Chinese tracker OBDII buffer error

Tomislav6 years ago

Hello,

I am testing traccar with chinese tracker (unbranded) that hooks on OBDII.

According to this post: https://www.traccar.org/forums/topic/dodgy-chinese-tracker/ it seems that it uses GT06,
since my messages also begin with 78780d010.

It connected properly and I could se messages like:

2018-08-02 12:41:35 INFO: [f938725c] id: xxx, time: 2018-08-02 12:36:44, lat: xx, lon: xx, speed: xx, course: xx

In meantime I changed timezone since messages were sent with future time (don't know if that affected it).
Despite messages there was little to no movement of marker on map (it works correctly with test-genrator so it's not websocket).
But route report showed cca 10km almost correct route.

Then device started going online/offline every few secs and i started getting messages like:

2018-08-02 12:42:17  WARN: [4996fac4] error - readerIndex(17) + length(1) exceeds writerIndex(17): PooledSlicedByteBuf(ridx: 17, widx: 17, cap: 17/17, unwrapped: PooledUnsafeD$
2018-08-02 12:42:17  INFO: [4996fac4] disconnected
2018-08-02 12:42:18  INFO: [cff72d4d] connected
2018-08-02 12:42:19 DEBUG: [cff72d4d: 5023 < xxx.xx.xxxx.xxx] HEX: 78780d01035288707774647100204fcb0d0a
2018-08-02 12:42:19 DEBUG: [cff72d4d: 5023 > xxx.xx.xxxx.xxx] HEX: 787805010020e9570d0a
2018-08-02 12:42:22 DEBUG: [cff72d4d: 5023 < xxx.xx.xxxx.xxx] HEX: 78780c691208020c240c0100548ebe0d0a
2018-08-02 12:42:22  WARN: [cff72d4d] error - readerIndex(17) + length(1) exceeds writerIndex(17): PooledSlicedByteBuf(ridx: 17, widx: 17, cap: 17/17, unwrapped: PooledUnsafeD$
2018-08-02 12:42:22  INFO: [cff72d4d] disconnected
2018-08-02 12:42:23  INFO: [a9cf2e02] connected

Am I using wrong protocol?
I built traccar from source few days ago if it matters.

Thanks

Anton Tananaev6 years ago

Protocol is correct, but there are a lot of different GT06-based devices with slight differences. Looks like this one is not supported yet. You need to get protocol documentation from your device vendor.

Tomislav6 years ago

Thanks,

doubt that vendor is even known.
After day of testing, the device is behaving correctly and I didn't see anymore of those errors, and the routes/positions are suprisingly accurate.
I guess it will do alright for testing.