Skip to content

Hesai: HW Monitor's synchronous TCP communication leads to long callback times #303

Description

@mojomex

As discussed in TIER IV INTERNAL LINK (analysis)
and TIER IV INTERNAL LINK (discussion),
the fetch_diag_from_sensor timer callback takes up to 36 ms in some cases (mean is around 17 ms for QT128).

While this does not affect pointcloud decode performance or latency (pointcloud decode is running in its own thread), this can become a problem in containers such as the ones used for Autoware's pointcloud_preprocessor. The long-running callback could block other callbacks from running, even though it is not actively computing, but rather just waiting for TCP comms to finish.

The proposed solution is to use an asynchronous approach instead.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions