Matlab Import Export Vehicle Data

Import, Export, and Decode Vehicle Data

Import, decode, and export vehicle network log files in MATLAB using Vehicle Network Toolbox (VNT).

When to Use

• Reading .mf4/.mdf/.dat, .blf, .asc, or .txt vehicle bus log files

• Decoding raw CAN/CAN FD frames to signals via DBC or ARXML databases

• Demultiplex signals from raw CAN/CAN FD frames via DBC databases in the case of simple multiplexing (R2026a or earlier)

• Decoding LIN frames via LDF databases (R2025a+)

• Handling polymorphic return types from VNT read functions

• Writing MATLAB timetable data to MDF/MF4 files with mdfWrite

• Writing CAN/CAN FD message data to BLF files with blfwrite

• Converting unsupported timetable types (logical, categorical, cell, datetime columns) for MDF export

When Not To Use

• Simulink CAN configuration

• Live CAN channel access

• XCP/A2L workflows

• Demultiplexing of signals in the case of hierarchical or extended multiplexing

• Demultiplexing of signals via ARXML or LIN databases

• Writing LIN data to BLF (unsupported by VNT's blfwrite -- use MDF instead)

• Writing to ASC or TXT files (no canMessageExport exists -- use BLF or MDF instead)

Prerequisites

Before running any operations, determine the MATLAB version:

r = matlabRelease;
tok = regexp(r.Release, 'R(\d{4})([ab])', 'tokens');
matlabYear = str2double(tok{1}{1});
matlabLetter = tok{1}{2};

Before calling BLF/ASC/TXT import or CAN/CAN FD/LIN decode functions, call detect_matlab_toolboxes to confirm Vehicle Network Toolbox is installed. If unavailable, tell the user. Once confirmed in a session, do not re-check.

MDF functions (mdfRead, mdfChannelGroupInfo, mdfInfo, mdfChannelInfo) are base MATLAB (R2023a+) and do not require VNT. Do not use the deprecated mdf() object constructor -- always use the functional API. For reading all groups blindly, mdfRead alone is acceptable. To selectively read or inspect structure first, use mdfChannelGroupInfo or mdfInfo.

For R2026a and earlier, multiplexing is not natively supported in Vehicle Network Toolbox. Apply the demultiplexing modifications described below to decoded signals.

Decode Pipeline

Database Loading

db = canDatabase("path/to/file.dbc");       % CAN and CAN FD
db = arxmlDatabase("path/to/file.arxml");   % CAN only (not CAN FD)
linDb = linDatabase("path/to/file.ldf");    % LIN (R2025a+)

Load once, pass to all decode calls. Multiple databases: canMessageTimetable(tt, [db1, db2]).

Database inspection for DBC files and classification of the multiplexing scheme

In case of DBC files only, determine whether the file contains definitions of multiplexed signals that use extended or hierarchical multiplexing. For MATLAB R2026a or earlier, it is not possible to determine the multiplexing scheme from the canDatabase object directly. Therefore, inspect the raw DBC text to classify the multiplexing scheme, by reading the DBC file as text using fileread():

dbcText = fileread("path/to/file.dbc");

Check for extended multiplexing indicators. See references/dbc-multiplexing-syntax.md for the full syntax guide — consult it when parsing unfamiliar DBC signal markers or SG_MUL_VAL_ entries.

hasSgMulValEntries = ~isempty(regexp(dbcText, '^SG_MUL_VAL_\s+\d+', 'once', 'lineanchors'));
hasNestedMuxor = ~isempty(regexp(dbcText, 'SG_\s+\w+\s+m\d+M\s', 'once'));
hasExtendedMux = hasSgMulValEntries || hasNestedMuxor;

Note: SG_MUL_VAL_ may appear in the NS_ (namespace) section as a keyword declaration — this does NOT indicate extended multiplexing. Only match actual data entries that start at the beginning of a line followed by a numeric message ID. If hasExtendedMux is true, identify which specific messages are affected by parsing the message IDs from SG_MUL_VAL_ entries:

sgMulValMsgIds = regexp(dbcText, '^SG_MUL_VAL_\s+(\d+)', 'tokens', 'lineanchors');
extendedMuxMsgIds = unique(cellfun(@(x) str2double(x{1}), sgMulValMsgIds));

For R2026a and earlier, filter out messages with extended multiplexing (identified in extendedMuxMsgIds) and inform the user.

