FAQ

Issue	Common Cause	Solution
LocoClientInitError	Robot not fully started yet, or timeout too short	Wait for the robot to fully start (typically 30–60 seconds after boot) before connecting; or retry after increasing the timeout parameter, e.g., BoosterRobot(timeout=30)
Non-battery get_xxx() snapshot APIs raise DataNotReadyError	Data did not arrive within timeout; some data channels are unavailable in virtual robot or direct power supply scenarios	First try increasing timeout (e.g., BoosterRobot(timeout=10)) and retry; if it persists, this data channel is unavailable in the current robot configuration, can catch DataNotReadyError in code and handle gracefully
set_joints() raises ValueError	Incorrect number of joints passed	set_joints() only accepts two cases: ① Control all joints (see T1/K1 manual for joint count of each model) ② Control only 10 upper body joints (head 2 + arms 8). Use robot.list_joints() to get complete joint names and count
Robot does not move after set_velocity()	Velocity value too small, or robot is not in motion mode	Check if velocity value is too small (recommended to start from 0.1 and gradually increase)
Head does not move after set_head_angle()	Angle value exceeds joint limits, or value too small	Use robot.get_joint_states() to check current head joint angles; if limits not reached, gradually increase angle parameter and retry
Speech task startup failed	Speech capability unavailable, audio service abnormal, or microphone stream unavailable	① Check whether the audio service has started normally. ② Check whether the microphone and speakers in the current operating environment are working properly. ③ If the issue persists, restart the robot or the simulation environment and try again.
AudioData.save() transcoding failed	ffmpeg not installed when saving non-WAV format	Try again after installing ffmpeg
Cannot import Detection	Vision optional dependencies not installed	Run python3 -m pip install "boosteros[brain]"
Detection model initialization failed	Model path does not exist, optional dependencies missing, ONNX Runtime / CUDA / TensorRT environment unavailable	First use Detection(model="default", device="cpu") to verify basic pipeline, then switch to custom model or GPU backend
Hand-guiding recording/playback function abnormal	Hand-guiding function of dual-board models (e.g., T1) is only available on the motion control board	Deploy hand-guiding related code to the motion control board for execution