fatigue alarms
Distraction / phone usage
ADAS collision / lane departure
Driver abnormal behaviour
Cabin monitoring alerts (standing person, crowding, etc.)
If you confirm that Traccar can parse them with custom decoders, we can prepare a pull request adding support for these event IDs.Multimedia Upload
One last question:
Is JT808 multimedia upload (images / short video clips) expected to work, or does that also require additional implementation?
This will help us evaluate whether video snapshots can be integrated before real-time streaming.
Happy to share sample packets, device logs, and work on a contribution if we get basic guidance on the preferred architecture.
Thanks again!
Where did you find the existing JT808ProtocolDecoder? I also recommend to discuss things one at a time.
Thanks for pointing that out my mistake. I assumed there was an existing JT808 decoder because several Chinese device protocols look similar, but I understand now that Traccar doesn’t include a full JT808 implementation.
To keep things simple and focus on one topic as you suggested, here’s the first step:
Topic 1 — Support for Extended Event Messages (AI / ADAS / DMS)
Before attempting anything with live video or JT1078, I want to clarify just this:
If I provide raw HEX samples of the extended alarm/event messages our devices send
(fatigue, distraction, phone usage, lane departure, pedestrian warning, cabin monitoring, etc.),
would it be feasible to add support for them by extending an existing protocol handler, or would this require creating a completely new protocol from scratch?
I can share:
Full raw packet
Header breakdown
Event ID mapping
Manufacturer documentation
Once you confirm the recommended approach, I’ll post the packet samples.
It probably makes sense to extend an existing protocol.
Hello,
We are integrating JT808-2013 devices that support JT1078 live video streaming, and we are planning to process AI-based events (ADAS, DMS, fatigue, phone usage, people standing detection, etc.) on our server.
Before proceeding, we would like to clarify Traccar’s current support level and request guidance on the following:
JT1078 Live Streaming
Does Traccar support receiving real-time AV streams from JT808/JT1078 devices?
If not fully supported, is there any recommended approach or existing module for:
Video channel negotiation
Stream forwarding/proxying (RTMP/FLV/WebRTC)
Recording or server-side decoding
AI / Intelligent Event Support
Our devices generate additional event packets such as:
Driver fatigue / distraction
Mobile-phone usage
ADAS collision warnings
Cabin monitoring (standing student detection, etc.)
Is Traccar currently able to parse these extended JT808 alarm types, or do we need to implement a custom decoder?
Recommended Implementation Path
We plan to run all compute (AI + analytics) locally.
Looking for guidance on:
How Traccar handles JT808 multimedia uploads
Whether JT1078 video frames can be intercepted for external AI processing
Best way to extend the decoder for new event IDs
Any known limitations in the current JT808 pipeline.
We are ready to contribute or develop custom modules if needed.
Any direction or technical insight will be very helpful. Thank you!