SPOT REINFORCEMENT LEARNING
RESEARCHER KIT
PUSH THE BOUNDARIES OF SPOT’S MOBILITY
The Spot Reinforcement Learning (RL) Researcher Kit from Boston Dynamics turns the quadruped robot Spot into a powerful platform for AI and robotics research. With low-level motor control access, an onboard NVIDIA Jetson computer, and a realistic simulation environment (Isaac Lab), researchers can design, train, and test new locomotion strategies.
This kit empowers breakthroughs in robotics, enabling the creation of agile, robust, and adaptive behaviors that go far beyond Spot’s out-of-the-box capabilities. By bridging simulation and real-world deployment, it accelerates innovation in fields such as advanced locomotion, autonomy, and human–robot interaction.
WHAT’S INCLUDED
Spot Academia & Manipulation Kit
The complete Spot robot package for academic, research and manipulation applications, including the articulated arm and the Spot Core I/O processor.
Spot Researcher API License
Spot’s joint-level API makes it possible to use reinforcement learning (RL) to train unique behaviors and modes of locomotion. Several tools are available from Boston Dynamics and partners to support researchers and developers of RL applications.
- Built and maintained by Boston Dynamics, the jointlevel API expands on their publicly available developer tools to broaden research opportunities—from faster gaits to more complex movements
- The RL Researcher API is compatible with the NVIDIA Isaac Lab simulation environment, where a dynamic model of Spot (including masses, inertias, and actuator models) and a mimic of the API are available for download.
Developer Hardware
- NVIDIA AGX Orin Developer Kit*
- Mount, brackets, and cables to attach the NVIDIA AGX Orin Developer Kit—recommended for optimal performance—to Spot.
*The AGX Orin is not required for deploying the Researcher API, however, it is recommended for optimal computing speed and latency.
