Autonomous Robots

Autonomous Robots
Robotic systems capable of performing intended tasks with varying degrees of autonomy
Field	Robotics
Principal functions	Sensing, state estimation, planning and control
Autonomy	Task-dependent
Human involvement	Varies by system and task
Operating environment	Application-dependent
Applications	Manufacturing, logistics, services and exploration
Related disciplines	Control engineering, artificial intelligence, computer vision

Autonomous robots are robots capable of performing intended tasks with a degree of independence from direct human intervention. They combine mechanical systems, sensing, computation and control to act within a defined operating environment. Depending on their design, autonomous robots may navigate through an environment, manipulate objects, inspect equipment or carry out other physical tasks.

The International Organization for Standardization (ISO), in ISO 8373:2021, defines terminology for robots and their autonomy. The standard characterizes autonomy in relation to carrying out intended tasks using the system's current state and sensing without human intervention. The degree of autonomy can vary according to the application and the system's independence in making decisions.

Autonomy does not necessarily mean that a robot can operate in every environment or perform tasks outside its designed capabilities. A robot may operate independently during a particular activity while still requiring human assistance for configuration, supervision, maintenance or exceptional situations. Autonomous operation should therefore be understood in relation to a specified task and operating environment.

Definition and scope

Robotic autonomy concerns the ability of a robot to use available information to determine or adjust its actions while carrying out assigned objectives. The extent of autonomy depends on the robot's sensors, control architecture, computational capabilities and operational constraints.

Autonomous robots may be mobile or stationary. Mobile robots move through an environment, while robotic manipulators can perform physical operations using articulated mechanisms. Autonomy may be incorporated into either type of system, although the required perception, planning and control functions differ between applications.

Autonomy and automation

Automation and autonomy are related but distinct concepts.

An automated machine may execute a predefined sequence of operations under specified conditions. Autonomous operation additionally involves making or adjusting operational decisions using information about the system and its environment. However, autonomous systems may also use predefined rules, conventional control algorithms and programmed procedures.

The presence of autonomy does not, by itself, establish that a robot possesses general artificial intelligence. The system's demonstrated abilities must be evaluated in relation to its intended tasks and operating conditions.

Degrees of autonomy

Robotic systems can differ substantially in the amount of human involvement required. Human operators may establish goals, supervise execution, approve actions, intervene during exceptional situations or directly control particular functions.

A robot may perform one task independently while relying on human control for another. Consequently, describing a system as autonomous does not establish that every function operates without human intervention.

Fundamental components and operation

Autonomous robotic systems integrate several technical disciplines, including mechanical engineering, electrical engineering, computer science, control theory and artificial intelligence. Their architectures vary according to application, but commonly involve sensing, state estimation, planning and action execution.

Sensing and perception

Sensors provide measurements of the robot's condition and surrounding environment.

Depending on the application, autonomous robots may use cameras, laser rangefinders, inertial measurement units, wheel encoders and other sensing devices.

Perception systems process sensor measurements to extract information relevant to robot operation, such as obstacles, surfaces, objects or environmental features.

Sensor measurements are affected by uncertainty. Their interpretation may require estimation methods that account for measurement errors, incomplete observations and changing conditions.

Localization and mapping

Localization is the process of estimating a robot's position and orientation relative to an environment or coordinate system.

Mobile robots may localize using information obtained from sensors, previously created maps and estimates of their own motion. Localization is an important component of navigation because planning and executing movement generally require an estimate of the robot's current pose.

Simultaneous localization and mapping (SLAM) is a class of methods in which a robot estimates its location while constructing or updating a representation of its environment.

SLAM has been an important research topic in autonomous mobile robotics, particularly where a sufficiently detailed map is not available in advance.

Planning and decision-making

Planning determines actions or sequences of actions intended to achieve specified objectives.

In mobile robotics, motion planning concerns the selection of a path or trajectory between locations while accounting for the robot's movement constraints and the surrounding environment.

