Our Rally AI platform combines robotics hardware, computer vision and reinforcement learning to create a training partner that never gets tired. In this post we explore some of the design decisions behind the system.
Hardware Design
Rally AI features a mobile base capable of moving in any direction, and custom arms capable of replicating a wide variety of shots. Depth sensors and high‑speed cameras track the trajectory of balls and adjust the robot’s positioning in real time.
Software & AI
Under the hood, we run physics simulations to predict ball trajectories and use reinforcement learning to adapt difficulty on the fly. The system logs your performance and uses it to personalise your next session.
We’ll be sharing more technical articles soon, including code snippets and tutorials on how to extend the platform. If you have questions or want to collaborate, feel free to contact us.