English

简体中文

English

简体中文

Appearance

On-Robot Host and Ubuntu ROS 2 Dev PC ​

On-Robot Host ​

  • Shell Execution Context: On-robot host shell.
  • Risk Level: Sourcing environments and read-only subscriptions are Read-Only; invoking recording or control services escalates risk per relevant sections.
bash
source /opt/ros/humble/setup.bash
source /home/realman/workspace/rm_robot_ws/install/setup.bash
python3 -c 'import rclpy; print("rclpy ready")'

  • Business application processes MUST decouple into distinct architecture layers: ROS 2 communication adapters, operational state machines, and file transfer modules.

  • Service call timeouts and business logic execution failures MUST be logged and tracked separately.

  • All write requests that modify system state MUST be deduplicated using a globally unique business run ID; on timeout, state MUST be queried prior to determining retry eligibility.

Ubuntu ROS 2 Dev PC ​

INFO

Cross-machine ROS 2 communication lacks complete structured re-verification within target customer developer environment baselines. Therefore, this section defines admission criteria only and does NOT guarantee "out-of-the-box connectivity by simply installing Humble".

  1. Dev PC and robot host reside within the same trusted local area network segment, with multicast/unicast discovery and DDS data ports fully reachable;
  2. ROS 2 Humble, RMW implementations, and ROS_DOMAIN_ID settings are fully compatible with on-site configurations;
  3. Dev PC possesses workspace-built rm_robot_interfaces matching the baseline version; blind cross-version file copying is strictly PROHIBITED;
  4. Read-only Topics MUST be used first to sequentially verify node discovery, message types, QoS profiles, and reception frequencies before executing any write operations.
  • Success Criteria: On-host pathway successfully imports rclpy and custom message interfaces; cross-machine pathway continuously receives expected Topics with matching types/QoS, free from misjudgments caused by stale ros2 daemon cache entries.
  • Failure Handling: Use on-robot direct reading as a benchmark control group; if the robot host receives data while the Dev PC does not, troubleshoot in order: Domain ID, RMW, multicast/firewall, and message package compatibility. Revert to on-robot execution or offline MCAP files during troubleshooting.