J2735 (2016 version) MessageFrame cannot be decoded?

[Yufei-Yan]

Hi, I'm current facing this issue: MessageFrame cannot be decoded after encoded.

A BSM is generated with some customized fields. When testing to encode and decode BSM without encapsulating in MessageFrame, both der_encode_to_buffer() and ber_decode() functions work well.

But after encapsulating BSM in MessageFrame, encoding function seems to work (I'm not sure if it is working correctly), decoding function does not work. It gives the result code: 2, which is failure.

All messages are properly written, since I can verify them by decode the byte sequence into xml format for verification.

The message below is xml format:

<MessageFrame>
    <messageId>20</messageId>
    <value>
        <BasicSafetyMessage>
            <coreData>
                <msgCnt>0</msgCnt>
                <id></id>
                <secMark>0</secMark>
                <lat>0</lat>
                <long>0</long>
                <elev>0</elev>
                <accuracy>
                    <semiMajor>0</semiMajor>
                    <semiMinor>0</semiMinor>
                    <orientation>0</orientation>
                </accuracy>
                <transmission><neutral/></transmission>
                <speed>0</speed>
                <heading>470220800</heading>
                <angle>847118336</angle>
                <accelSet>
                    <long>2140209152</long>
                    <lat>2139422720</lat>
                    <vert>0</vert>
                    <yaw>1908211712</yaw>
                </accelSet>
                <brakes>
                    <wheelBrakes>
                        0000000000000000000000000000000000000
                    </wheelBrakes>
                    <traction><unavailable/></traction>
                    <abs><unavailable/></abs>
                    <scs><unavailable/></scs>
                    <brakeBoost><unavailable/></brakeBoost>
                    <auxBrakes><unavailable/></auxBrakes>
                </brakes>
                <size>
                    <width>0</width>
                    <length>0</length>
                </size>
            </coreData>
            <platoon>
                <status>00</status>
                <position>00</position>
                <ecef>
                    <ecefX>D6 9A 07 91</ecefX>
                    <ecefY>D6 9A 07 91</ecefY>
                    <ecefZ>D6 9A 07 91</ecefZ>
                </ecef>
                <brakePercentage><minPressure/></brakePercentage>
                <platoonPos>
                    <headway>-2012020736</headway>
                    <range>604307456</range>
                    <rangeRate>2138636288</rangeRate>
                </platoonPos>
                <goal>
                    <desiredAcc>2138374144</desiredAcc>
                    <goalHeadway>1614479360</goalHeadway>
                    <goalSpeed>1476395008</goalSpeed>
                </goal>
            </platoon>
        </BasicSafetyMessage>
    </value>
</MessageFrame>

What could be wrong?

[DavidKelleySCSC]

Please send your ASN spec along as well (or by private mail to me if you wish). Your addition of the is valid with some resections, but this data looks rather fishy to me. So I am sure that your inner BSM /message/ is not valid to start with. Valid BSM Part II data would not be encoded that way, but is found in the regional or the partII elements. Once that works, you can work on encoding it in the message frame, which to my human eye looks fine. Regards, David Kelley, SAE J2735 DSRC Msg Set Editor

[brchiu]

@Yufei-Yan , compile your program with -g flag and use gdb stepping into function call then you can find where it failed.

[Yufei-Yan]

Hi, @DavidKelleySCSC Can I have your email please? I cannot post it here. sorry.

I admit that the ASN file could be wrong or some fields may not be defined correctly. But the wired thing is that this BSM can be successfully encoded and decoded by der_encode_to_buffer() and ber_decode(). When this BSM is encapsulated into MessagaFrame, ber_decode() cannot decode the MessageFrame, and it gives an error at value in MessageFrame. This makes me confused.

[Yufei-Yan]

Hi, @brchiu

Thanks for you reply. The program is running on an embedded platform, and the OS is rather minimal. I doubt it has gdb. And also, the function fails in ber_decode(), it gives error at value in MessageFrame when decoding. The program does not crash.

What else can I do to debug this issue?

Thank you.

[brchiu]

@Yufei-Yan , could you have a trial using https://github.com/brchiu/asn1c/tree/velichkov_s1ap_plus_option_group ?

asn1c community members have fixed several problems found in master branch of @vlm's repository.

By the way, under examples/sample.source.J2735 directory, you can compile a sample utility program called j2735-dump with debug flag -g on Linux platform. Then feed it with the data you want to encode or decode then step into each function call on Linux platform.

For example, I use it to process the xml you post with ./j2735-dump -dd -ixer message1.xer -oder message1.ber then I get the following debug message.

https://gist.github.com/brchiu/a56c690e4c159f36ac574fbb8c387632

After AD: Encoded in 113 bytes of DER, at line 1200 of this log, there are still more data in your <platoon> tag which does not exist in the default J2735 ASN.1 which I purchased. So it displays AD: RC_WMORE, continuing read=0, cons=1 with 1..0-0..1 in line 1238.

[Yufei-Yan]

@brchiu , thank you so much for your help.

<platoon> tag is our customized tag. It's not in standard J2735 ASN file. I'll try the tool you mentioned for sure.

The weird thing is when I encode and decode a simple BSM with <platoon> tag, it is successful. But when I embed this BSM in a MessageFrame, the function ber_decode() gives an error, failing at value in MessageFrame.

Again, I appreciate your effort.

[ronMilne]

I'm facing the exact same error has this been fixed?

[Yufei-Yan]

@ronMilne , not yet confirmed if this is a bug or not. Haven't got any time to test it since then. Can I ask if you are working on standard J2735 ASN file or a customized ASN file?

I was working on a customized ASN file, and came across this issue. I might have time to test it again in the next couple of weeks.

Thanks.

[DavidKelleySCSC]

To repeat a prior offer...

Please send your ASN spec along as well (or by private mail to me if you wish).

Can take a look at it for issues, it is easily easy to 'break' the frame and I am aware the SAE TC needs to get a suitable example of how to do this out into the world. Regards, David Kelley, SAE J2735 DSRC Msg Set Editor

[ronMilne]

My ASN is the standard J2735 (latest). I'm successfully encoding/decoding a BSM but not the MessageFrame.

MessageFrame_t* msgFrame = (MessageFrame_t*)calloc(1, sizeof *msgFrame);
msgFrame->messageId = 100;
msgFrame->value.present = (MessageFrame__value_PR)MessageFrame__value_PR_BasicSafetyMessage;
msgFrame->value.choice.BasicSafetyMessage = *bsm;

asn_encode_to_new_buffer_result_t rsltMsg = asn_encode_to_new_buffer(0, ATS_DER, &asn_DEF_MessageFrame, msgFrame);

[DavidKelleySCSC]

Okay, but 100 is not the defined ID for the BSM, 0x20 is. You have got to match up the message Id to an anticipated message content for this to work. If you were to define your own content (in the provided ranges) you have to provide that definition to the ASN tool when it is complied as well. In the absence of defining your own message types, use this to select into what ASN packed strut the source C strut will be placed..

DSRCmsgID ::= INTEGER (0..32767)
   --  content removed here , dck
   --
   -- UPER forms
   --
   mapData                           DSRCmsgID ::=   18 -- MAP, intersections 
   signalPhaseAndTimingMessage       DSRCmsgID ::=   19 -- SPAT
   -- Above two entries were adopted in the 2015-04 edition
   -- Message assignments added in 2015 follow below   < and that n means for 2016 as well, dck
   basicSafetyMessage                DSRCmsgID ::=   20 -- BSM, heartbeat msg
   commonSafetyRequest               DSRCmsgID ::=   21 -- CSR
   emergencyVehicleAlert             DSRCmsgID ::=   22 -- EVA
   intersectionCollision             DSRCmsgID ::=   23 -- ICA
   nmeaCorrections                   DSRCmsgID ::=   24 -- NMEA
   probeDataManagement               DSRCmsgID ::=   25 -- PDM
   probeVehicleData                  DSRCmsgID ::=   26 -- PVD
   roadSideAlert                     DSRCmsgID ::=   27 -- RSA
   rtcmCorrections                   DSRCmsgID ::=   28 -- RTCM
   signalRequestMessage              DSRCmsgID ::=   29 -- SRM
   signalStatusMessage               DSRCmsgID ::=   30 -- SSM
   travelerInformation               DSRCmsgID ::=   31 -- TIM
   personalSafetyMessage             DSRCmsgID ::=   32 -- PSM

   --
   -- The Below values are reserved for local message testing use
   -- 
   testMessage00                     DSRCmsgID ::=  240 -- Hex 0xF0 
   testMessage01                     DSRCmsgID ::=  241  
   testMessage02                     DSRCmsgID ::=  242  
   testMessage03                     DSRCmsgID ::=  243  
   testMessage04                     DSRCmsgID ::=  244   
   testMessage05                     DSRCmsgID ::=  245  
   testMessage06                     DSRCmsgID ::=  246  
   testMessage07                     DSRCmsgID ::=  247  
   testMessage08                     DSRCmsgID ::=  248  
   testMessage09                     DSRCmsgID ::=  249  
   testMessage10                     DSRCmsgID ::=  250 
   testMessage11                     DSRCmsgID ::=  251  
   testMessage12                     DSRCmsgID ::=  252  
   testMessage13                     DSRCmsgID ::=  253  
   testMessage14                     DSRCmsgID ::=  254  
   testMessage15                     DSRCmsgID ::=  255-- Hex 0xFF  
   --
   -- All other values are reserved for std use
   --

[ronMilne]

One other thing I had to do was set coreData.msgCnt = 1; on the BSM. This was picked up asn_check_constraints on the MessageFrame. Thanks greatly, now have a message that transmits.

[DavidKelleySCSC]

And please remember to increment it (modulo 0~127) every time the BSM is updated.

I would suggest that as a best practice you write a utility routine that sets ALL the values found in the BSMcoreData to their default invalid values (and in that data frame they are all required, so for ascn1c you will not have to deal with various pointer issues). If you are wrapping this code in C++, this would be a good constructor step. I would estimate that at least half the support queries I get on that standard deal in one way or another with some of the data elements not set to valid values and then being forgotten.

[ronMilne]

Given the following BasicSafetyMessage_t bsm; bsm = (BasicSafetyMessage_t)calloc(1, sizeof(BasicSafetyMessage_t));

What is the correct way to create the TemporaryID id field? When calling ASN_STRUCT_FREE? Is there a better way of doing the following

    TemporaryID_t* tempId = (TemporaryID_t*)calloc(1, sizeof(TemporaryID_t));
    tempId->size = 4;
    tempId->buf = (uint8_t*)calloc(1, 4);

    const uint8_t tempMacId[] = { 0x00, 0x00, 0x00, 0x02 };
    memcpy(tempId->buf, tempMacId, 4);

             bsm->coreData.id = *tempId;

Thanks.

[nprobert]

I have my J2735_201603SA decoder working now. Don't know why folks here are doing BER/DER as UPER is the chosen standardized format. I do find that ASN_STRUCT_FREE crashes for J2735, but not for the P1609.2 decoder I have (OER).

[Hutch07]

Neal - I am trying to encode a J2735 Map file for use in an RSU. What J2735_201603SA decoder are you using?