Key Constraints

• Polymorphic returns: All VNT read functions (canSignalImport, blfread, mdfRead, canMessageImport) return different types based on channel count and parameters. Always guard with iscell()/istimetable()/isstruct(). When checking multi-channel results, use iscell() as the primary guard to branch between cell array (multi-channel) vs direct timetable (single channel), then iterate or extract with {n} or {i}.

• ARXML + CAN FD: arxmlDatabase does not support CAN FD in MATLAB. It cannot be used with canFDMessageTimetable. When the user provides an ARXML file for CAN FD data, always explain that ARXML does not support CAN FD and they must use canDatabase with a DBC file instead. Do not silently switch to DBC without explaining why.

• ARXML for CAN (non-FD): When the user provides an ARXML file for standard CAN decode, use arxmlDatabase exclusively. Do not also load canDatabase as a fallback -- the ARXML database object works identically in canMessageTimetable and canSignalTimetable. Only use canDatabase when the user provides a DBC file.

• No canFDSignalTimetable: This function does not exist. Use canSignalTimetable for both CAN and CAN FD signals.

• BLF LIN requires ProtocolMode="LIN": Without it, blfread returns only CAN data.

• No signal-level LIN decode: Only linMessageTimetable exists. Access signal values from timetable columns directly.

• MDF raw CAN detection: Use channel name prefixes CAN_DataFrame / LIN_Frame (ASAM standard). NEVER use SourceBusType or AcquisitionSourceType metadata -- not even as a fallback. These fields are vendor-specific and unreliable.

• canMessageImport default format: Returns legacy can.Message array. Always pass OutputFormat="timetable".

• Cell extraction: Multi-channel calls to blfread, mdfRead, and canMessageImport return cell arrays. Extract with {n} before processing.

• Use canSignalImport for ASC/TXT signals: It decodes in one call. Only use canMessageImport when raw message timetables are needed.

File-Specific References

• ASC import -- canSignalImport/canMessageImport with Vector and Kvaser vendors, polymorphic return type handling for multi-channel files

• BLF import -- blfinfo for file inspection, blfread with and without ChannelID, CAN vs LIN protocol selection

• MDF import -- mdfRead/mdfChannelGroupInfo/mdfChannelInfo, detecting raw CAN/LIN groups by channel name, reading specific groups

Protocol References

• CAN decode -- canMessageTimetable and canSignalTimetable pipelines, DBC and ARXML database usage

• CAN FD decode -- canFDMessageTimetable pipeline, why ARXML does not work for CAN FD

• LIN decode -- linMessageTimetable usage, accessing signal values from timetable columns (no signal-level function exists)

• Database objects -- canDatabase vs arxmlDatabase vs linDatabase, which decode functions accept which database type

Demultiplexing References

• Simple demultiplexing -- canDatabase, canMessageTimetable, canFDMessageTimetable, canSignalTimetable pipelines and signal demultiplexing

Export Workflows

MDF Writing

Writing MATLAB timetable data to MDF/MF4 files with mdfWrite.

BLF Writing

Writing CAN and CAN FD message data to BLF files with blfwrite.

LIN is not supported by VNT's blfwrite. Use MDF for LIN logging. When the user asks to write LIN data to BLF, state this limitation immediately -- do not attempt workarounds or partial solutions before explaining the constraint.

ASC/TXT writing is not supported. There is no canMessageExport counterpart to canMessageImport. Use BLF or MDF for export.

Export References

• MDF writing -- mdfWrite/mdfCreate, type conversion, ASAM CAN/LIN naming, overwriting groups, metadata preservation

• BLF writing -- blfwrite with CAN/CAN FD, canMessageTimetable/canFDMessageTimetable formatting, channel callbacks

Copyright 2026 The MathWorks, Inc.