← Catalog of 100 simulationsAI TRAINING LAB · THREE.JS

NAV-01

Adaptive Maze

Navigation · autonomous training environment

NAVIGATION

3D ENVIRONMENTEPISODE 001

POLICY PPOSEED 0000

STEP 0WEBGL ACTIVE

EPISODE PROGRESS0 / 600

TRAINING TELEMETRY

Reward0.00reward

Success Rate0%goal rate

Loss1.000policy loss

Exploration100%entropy

REWARD CURVE

CONTROLS

Speed1.0×Agents24

TASK · DIFFICULTY 5/10

+1.0 goal−0.3 collision

← Back Next →