Abstract: Agricultural robot competitions, serving as important scenarios for validating intelligent agricultural machinery technologies, generally faced technical bottlenecks such as high hardware development thresholds and low multi-module collaboration efficiency. Although existing simulation platforms supported high-precision physical simulation, their poor cross-platform compatibility and reliance on local deployment made low-latency remote collaboration difficult to achieve. To address this issue, a lightweight simulation and visualization method based on the ros3djs framework was proposed, which constructed a browser/server（B/S） architecture simulation system and achieved three-dimensional dynamic mapping of the robot's motion state, sensor data streams, and the competition scene. This system utilized the WebSocket protocol to capture multi-source topic data, such as navigation positioning and environmental perception, from the ROS environment in real-time, and accomplished synchronous rendering and interactive control of the robot's pose, operation trajectory, and scene elements. Validation experiments showed that this platform achieved low-latency visual monitoring and remote command transmission of the robot's motion state on the web client, effectively reduced reliance on physical hardware, solved the problem of difficult remote multi-robot interaction in traditional local platforms, and provided an extensible virtualization technology verification solution for agricultural robot competitions.

Key words: ROS3DJS, agricultural robot, competition simulation, WebSocket, visualization technology