ROS Log Files (rosbags)

Introduction

A rosbag or bag is a file format in ROS for storing ROS message data. These bags are often created by subscribing to one or more ROS topics, and storing the received message data in an efficient file structure. MATLAB® can read these rosbag files and help with filtering and extracting message data. The following sections detail the structure of rosbags in MATLAB® and the workflow for extracting data from them.

MATLAB rosbag Structure

When accessing rosbag log files, call rosbag and specify the file path to the object. MATLAB then creates a BagSelection object that contains an index of all the messages from the rosbag.

The BagSelection object has the following properties related to the rosbag:

  • FilePath: a character vector of the absolute path to the rosbag file.

  • StartTime: a scalar indicating the time the first message was recorded

  • EndTime: a scalar indicating the time the last message was recorded

  • NumMessages: a scalar indicating how many messages are contained in the file

  • AvailableTopics: a list of what topic and message types were recorded in the bag. This is stored as table data that lists the number of messages, message type, and message definition for each topic. For more information on table data types, see Access Data in Tables (MATLAB). Here is an example output of this table:

    ans = 
    
                               NumMessages         MessageType          MessageDefinition
                               ___________    ______________________    _________________
    
        /clock                 12001          rosgraph_msgs/Clock       [1x185  char]    
        /gazebo/link_states    11999          gazebo_msgs/LinkStates    [1x1247 char]    
        /odom                  11998          nav_msgs/Odometry         [1x2918 char]    
        /scan                    965          sensor_msgs/LaserScan     [1x2123 char]    
  • MessageList: a list of every message in the bag with rows sorted by time stamp of when the message was recorded. This list can be indexed and you can select a portion of the list this way. Calling select allows you to select subsets based on time stamp, topic or message type.

Also, note that the BagSelection object contains an index for all the messages. However, you must still use functions to extract the data. For extracting this information, see readMessages for getting messages based on indices as a cell array or see timeseries for reading the data of specified properties as a time series.

Workflow for rosbag Selection

When working with rosbags, there is a general procedure of how you should extract data.

  • Load a rosbag: Call rosbag and the file path to load file and create BagSelection.

  • Examine available messages: Examine BagSelection properties (AvailableTopics, NumMessages, StartTime, EndTime, and MessageList) to determine how to select a subset of messages for analysis.

  • Select messages: Call select to create a selection of messages based on your desired properties.

  • Extract message data: Call readMessages or timeseries to get message data as either a cell array or time series data structure.

  • Visualize, analyze or process data: Use the extracted data for your specific application. You can plot data or develop algorithms to process data.

The following figure also shows the workflow.

Limitations

There are a few limitations in the rosbag support within MATLAB:

  • MATLAB can only parse uncompressed rosbags. See the ROS Wiki for a tool to decompress a compressed rosbag.

  • Only rosbags in the v2.0 format are supported. See the ROS Wiki for more information on different bag formats

  • The file path to the rosbag must always be accessible. Because the message selection process does not retrieve any data, the file needs to be available for reading when the message data is accessed.

See Also

| |

Related Topics