Robotics Glossary

We’ve compiled a list of commonly used robotics terms and tried our best to explain them. If you don’t see what you’re looking for, email us and we’ll do our best to add an explanation.

Active movement: A movement which is generated by a person wearing a robot (e.g. exoskeleton) which is generated by the user, not by the robot.

Actuator: A type of motor that produces movement in a robot.

Algorithm: Essentially a mathematical formula that is used in a computer programme to dictate how a robot behaves.

Autonomous: Able to function without direct human input.

Brain-Computer Interface (BCI): A method to enable a robot to understand what its driver wants without them having to move but by measuring brain activity. This can be achieved by surgically implanting something in the nervous system (invasive: e.g. to allow a prosthetic arm to be controlled by the brain) or by placing electrodes on the skin to pick up signals (non-invasive: e.g. wearing a cap to control a computer game).

Compliance: A compliant robot is somehow “soft” and therefore capable of passively adapting to external disturbances (e.g. an unforeseen obstacle during locomotion) and to safely interact with humans.

Impedance control: A way of robots and humans interacting with each other but taking position, velocity and acceleration into account – within the NCCR impedance control is particularly useful when building robots for physiotherapists as this means that a patient makes a movement without the robot stopping them but if the patient stumbles the robot will catch them as the velocity will increase sharply prompting action from the robot.

Inertial Measurement Unit (IMU): A device consisting of accelerometers and gyroscopes commonly used to measure accelerations and angular rate (amount of spin) on a robot.

Kalman filter: An algorithm used for navigation and route planning in robots. Kalman filters differ from other methods of control as they update over time.

Model Predictive Control (MPC): A type of model (mathematical equation) that represents behaviour of a robot in a system with many obstacles which may be shifting position. MPCs predict changes in behaviour or a robot when an unpredicted input (e.g. someone crossing the path of a robot) occurs.

Orthosis: A device worn outside the body (e.g. exoskeleton) to help someone to move after they have lost or reduced mobility e.g. after spinal cord injury or stroke. Orthoses are often used in physiotherapy to help patients to rebuild muscle function.

Passive Movement: A movement from a wearable robot (e.g. exoskeleton) which is generated by the robot, not by the person wearing it.

Prosthesis: Worn by somebody with an amputation to replace the look and/or function of the missing limb.

Sensor: A sensor is an electro-mechanical device that translates stimuli to the robot into a meaningful information that the robot can process. For example a camera is the artificial equivalent of human eye. Many sensors are produce low quality images, so a robot may not be able to see what exactly is in front of it, just an outline of something that it knows it must avoid when moving.

State-estimation: How a robot knows what kind of environment it is in at any given moment, e.g. if a robot is walking on an uneven surface, it must estimate where it should put its foot to prevent it from falling over.

Torque-controlled: Essentially a method of achieving a desired force/torque by controlling the movements of an actuator.

Quadcopter/Quadrotor: A flying robot that uses four sets of propellors to lift its self off the ground, much like a helicopter.

More Resources