Planning systems may incorporate geometric representations, search algorithms, probabilistic methods and other computational techniques. The appropriate approach depends on the task, the available information and the uncertainty involved.

Autonomous navigation may also require a robot to modify its planned movement when conditions change.

Motion control and actuation

Actuators convert control commands into mechanical action. Depending on the robot, these may operate wheels, joints, grippers or other mechanisms.

Motion-control systems use control algorithms and feedback to regulate movement and other physical actions. For mobile robots, this may include controlling vehicle speed, heading and trajectory. For manipulators, it may involve controlling joint positions, velocities or forces.

Control-system design must account for the robot's mechanical characteristics, operating requirements and physical constraints.

Robotic manipulation

Robotic manipulation involves interacting with objects through controlled physical movement.

Manipulation systems commonly include robotic arms, end-effectors and control mechanisms. Tasks may include grasping, placing, handling or positioning objects.

Autonomous manipulation can require perception, motion planning and feedback control to be combined within a coordinated system.

Types and applications

Autonomous robots are used or studied in a range of commercial, industrial and scientific contexts. The degree of autonomy and technical requirements vary significantly between applications.

Industrial and logistics robots

Mobile robots are used in manufacturing and logistics for activities such as material movement and internal transport.

The International Federation of Robotics describes autonomous mobile robots as systems capable of navigating using environmental information and responding to obstacles encountered during operation. Such systems may be coordinated through fleet-management software.

Driverless industrial trucks are addressed by ISO 3691-4:2023, which includes autonomous mobile robots and automated guided vehicles among the examples within its scope.

Service and personal-care robots

Service robots may perform tasks involving assistance, transportation or the delivery of services.

Some operate in settings shared with people, creating requirements for navigation, interaction and physical safety.

ISO 13482:2014 addresses safety requirements for specified categories of personal-care robots, including mobile servant robots, physical assistant robots and person carrier robots.

Earth-moving and mining machinery

Autonomous and semi-autonomous machinery is also used in earth-moving and mining applications.

ISO 17757:2019 establishes safety requirements for autonomous and semi-autonomous machines and their associated systems in these sectors. Its scope includes relevant machinery, system infrastructure and considerations throughout the operational life cycle.

Space exploration

Autonomous navigation is an established capability in planetary robotic exploration.

NASA's Perseverance Mars rover uses an autonomous navigation system known as AutoNav. According to NASA, AutoNav allows the rover to construct three-dimensional representations of nearby terrain, identify obstacles and plan routes toward destinations established by mission controllers.

This illustrates task-specific autonomy: mission personnel establish operational objectives, while the rover can make certain local navigation decisions without requiring a separate human command for each movement.

Safety and operational limitations

Autonomous robots operate in physical environments where mechanical movement and interaction can create hazards.

Relevant risks depend on the robot's design and application and may involve collisions, unexpected movement, equipment failures, environmental obstacles or failures in sensing and control.

Risk assessment

ISO 12100:2010 establishes general principles for machinery risk assessment and risk reduction. Its methodology addresses hazard identification, risk estimation, evaluation and measures for reducing identified risks.

For robotic systems, safety requirements must be considered in relation to the machine's intended use and the hazards arising from its operation.

Application-specific standards

Different classes of robotic equipment are subject to different technical safety requirements.

• ISO 10218-1:2025 addresses safety requirements for industrial robots.
• ISO 10218-2:2025 addresses industrial robot applications and robot cells.
• ISO 3691-4:2023 concerns driverless industrial trucks and their systems.
• ISO 13482:2014 addresses specified personal-care robots.
• ISO 17757:2019 concerns autonomous and semi-autonomous machinery in earth-moving and mining.

These standards have different scopes and exclusions. A safety standard for one robot category should not automatically be assumed applicable to another.

Reliability and environmental uncertainty

Autonomous robots depend on the quality of information available to their sensing, estimation and control systems.

Uncertainty may arise from measurement noise, imperfect models and incomplete environmental information. Probabilistic approaches to localization, mapping and decision-making provide methods for representing and reasoning about some of these uncertainties.

