Robotics glossary: terms every robot buyer should know

Plain-language explanations of robotics terms: DOF, AMR vs AGV, SLAM, LiDAR, payload, RaaS, VLA models, teleoperation and more, with why they matter when you choose a robot.

Actuator: The unit that creates movement in a joint, usually a motor combined with a gearbox.
AGV (automated guided vehicle): A transport vehicle that follows a fixed route marked by tape, magnets, or QR codes.
AMR (autonomous mobile robot): A mobile robot that plans its own route and drives around obstacles.
AS/RS (automated storage and retrieval system): An automated system that stores and retrieves goods from racks, often in tall dense storage.
Autonomy level: How much a robot does on its own, from full manual control to full independence.
Battery runtime and hot swap: How long a robot works per charge, and whether the battery can be replaced without shutting it down.
Bipedal locomotion: Balancing and walking on two legs, the core ability of a humanoid robot.
Cobot (collaborative robot): A robot arm designed to work next to people safely, usually without a safety fence.
Dexterous hand: A multi-fingered robot hand with many degrees of freedom for fine manipulation.
Digital twin: A virtual copy of a robot or a facility used for testing before touching the real thing.
DOF (degrees of freedom): The number of independent movements a joint or a whole robot can make.
Embodied AI: AI that acts in the physical world through the body of a robot, not only on a screen.
Emergency stop (E-stop): A physical or wireless control that immediately cuts power to the robot's motion.
End effector: The tool at the end of a robot arm, such as a gripper or a suction cup.
Exoskeleton: A wearable robotic frame that supports or amplifies a person's movements.
Fleet management: Software that coordinates many robots: tasks, traffic, charging and status in one place.
Foundation model for robotics: A large general-purpose model that can be adapted to many robots and tasks.
Goods-to-person (GTP): A warehouse setup where robots bring the goods to a worker instead of the worker walking to the shelf.
Gripper: The most common end effector: a device that grips, pinches or holds objects.
Harmonic drive: A compact, precise gearbox with almost no backlash, common in robot joints.
Imitation learning: A robot learns a task by copying demonstrations given by a human.
IMU (inertial measurement unit): A sensor with accelerometers and gyroscopes that measures tilt and acceleration.
IP rating: A code such as IP54 that states how well a device is protected against dust and water.
ISO 10218 / ISO 13482: International safety standards for industrial robots (10218) and personal care robots (13482).
LiDAR: A laser rangefinder that gives the robot a 2D or 3D cloud of points describing its surroundings.
Navigation stack: The set of software modules for map, localization and path planning that lets a robot move.
Palletizing: Automatically stacking boxes or bags onto a pallet in a stable pattern.
Path planning: Calculating a safe route from the current position to the goal.
Payload: The maximum weight a robot can carry or lift while working normally.
Pick and place: Picking up an object and putting it down in a defined spot.
Quadruped robot: A robot that walks on four legs, usually called a robot dog.
RaaS (Robots as a Service): Paying a monthly fee to use a robot instead of buying it, often with maintenance included.
Reinforcement learning: Learning by trial and error, where good results are rewarded and bad ones are penalized.
ROS / ROS 2: An open-source framework of tools and libraries for programming robots.
Sim-to-real: Transferring skills a robot learned in simulation to the real machine.
SLAM: A method by which a robot builds a map and works out its own position on it at the same time.
Tactile sensor: A sensor that detects touch and pressure on a robot finger or skin.
TCO and ROI: The full cost of owning a robot over its life, and how long it takes to earn that money back.
Teleoperation: A person controls the robot remotely, often with a VR headset or motion-capture suit.
ToF camera (time of flight): A depth camera that measures distance by how long emitted light takes to return.
Torque density: How much turning force an actuator delivers for each kilogram of its own weight.
VLA model (vision-language-action): A neural network that takes camera images and a text instruction and outputs robot actions.
Whole-body control: A control approach that coordinates all joints of a robot together as one system.
WMS integration: Connecting a robot to the warehouse management system so it knows what to move and where.
World model: An AI model that predicts how the physical world will change after an action.