kinova-gen3

PPO reinforcement learning for a 7-DoF Kinova Gen3 arm in MuJoCo: reaching and obstacle avoidance across three environments.

Finetuning and deploying Octo, a VLA on a Kinova Gen3 (7-DoF) robot for goal-image-conditioned bottle pick-and-lift using diffusion arm actions, a BCE gripper head, and latency-aware receding-horizon control.

A simple and explicit vision guided pick and place using Kinova Gen3, Intel RealSense, Kortex API, and ArUco marker

ROS2 Jazzy and MoveIt2 package for the Kinova Gen3 7DOF robot arm integrated with a Panda gripper. Simulated in Gazebo Harmonic.

Entorno avanzado de robótica colaborativa (ROS 2 Jazzy). Manipulación con Kinova Gen3, robots móviles, AprilTags, MoveIt 2 e integración de modelos IA (Gemini-Robotics).

A lightweight command interface for executing Kinova Gen 3 7-DoF robot arm skills using typed natural-language-style commands, cartesian motion primitives, gripper control, and ArUco-based object pose detection.

Fine-tuned π₀.₅ on 54 own-collected dataset for language-conditioned pick-and-place on a Kinova Gen3 robotic arm. Full pipeline: teleop data collection, LeRobot conversion, LoRA training, and closed-loop deployment with remote inference. A vision-language-action policy that picks up a block and puts it in a bowl.

ROS-based Kinova Gen3 Lite demo for unseen-object grasping using RGB-D perception and calibration

A lightweight wrapper around Kinova Kortex API to work with Gen3 Lite robotic arms.