Autonomous navigation capabilities should therefore be assessed in the environments and conditions for which the systems are intended.

Cybersecurity, identity and control

Connected robotic systems may exchange commands, operational information and software updates through digital communication systems.

Where robots form part of operational technology environments, cybersecurity considerations must account for the interaction between digital systems and physical processes.

The United States National Institute of Standards and Technology (NIST) addresses such concerns in Special Publication 800-82 Revision 3, which discusses operational technology security and the associated requirements for performance, reliability and safety.

Device identity and authentication

Authentication establishes the identity of an entity attempting to access a system or resource. Authorization determines the operations or resources that the authenticated entity is permitted to access.

For network-connected robotic systems, these concepts are relevant to access involving operators, devices, applications and management services.

NIST's Zero Trust Architecture, SP 800-207, describes an approach that does not grant implicit trust solely because an entity is connected to a particular network or is situated within a network boundary.

These are general cybersecurity principles rather than a robot-specific certification of safe autonomous behavior.

Protection of operational commands

Unauthorized access to a robot's control environment can create cybersecurity risks. In systems where digital commands produce physical actions, the consequences of security failures may extend beyond information systems.

Security controls should therefore be evaluated alongside the system's operational and safety requirements. NIST SP 800-82 Rev. 3 provides guidance for managing cybersecurity risks in operational technology environments.

Cybersecurity measures and physical machinery safeguards address different aspects of system risk. Authentication or access control alone does not establish that a robot's physical behavior will be safe.

Evaluation of autonomous capabilities

The capabilities of an autonomous robot should be described in relation to its intended tasks, operating conditions and the level of human involvement.

For mobile systems, technically relevant areas of evaluation include localization, mapping, motion planning and control. These areas are treated in established mobile robotics and probabilistic robotics literature.

Evaluation should distinguish between what a system is designed to do, what it has demonstrated under specified conditions and what remains outside its tested capabilities.

A system's ability to complete a route or manipulation task autonomously is not, by itself, evidence of unrestricted operation or general-purpose intelligence.

Relationship to artificial intelligence

Autonomous robotics draws on methods from several fields, including artificial intelligence, computer vision, probability theory, motion planning and control engineering.

Robot perception and decision-making may use machine-learning methods, but autonomous operation does not necessarily depend on machine learning. Established approaches also include mathematical modeling, state estimation, geometric planning and feedback control.

The relationship between robotics and artificial intelligence therefore depends on the functions implemented in a particular system rather than the mere presence of autonomous movement or automated physical action.

See also

• Robotics
• Autonomous systems
• Artificial intelligence
• Physical AI
• Robot perception
• Computer vision
• Machine learning
• Simultaneous localization and mapping
• Motion planning
• Edge AI
• Cybersecurity

Further reading

Books and academic references

• Siegwart, Roland; Nourbakhsh, Illah Reza; Scaramuzza, Davide (2011). Introduction to Autonomous Mobile Robots. 2nd edition. MIT Press. ISBN 978-0-262-01535-6. Publisher.
• Thrun, Sebastian; Burgard, Wolfram; Fox, Dieter (2005). Probabilistic Robotics. MIT Press. ISBN 978-0-262-20162-9. Publisher.
• Siciliano, Bruno; Khatib, Oussama (eds.) (2016). Springer Handbook of Robotics. 2nd edition. Springer. ISBN 978-3-319-32552-1. Publisher.

Technical standards and publications

• ISO 8373:2021 — Robotics — Vocabulary.
• ISO 12100:2010 — Safety of machinery — Risk assessment and risk reduction.
• ISO 3691-4:2023 — Driverless industrial trucks and their systems.
• ISO 10218-1:2025 — Safety requirements for industrial robots.
• ISO 10218-2:2025 — Industrial robot applications and robot cells.
• NIST SP 800-82 Rev. 3 — Guide to Operational Technology Security.
• NIST SP 800-207 — Zero Trust Architecture.