Self-Evolving AI Robot System for Lunar Exploration and Human Outpost Construction. | Moon Shot Project Goal 3 Realization of AI robots that autonomously learn, adapt to their environment, evolve in intelligence and act alongside human beings, by 2050.

Integrated control of base construction with multiple robots

Summary

Based on a robot manipulator system for motion verification, we created a situation in which multiple robots with different structures cooperate by restricting the active degrees of freedom of each arm and by adding variations to the end-effectors. A research environment (Sim2Real) was constructed to seamlessly link dynamics simulations and hardware experiments, and the following three items were verified.

Integrated control of base construction with multiple robots

1. To construct a validation model of a multi-limbed modular robot, and to build control models for each of more than 5 different structures/forms.

2. Verify that the AI control system built for the robot task in a specific structure/form can perform the same task with 50% fewer trials by utilizing the learning results already accumulated in the repository.

3. Demonstrated that a robot arm can collect arbitrary sand and gravel samples from the surface and 10 cm below the surface in a simulated sand and gravel site (randomly distributed rocks with grain sizes ranging from 1 mm to 50 cm).

RESEARCH & DEVELOPMENT ITEMS

Design, prototyping, and functional analysis of a self-regenerating modular robot

Functional Analysis Hierarchical Reinforcement Learning for Distributed AI

Integrated control of base construction with multiple robots