Inspire Hand SDK Setup Guide (LIDAR Lab @ GT)

This guide describes the setup and usage of the Inspire Hand SDK for the LIDAR Lab at Georgia Tech.

---

Inspire Hand Details

Virtual Environment Management (Conda)

We use conda to manage the Python virtual environment.

1. Create and Activate Environment

conda create -n inspire_hand python=3.8
conda activate inspire_hand

---

Installation

1. Clone Repository (if not already done)

git clone <repo_url>
cd InspireHand-Tactile-Sensing

2. Initialize and Update Submodules

git submodule init
git submodule update

3. Install SDKs (Editable Mode)

⚠️ We do NOT use pip install -r requirements.txt.

Instead, install the two SDKs in editable mode:

cd unitree_sdk2_python
pip install -e .

cd ../inspire_hand_sdk
pip install -e .

After installation, return to the root directory:

cd ..

---

Configuring Wired Network Connection for Inspire Hands (One-time setup in Ubuntu)

To communicate with the Inspire Hands via Ethernet/USB-C:

1. Open Wired Network Settings

   • Click the network icon in the top-right corner of Ubuntu
   • Select Wired Settings

2. Create a New Profile

   • Click the gear icon next to your wired interface
   • Select IPv4 settings
   • Set Method to Manual
   • Assign a static IP in the same subnet as the Inspire Hands, e.g.:
     • Example for PC: 192.168.123.100
     • Subnet Mask: 255.255.255.0
     • Gateway: leave blank

3. Connect to the Hands

   • The default IPs are:
     • Left hand: 192.168.123.210
     • Right hand: 192.168.123.211
   • Save and apply the settings
   • Open a terminal and test connectivity:

     ping 192.168.123.210
     ping 192.168.123.211

   • If replies are received, communication is normal

Note: Ensure your PC is directly connected to the hand via Ethernet or USB-C. Both the driver device and the hands must be on the same subnet to communicate properly.

Running the Inspire Hands (LIDAR Lab Workflow)

All runtime scripts are located in:

/home/lidar/InspireHand-Tactile-Sensing/inspire_hand_sdk/example

⚠️ Important: Multi-Terminal Setup Required

You must run the driver in one terminal, and all visualization or other scripts in separate terminals.

---

Step 1: Start the Headless Driver (Required)

In Terminal 1, activate conda and run:

conda activate inspire_hand
python inspire_hand_sdk/example/Headless_driver_double.py

This script:

• Initializes the hands
• Starts DDS communication
• Publishes and subscribes to required topics

This must remain running before launching any other scripts.

---

Step 2: Run Visualization or Other Scripts

In Terminal 2 (or more terminals):

conda activate inspire_hand
python inspire_hand_sdk/example/plot_force_hand_map_updated.py

This script:

• Subscribes to tactile sensor data
• Displays a GUI
• Plots force values on the tactile sensor map

A new set of scripts has been added for tactile data collection and robot hand pose toggling. These scripts are located in:

/home/lidar/InspireHand-Tactile-Sensing/data_collection_scripts

These scripts are designed to run while the Headless_driver_double.py is active (multi-terminal setup still applies).

Control Modes

The Inspire Hand SDK supports multiple control modes:

• Mode 0: 0000 (No operation)
• Mode 1: 0001 (Angle)
• Mode 2: 0010 (Position)
• Mode 3: 0011 (Angle + Position)
• Mode 4: 0100 (Force control)
• Mode 5: 0101 (Angle + Force control)
• Mode 6: 0110 (Position + Force control)
• Mode 7: 0111 (Angle + Position + Force control)
• Mode 8: 1000 (Velocity)
• Mode 9: 1001 (Angle + Velocity)
• Mode 10: 1010 (Position + Velocity)
• Mode 11: 1011 (Angle + Position + Velocity)
• Mode 12: 1100 (Force control + Velocity)
• Mode 13: 1101 (Angle + Force control + Velocity)
• Mode 14: 1110 (Position + Force control + Velocity)
• Mode 15: 1111 (Angle + Position + Force control + Velocity)

---

Summary (LIDAR Lab Standard Workflow)

1. Activate conda environment
2. Start Headless_driver_double.py (Terminal 1)
3. Run GUI or control scripts in separate terminals