Version History

1.Overview

The Voice Interaction Kit SDK Documentation is designed to assist users in rapidly deploying the code for the Voice Interaction Kit after purchase. Additionally, this documentation enables users to perform simulator and real-device deployments of the Voice Interaction Kit on TRON1.

2. Simulator and Real-Device Deployments

💡Prompt
Jetson Orin NX (hereinafter referred to as the development board) comes pre-installed with the Ubuntu 22.04 system, and its IP address is fixed at 10.192.1.3.

2.1 Hardware Environment Setup

Remove the voice interaction kit from its packaging. Plug the power cord into the robot's 24V power output port and the Ethernet cable into the robot's Ethernet port. Power on the robot and wait for the network port light on the Jetson Orin NX to illuminate before proceeding with environment setup.

2.2 Development Board (Jetson Orin NX) Environment Setup

2.2.1 Development Board Environment Setup

When setting up the development environment, you can either directly connect a monitor and Ethernet cable to the development board or use your computer to connect to TRON1 via WiFi SSH for environment setup. We recommend the wireless connection method.

• When using SSH for environment setup, ensure that your computer and the development board are on the same IP subnet.
• Open the local terminal on your computer and enter the following commands:

ssh guest@10.192.1.3 #password：123456
nmcli dev wifi list  #Connect to a network during development board setup
sudo nmcli dev wifi connect "wifiname" password "wifipassword" #Replace wifi name and password with your own
#Upon successful connection, the message Device 'wlan0' successfully activated with "device number" will be displayed”

• You can also download vs code and enter commands and passwords in the Remote Explorer for direct code development.

2.2.2 Python Environment Setup

• For setting up the Python development environment, we recommend using the Python 3.10 environment on Ubuntu 22.04. To facilitate automatic startup, it is recommended to deploy in the native environment. However, during debugging and development, it is recommended to work within a conda environment to avoid disrupting the Ubuntu system. The following steps outline how to deploy on the native environment.

  Enter the following commands in the terminal of your computer to install the required dependent library.

ssh guest@10.192.1.3  #password：123456
sudo su
sudo apt upgrade
sudo apt update
sudo apt dist-upgrade
sudo apt install -y nano cmake curl python3-dev python3-pip libpython3-dev build-essential zlib1g-dev libncurses5-dev libgdbm-dev libnss3-dev libssl-dev libreadline-dev libffi-dev libsqlite3-dev wget libbz2-dev liblzma-dev libncursesw5-dev libreadline6-dev libsqlite3-dev tk-dev uuid-dev python3-dev libncurses5-dev libdb-dev libgdbm-dev libgdbm-compat-dev curl portaudio19-dev libgirepository1.0-dev libopenblas-dev cuda-toolkit nvidia-jetpack libportaudio2 libportaudiocpp0 portaudio19-dev git-lfs tar aptitude sysvinit-utils

Install PiP

mkdir -p /home/guest/tron #Requires sudo environment
cd /home/guest/tron
curl https://bootstrap.pypa.io/get-pip.py -o get-pip.py
sudo python3 get-pip.py
# Upgrade setuptools、packaging and pip
pip install --upgrade setuptools packaging pip wheel #Upgrade setuptools、packaging and pip

Install the Jetson management tool jetpack.

• Jetpack is NVIDIA's all-in-one software bundle containing development tools for NVIDIA Jetson embedded platforms.

ssh guest@10.192.1.3  #password：123456
sudo apt upgrade #Requires sudo environment
sudo apt update
sudo apt dist-upgrade
sudo apt install nvidia-jetpack
sudo -H pip3 install -U jetson-stats

• Clone github repository and install Python dependencies.

