Time to Play: Simulating Early-Life Animal Dynamics Enhances Robotics Locomotion Discovery

2025-09-15

AI Summary

Conventional robot training relies on static physical parameters, overlooking the critical role of dynamic power-to-mass ratio changes—observed during biological development—in shaping locomotor evolution. Method: We propose SMOL, a curriculum learning framework that, for the first time, incorporates ontogenetic-scale actuator mechanical output variations (e.g., torque modulation induced by growth or aging) into robotic reinforcement learning, explicitly modeling physiological stages such as puberty. SMOL integrates with the MAP-Elites quality-diversity framework to dynamically modulate actuator torque in standard locomotion tasks and extends to a human-inspired variant, SMOL-Human. Contribution/Results: Experiments demonstrate substantial improvements in locomotion performance and behavioral diversity. Crucially, policies learned under SMOL transfer efficiently to robots with fixed morphologies, establishing a novel developmental paradigm for embodied intelligence training.

Technology Category

• Intelligent Robots: Learning & Optimization for ROB
• Machine Learning: Evolutionary Learning
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Developmental changes in body morphology profoundly shape locomotion in animals, yet artificial agents and robots are typically trained under static physical parameters. Inspired by ontogenetic scaling of muscle power in biology, we propose Scaling Mechanical Output over Lifetime (SMOL), a novel curriculum that dynamically modulates robot actuator strength to mimic natural variations in power-to-weight ratio during growth and ageing. Integrating SMOL into the MAP-Elites quality-diversity framework, we vary the torque in standard robotics tasks to mimic the evolution of strength in animals as they grow up and as their body changes. Through comprehensive empirical evaluation, we show that the SMOL schedule consistently elevates both performance and diversity of locomotion behaviours across varied control scenarios, by allowing agents to leverage advantageous physics early on to discover skills that act as stepping stones when they reach their final standard body properties. Based on studies of the total power output in humans, we also implement the SMOL-Human schedule that models isometric body variations due to non-linear changes like puberty, and study its impact on robotics locomotion.

Problem

• Simulating animal growth dynamics improves robot locomotion discovery
• Dynamically modulating actuator strength mimics biological power-to-weight changes
• Enhancing performance and diversity through developmental-inspired torque variation

Innovation

• Dynamic actuator strength modulation for robots
• Integration with MAP-Elites quality-diversity framework
• SMOL schedule mimicking biological growth patterns