Physical AI
AI systems that perceive, reason, and act through physical machines in the real world.
Open topicCurriculum index
Physical AI topics
Browse foundational concepts, core AI methods, robotics systems, tools, resources, and deployment concerns.
Foundations
The mathematics, programming, machine learning, vision, and robotics concepts needed to study Physical AI.
Open topicEmbodied Intelligence
Intelligence that emerges through an agent's body, sensors, actions, and interaction with an environment.
Open topicWorld Models
Learned models that predict environment dynamics and possible future outcomes.
Open topicVLA Models
Models that map visual observations and language instructions to robot actions.
Open topicRobot Learning
Methods that allow robots to acquire behavior from demonstrations, rewards, interaction, or generated experience.
Open topicManipulation
Perception and control for grasping, moving, assembling, and using objects.
Open topicLocomotion
Control methods that enable legged, wheeled, and humanoid robots to move reliably.
Open topicNavigation
Localization, mapping, planning, and control for moving through physical environments.
Open topicSimulation
Virtual environments used to train, test, and evaluate embodied agents before real-world deployment.
Open topicDatasets
Collections of robot trajectories, observations, actions, language instructions, and environment interactions.
Open topicSpatial Intelligence
Models that understand, generate, and reason about persistent 3D spaces rather than isolated images.
Open topicEgocentric Data
First-person video, audio, gaze, hand-object interaction, and wearable sensor data used to teach AI what humans see and do.
Open topicRobot Data Flywheels
The loop of collecting demonstrations, mining failures, annotating trajectories, generating synthetic data, retraining policies, and redeploying robots.
Open topicHumanoid Foundation Models
General-purpose policies and model stacks designed for humanoid bodies, dexterous hands, and whole-body mobile manipulation.
Open topicData Annotation Pipelines
Human-in-the-loop and model-assisted systems that convert raw video, sensor streams, and robot logs into training data.
Open topicBenchmarks
Standard tasks and metrics used to compare embodied AI systems.
Open topicSafety
Methods for keeping embodied systems reliable, controllable, and compatible with human environments.
Open topicNVIDIA Halos
NVIDIA Halos is a full-stack functional-safety system for humanoids, industrial robots, and autonomous mobile robots; Halos OS is its operating-system layer, not a general robot-development OS.
Open topicComputer Vision
Methods that let machines extract useful structure from images and video.
Open topicReinforcement Learning
Learning behavior by maximizing expected reward through interaction.
Open topicSim-to-Real Transfer
Techniques for transferring policies trained in simulation to physical robots.
Open topicDeployment
Engineering reliable Physical AI systems under latency, compute, power, hardware, and operational constraints.
Open topic