git clone https://github.com/limxdynamics/tron1-agent.git
cd tron1-agent #Requires sudo environment
sudo apt-get remove -y python3-blinker
sudo apt-get remove -y python3-sympy
pip install -r requirements.txt -i 

mkdir /home/guest/tron/tron1-agent/tools/pynini && cd /home/guest/tron/tron1-agent/tools/pynini
wget  [https://www.openfst.org/twiki/pub/FST/FstDownload/openfst-1.8.2.tar.gz](https://www.openfst.org/twiki/pub/FST/FstDownload/openfst-1.8.2.tar.gz)  &&  tar -zxvf openfst-1.8.2.tar.gz && cd openfst-1.8.2
./configure --enable-grm --enable-static --enable-shared --host=arm-linux-gnueabihf && make -j$(nproc) && make install && sudo ldconfig
echo 'export CPLUS_INCLUDE_PATH="/usr/local/include:$CPLUS_INCLUDE_PATH"' >> ~/.bashrc
echo 'export LIBRARY_PATH="/usr/local/lib:$LIBRARY_PATH"' >> ~/.bashrc
echo 'export LD_LIBRARY_PATH="/usr/local/lib:$LD_LIBRARY_PATH"' >> ~/.bashrc
source ~/.bashrc
pip install setuptools==62.3.4
pip install cython
pip install pynini==2.1.5 -i --no-build-isolation
pip install WeTextProcessing==1.0.3 --no-deps 

mkdir /home/guest/tron/tron1-agent/tools/torch && cd /home/guest/tron/tron1-agent/tools/torch
# jetpack6 and cuda12.6
wget https://github.com/limxdynamics/tron1-agent/releases/download/v1.0.0/torch-2.8.0-cp310-cp310-linux_aarch64.whl
wget https://github.com/limxdynamics/tron1-agent/releases/download/v1.0.0/torchaudio-2.8.0-cp310-cp310-linux_aarch64.whl
wget https://github.com/limxdynamics/tron1-agent/releases/download/v1.0.0/torchvision-0.23.0-cp310-cp310-linux_aarch64.whl
pip install torch-2.8.0-cp310-cp310-linux_aarch64.whl
pip install torchaudio-2.8.0-cp310-cp310-linux_aarch64.whl
pip install torchvision-0.23.0-cp310-cp310-linux_aarch64.whl

mkdir /home/guest/tron/tron1-agent/tools/spacy && cd /home/guest/tron/tron1-agent/tools/spacy
wget https://github.com/limxdynamics/tron1-agent/releases/download/v1.0.0/en_core_web_trf-3.8.0-py3-none-any.whl
wget https://github.com/limxdynamics/tron1-agent/releases/download/v1.0.0/zh_core_web_trf-3.8.0-py3-none-any.whl
pip install en_core_web_trf-3.8.0-py3-none-any.whl
pip install zh_core_web_trf-3.8.0-py3-none-any.whl -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install scipy --upgrade -i https://pypi.tuna.tsinghua.edu.cn/simple

• Model download (Only default model configurations are provided. Explore others as needed.)

cd /home/guest/tron/tron1-agent/models
pip install modelscope
modelscope download --model iic/SenseVoiceSmall --local_dir SenseVoiceSmall
modelscope download --model BAAI/bge-small-zh --local_dir rag/bge-small-zh
modelscope download --model BAAI/bge-small-en --local_dir rag/bge-small-en
modelscope download --model iic/speech_fsmn_vad_zh-cn-16k-common-pytorch --local_dir fsmn_vad
git lfs install
sudo apt-get install git-lfs
mkdir -p translation/Helsinki-NLP
cd translation/Helsinki-NLP
git clone [https://huggingface.co/Helsinki-NLP/opus-mt-zh-en](https://huggingface.co/Helsinki-NLP/opus-mt-zh-en)
git clone [https://huggingface.co/Helsinki-NLP/opus-mt-en-zh](https://huggingface.co/Helsinki-NLP/opus-mt-en-zh)
cd ../.. && mkdir -p tts/piper && cd tts/piper #Return to the models directory
# Model repository locations https://huggingface.co/rhasspy/piper-voices/tree/v1.0.0
wget https://huggingface.co/rhasspy/piper-voices/blob/v1.0.0/en/en_US/amy/medium/en_US-amy-medium.onnx
wget https://huggingface.co/rhasspy/piper-voices/blob/v1.0.0/en/en_US/amy/medium/en_US-amy-medium.onnx.json
wget https://huggingface.co/rhasspy/piper-voices/blob/v1.0.0/zh/zh_CN/huayan/medium/zh_CN-huayan-medium.onnx
wget https://huggingface.co/rhasspy/piper-voices/blob/v1.0.0/zh/zh_CN/huayan/medium/zh_CN-huayan-medium.onnx.json

• To complete the tron1-agent environment setup, download your selected model weights according to the configuration file.

