AI Robots Trained to Replicate Complex Athletic Movements

Carnegie Mellon University and NVIDIA have pioneered a breakthrough training method that empowers humanoid robots to perform intricate athletic feats with remarkable agility—from Cristiano Ronaldo’s mid-air spin to Kobe Bryant’s iconic fadeaway.

Their framework, Aligning Simulation and Real Physics (ASAP), bridges the gap between virtual training and real-world execution, enabling robots to master movements once deemed too complex for machines.

The researchers noted in their paper:

"Humanoid robots hold the potential for unparalleled versatility for performing human-like, whole-body skills. However, achieving agile and coordinated whole-body motions remains a significant challenge due to the dynamics mismatch between simulation and the real world.”

Complicated Algorithm Involved in Training AI Robots

ASAP addresses the challenge of bridging simulation and real-world physics through a two-stage process.

First, it pre-trains motion tracking policies using human motion data in simulation.

Then, these policies are deployed in the real world, collecting data to refine movement accuracy.

The result is a humanoid robot capable of replicating signature athletic feats—from Ronaldo’s 180-degree mid-air "Siu" celebration to LeBron’s precise single-leg "Silencer" and Kobe Bryant’s fadeaway jump shot.

Beyond sports moves, the robots demonstrated impressive agility, executing forward and lateral jumps exceeding one meter.

While their movements may still appear clumsy due to hardware limitations, they possess greater dexterity than previous humanoids, thanks to the "delta action model"—a correction mechanism that compensates for discrepancies between simulated and real-world physics.

This innovation reduced tracking errors by up to 52.7%, significantly enhancing agility and whole-body coordination.

The researchers emphasized its potential to pave the way for more versatile humanoid robots in real-world applications:

"Our approach significantly improves agility and whole-body coordination across various dynamic motions."

Fluid Agility is a Constant Issue for Robots

Achieving human-like dexterity in robots has long been a major challenge in robotics.

The researchers wrote:

"For decades, we have envisioned humanoid robots achieving or even surpassing human-level agility. However, most prior work has primarily focused on locomotion, treating the legs as a means of mobility."

ASAP tackles this by pre-training movement policies in simulation, allowing robots to adapt seamlessly to real-world conditions.

By mimicking human biomechanics, their extremities function dynamically—supporting movement, balance, counterweighting, and even expression.

The research team continues refining ASAP, addressing limitations such as hardware strain from executing complex maneuvers.

Noting that some models broke whilst executing complex movements, they said:

“Future directions could focus on developing damage-aware policy architectures to mitigate hardware risks.”

Future efforts will explore markerless pose estimation and onboard sensor fusion to reduce reliance on motion capture systems, enhancing adaptability and efficiency.

With these advancements, how long before robots compete—and dominate—at the Olympics?