Walker Tienkung Operation Guide

This document describes the basic flow for using Thinker studio v2.0.x with Walker Tienkung for Pico teleoperation, motion recording, data collection, quality inspection, and export.

For general installation configuration, service management, and Pico-side operations, please refer to:

• Documentation Index and Introduction
• Pico Operation Guide

Applicable Environment​

Network Configuration​

After connecting the target machine and Walker Tienkung with an Ethernet cable, ensure both sides are on the same network segment.

Walker Tienkung factory network segment:

192.168.41.x

Configure the target machine's physical Ethernet port to a static IP in the same network segment, for example:

IP Address: 192.168.41.100
Subnet Mask: 255.255.255.0

Use the following command to check whether the target machine can access the Orin board:

ping 192.168.41.2

Ubuntu wired network configuration entry

Ubuntu IPv4 manual configuration example

Pre-use Preparation​

Please confirm:

• The target machine has Thinker studio installed.
• The target machine has completed network configuration and can access 192.168.41.2.
• Walker Tienkung is powered on.
• Pico, body trackers, and remote controller are ready.

Select Robot​

Find the Thinker studio icon in the application center on the target machine and click it to open the client.

On the robot selection page, select "Walker Tienkung" to enter the teleoperation page.

Thinker studio client icon

Start Pico Teleoperation​

Start Teleop Service​

Start the Teleop Service service in the Thinker studio client.

Start Teleop Service

This service is used to receive tracking and control data sent from the Pico side. After the service starts, keep the client page open.

Teleop Service running

Pico-side Operations​

For tracker wearing, XRoboToolkit connection, checking Send, and other Pico-side operations, please refer to the Pico Operation Guide.

Start Retargeting Monitor​

Start the Retargeting Monitor service in the Thinker studio client.

Start Retargeting Monitor

This service is used for motion retargeting monitoring during teleoperation.

Retargeting Monitor running

Enter Robot Upper Limb Control Mode​

Before starting teleoperation, you need to put Walker Tienkung into upper limb control mode, also known as half-body control mode.

Please operate in the following order:

1. Confirm that Walker Tienkung is powered on. Open the main power switch on the left side of the robot's back, then short-press the power button on the right side.
2. Turn on the remote controller. Short-press the remote controller switch, then long-press the remote controller switch.
3. Press A on the remote controller to make the robot perform self-check.
4. After self-check is complete and a "beep" is heard, wait for 5 seconds.
5. Toggle the F switch down.
6. Long-press A on the remote controller until a "beep" is heard.
7. Toggle the E switch up.

After completing the above steps, click Start in the teleoperation area of the Thinker studio client.

Click to start teleoperation

Safety Reminder

After clicking Start, the robot will begin to follow movements. Please confirm that the robot's surrounding environment is safe and spacious, and avoid people or objects entering the robot's motion range before clicking.

Teleoperation started successfully

End Teleoperation​

When ending teleoperation, please operate in the following order:

1. Click Stop in the teleoperation area of the Thinker studio client.
2. Stop related services, such as Teleop Service and Retargeting Monitor.
3. If you need to pause Pico control data sending, you can uncheck Send under Data & Control in XRoboToolkit on the Pico side.

Stop teleoperation

Stop related services

Motion Recording​

Start Recording​

Motion recording requires the following prerequisites:

• Currently in teleoperation mode.
• Data collection is not currently started.

Click Start in the motion recording area of the Thinker studio client to begin motion recording.

Start motion recording

During recording, the Send status of XRoboToolkit on the Pico side affects whether the robot follows the operator's movements:

• Check Send: Pico sends control data, and the robot follows the movements.
• Uncheck Send: Pico does not send control data, and the robot does not follow the movements.

XRoboToolkit Send status

End Recording​

Click End in the motion recording area to complete the recording.

After the recording is successful, the recorded motion will be displayed in the recording history area on the right.

Motion recording history

Play Recorded Motion​

Motions in the recording history can be played back.

Play recorded motion

Safety Reminder

Before playing a recorded motion, please confirm that there is enough space around the robot to avoid collisions with people, equipment, or surrounding objects.

Data Collection​

Data collection is used to collect camera data and arm joint angles.