2.2.3 Download Large Models

• Open a new terminal and execute the following commands to download large models.

ssh guest@10.192.1.3  #password：123456
sudo su
cd /home/guest/tron 
curl -fsSL https://ollama.com/install.sh | sh 
# Select an appropriate model, such as qanwen3:1.7b
ollama run modelscope.cn/unsloth/Qwen3-1.7B-GGUF

2.2.4 Sound Pickup Module Configuration

• Execute the following commands to configure the module.

ssh guest@10.192.1.3  #password：123456
sudo su
nano /etc/asound.conf

• Use # to comment out the contents in the red box.

• Check whether the sound pickup module has been connected to the development board. If connected, you can directly run the following program.

ssh guest@10.192.1.3  #password：123456 
ll /dev    # After the commands are executed, you can observe the ttyACM0 appearing

• Once ttyACM0 is detected, you can configure rules for the pickup module.

cd /home/guest/tron/tron1-agent/tools/iflytek
sudo chmod +x iflytek.sh
sudo bash iflytek.sh

• After configuring the above commands, disconnect the power supply of the pickup module for 3s, and then reconnect it.

ssh guest@10.192.1.3  #password：123456 
ll /dev    # After the commands are executed, you can observe the wheeltec_mic sppearing, indicating that the rule has been configured.

• After the configuration, run the following program.

sudo apt update
sudo apt install libcjson-dev

• Open /etc/ld.so.conf and append the following content at the end of the file.

sudo nano /etc/ld.so.conf
/usr/local/lib #Add the content

• Save (Ctrl+X) and exit, then execute.

sudo /sbin/ldconfig

• Check whether the audio device is configured successfully.

cd /home/guest/tron/tron1-agent/tools/iflytek
python check.py

• Configure the wake-up word.

cd /home/guest/tron/tron1-agent/tools/iflytek
python change_wakeup.py --keyword "hey tron" --threshold 500
sudo reboot

• When the wake-up word configuration is complete, unplug the sound pickup module and disconnect its power supply for 3s. Then, reboot the system and proceed to the next step.

2.3 Environment Detection

• Run the following program in the terminal.

ssh guest@10.192.1.3  #password：123456 
sudo su   # Enter your password
cd /home/guest/tron/tron1-agent
python main.py

• If you hear the playback "Hello, I'm TRON!". Then the program is deployed successfully.

2.4 Simulation Environment Setup

💡Prompt
If the voice interaction kit has been installed on a real machine, there is no need to run the simulation environment. You can directly run the auto-start program.

• Download and extract.

• tron_simulator.tar.gz

• Configure the operating environment

  • You can configure the operating environment with the following command:

# Install UV Tools
sudo apt install curl libwebsockets-dev
curl -LsSf https://astral.sh/uv/install.sh | sh && bash

# Navigate to the extracted directory
cd tron_simulator

# Create a Python virtual environment
uv venv .venv

# Install Python 3.8
source .venv/bin/activate
uv python install 3.8
uv venv --python 3.8

# Install limxsdk
source .venv/bin/activate
uv pip install limxsdk-3.0.17-py3-none-any.whl

• Run Tron Simulator

cd tron_simulator
sudo bash start.sh

2.5 Auto-Start on Boot

• Enter the following command.

chmod +x /home/guest/tron/tron1-agent/start.sh
sudo cp /home/guest/tron/tron1-agent/tools/auto_start/tronagent.service /etc/systemd/system/tronagent.servce
sudo systemctl enable tronagent.service
sudo systemctl daemon-reload

• Close the terminal, restart the robot, and wait for about 1 min. If the device outputs "Hello, I am TRON", the auto-start has been successful.