Kinova GEN3 experiment using Vector Field strategy in SE(3)

API installation

Check Kinova Kortex repository for latest API Release. This will install the 2.8.0 version of the Python API.

For non-linux users other files are available here

curl --output-dir "/tmp" -O https://artifactory.kinovaapps.com/artifactory/generic-local-public/kortex/API/2.8.0/kortex_api-2.8.0.post5-py3-none-any.whl

python -m pip install /tmp/kortex_api-2.8.0.post5-py3-none-any.whl
python -m pip install uaibot

Note: Tested with Python 3.14 and uaibot==1.2.7.

Setup

Initialize Kinova Gen 3 (hold button for ~3s) and connect to the robot via Ethernet cable. Run the following command to configure the robot IP address and port:

Note: you might have to change the interface name. Check with ip link show

chmod +x ./kinova_network_setup.sh
./kinova_network_setup.sh

To reverse the changes, run:

chmod +x ./reset_network.sh
./reset_network.sh

Run basic tests

This will set q0=0 and rotate each joint of the robot by 5 degrees/s and save configuration data. Finally, the robot returns to the default pose.

Log files will be tests_config.npy and tests_time.npy and should contain the configuration values and timestamps of the tests.

python ./kinova_basic_tests.py

If everything works, you should now perform the experiment by:

Run the experiment

The kinematic control uses the Advanced Interface of BaseCyclic which has a sampling rate of 1kHz. The old script (control_40hz.py) runs at 40Hz only, but is safer.

You should first create the curve .npy file using the precompute_curve.py script

The robot will go to the configuration q=[0, 10, 0, 15, 0, 40, 30] and wait for 5 seconds.

python ./control.py

Results

Expected movement can be checked agains the simulation in expected_movement.py script.

Data can be analyzed using the check_experiment_results.py, although it will consider only data used in the previous work.

You can animate the experiment data using the experiment_animation.py script, although you it will probably need heavy modifications.

Kinova Tutorials Playlist