Before starting data collection, you must first start the head camera service on the Orin board of Walker Tienkung at 192.168.41.2. Otherwise, data collection will fail or no camera data will be collected.

Start Head Camera Service​

SSH into the Orin board from the target machine:

ssh nvidia@192.168.41.2

Password:

nvidia

Start the head camera service:

sudo systemctl start orbbec_head.service

Confirm that the camera compressed topic exists and has data:

ros2 topic echo /ob_camera_head/color/image_raw/compressed --no-arr

If the command continuously outputs data, the compressed image topic is normal. After confirming, you can press Ctrl+C to exit the view.

Start Collection​

Data collection can only be started when no motion is being recorded.

Safety Reminder

Before starting data collection, please confirm that the robot's surrounding environment is safe and spacious, and avoid people or objects entering the robot's motion range.

Click Start in the data collection area of the Thinker studio client to begin collection.

Start data collection

Data collection started successfully

End Collection​

Click End in the data collection area to stop collection.

After successful collection, the collected data will be displayed in the collection history area on the right.

Collection history

Data Quality Inspection and Export​

Data Quality Inspection​

Select the data to be processed in the collection history and click Quality Check to enter the data quality inspection page.

Enter quality inspection

Quality inspection page

The quality inspection page supports adding multiple region frames and adding descriptions for the region frames. Click Submit when finished.

Submit quality inspection description

Data Export​

Data that has completed quality inspection will display an Export button.

Export button

After clicking Export, select the desired export format. The larger the data volume, the longer the export time.

Select export format

After successful export, refresh the page. The page will display the exported data directory, and you can click to copy the directory path.

Exported data directory

FAQs​

Data collection fails to start​

Please check the following items first:

• Whether Walker Tienkung is powered on.
• Whether the target machine can access 192.168.41.2.
• Whether orbbec_head.service on the Orin board is started.
• Whether the /ob_camera_head/color/image_raw/compressed topic exists and continuously outputs data.

Appendix: Manual Installation of Camera Compressed Topic Components​

This section is mainly for R&D or operations personnel. Under normal circumstances, the installation script will automatically handle the camera compressed topic components when the target machine and Walker Tienkung are directly connected via Ethernet and Walker Tienkung is powered on.

Start Camera Service and Check Topic​

Log in to the Orin board:

ssh nvidia@192.168.41.2

Password:

nvidia

Start the head camera service:

sudo systemctl start orbbec_head.service

Check whether the following compressed image topic exists:

ros2 topic echo /ob_camera_head/color/image_raw/compressed --no-arr

If the topic already exists and has data, no further installation is needed.

Install Dependencies on the Orin Board​

export ROS_DISTRO=humble

sudo apt install libgflags-dev nlohmann-json3-dev \
  ros-$ROS_DISTRO-image-transport \
  ros-$ROS_DISTRO-image-transport-plugins \
  ros-$ROS_DISTRO-compressed-image-transport \
  ros-$ROS_DISTRO-image-publisher \
  ros-$ROS_DISTRO-camera-info-manager

Copy Compressed Component Files​

Please first download the following two files to the /home/ubuntu/Downloads directory on the target machine:

• rvl_codec.hpp
• libcompressed_depth_image_transport.so

Execute on the target machine:

scp /home/ubuntu/Downloads/rvl_codec.hpp nvidia@192.168.41.2:/home/nvidia/
scp /home/ubuntu/Downloads/libcompressed_depth_image_transport.so nvidia@192.168.41.2:/home/nvidia/

Install to Orin Board ROS Directory​

Log in to the Orin board again:

ssh nvidia@192.168.41.2

Execute:

sudo cp /home/nvidia/rvl_codec.hpp /opt/ros/humble/include/compressed_depth_image_transport/
sudo cp /home/nvidia/libcompressed_depth_image_transport.so /opt/ros/humble/lib/

After installation, it is recommended to power off and restart the robot.

After restarting, log in to the Orin board again and start the head camera service:

sudo systemctl start orbbec_head.service

Confirm that the compressed image topic exists and has data:

ros2 topic echo /ob_camera_head/color/image_raw/compressed --no-arr