Inferensys

Glossary

Power and Force Limiting (PFL)

Power and Force Limiting (PFL) is a collaborative robot safety mode defined in ISO/TS 15066 where the robot's inherent design limits its power and force to levels that will not cause injury in the event of contact with a human.
Stylish WeWork-like workspace with hot desks and document wall, professional searching through enterprise knowledge base on a mounted ultrawide display, warm industrial pendants overhead.
COLLABORATIVE ROBOT SAFETY

What is Power and Force Limiting (PFL)?

Power and Force Limiting (PFL) is a foundational safety methodology for collaborative robots, defined by international standards, that ensures safe physical contact between humans and machines.

Power and Force Limiting (PFL) is a collaborative robot safety mode, formally defined in ISO/TS 15066, where the robot's inherent design and control system cap its maximum output power and force to biomechanically safe thresholds. This engineering ensures that any unintended contact with a human—whether transient or prolonged—will not cause pain or injury. PFL is a core enabler of collaborative operation, allowing humans and robots to work in direct proximity without traditional safety cages or fences.

Implementing PFL requires a holistic system design integrating inherently safe mechanics (e.g., lightweight structures, rounded edges, compliant joints) with active control strategies that monitor joint torque and velocity. The system must be validated against standardized biomechanical limits for different body regions. This approach is distinct from reactive methods like Safety-Rated Monitored Stop, as PFL allows the robot to continue operating safely during contact, enabling true physical human-robot interaction (pHRI) and tasks like hand guiding.

COLLABORATIVE ROBOT SAFETY

Key Characteristics of PFL

Power and Force Limiting (PFL) is a foundational safety mode for collaborative robots defined by ISO/TS 15066. Its core function is to intrinsically cap the robot's kinetic energy to levels deemed safe for incidental contact with a human operator.

01

Intrinsic Design Limitation

PFL is not an add-on sensor; it is an inherent property of the robot's mechanical and control design. This is achieved through:

  • Low-inertia actuators and lightweight links to minimize kinetic energy.
  • Torque sensors in each joint to monitor and limit applied forces in real-time.
  • Backdrivable gearboxes that allow the arm to be easily pushed away by a human. This design-first approach ensures safety is always active, even in the event of a sensor or software failure.
02

Biomechanical Injury Thresholds

PFL limits are based on pain and injury thresholds defined in ISO/TS 15066. The standard provides maximum permissible values for different body regions, distinguishing between:

  • Transient Contact: Brief, impact-like contact (e.g., a bump). Limits are higher, focusing on peak force and pressure.
  • Quasi-Static Contact: Prolonged clamping or trapping. Limits are much lower to prevent ischemia and tissue damage. For example, the limit for transient contact on the forearm is 150 N, while for quasi-static contact it is 75 N. These values are derived from biomechanical studies.
03

Dynamic Risk Assessment

A PFL robot continuously performs an implicit risk assessment based on its own state. The safety controller evaluates:

  • Tool Center Point (TCP) speed and position relative to the workspace.
  • Kinetic energy calculated from mass and velocity.
  • Contact force and pressure measured at the joints. The system dynamically adjusts its monitored stopping distance and compliance to ensure that if contact occurs, the transmitted energy stays below the biomechanical limits, regardless of the robot's programmed task.
04

Integration with Other Safeguards

PFL is one of four collaborative operation modes in ISO 10218-1/ISO TS 15066. It is often used in combination with:

  • Safety-Rated Monitored Stop: The robot stops when a human enters the workspace but does not limit its inherent power.
  • Hand Guiding: The robot enters a zero-gravity mode for direct teaching.
  • Speed and Separation Monitoring (SSM): Uses external sensors to maintain a protective separation distance. A typical collaborative cell might use SSM for approach and PFL for the final collaborative task phase, creating a layered safety architecture.
05

Performance Validation & Testing

Verifying PFL compliance requires rigorous physical testing, not just simulation. According to ISO/TS 15066, this involves:

  • Using a force-pressure measurement device to simulate human body regions.
  • Conducting transient contact tests by moving the robot at maximum speed into the measurement device.
  • Conducting quasi-static contact tests by commanding the robot to push against the device. All tests must be performed at the point of greatest possible force (usually the TCP with the heaviest approved tool) to certify the worst-case scenario is safe.
06

Application in Physical HRI (pHRI)

PFL is the enabling technology for Physical Human-Robot Interaction (pHRI), where contact is not just possible but part of the task. Key applications include:

  • Cooperative Carrying: Human and robot lift an object together, with the robot complying to the human's lead.
  • Assisted Assembly: The robot presents a part, and the human pushes or guides the arm into final position.
  • Tactile Feedback Tasks: The robot performs a task like sanding or polishing, where the human may need to adjust the tool's pressure by hand. This transforms the robot from an isolated automaton into a compliant, responsive team member.
COLLABORATIVE ROBOT SAFETY

How Power and Force Limiting Works

Power and Force Limiting (PFL) is a foundational safety mode for collaborative robots, enabling safe physical contact between humans and machines by design.

Power and Force Limiting (PFL) is a collaborative robot safety mode, defined in ISO/TS 15066, where the robot's inherent design—through a combination of lightweight materials, backdrivable actuators, and sensitive joint torque sensors—intrinsically caps its kinetic energy and applied force. This engineering ensures that any unintended contact with a human will not exceed biomechanical pain and injury thresholds for transient (brief) or quasi-static (pinching) contact. PFL is the primary technical enabler for robots to work alongside humans without traditional safety cages or fences.

Implementation relies on a safety-rated control system that continuously monitors joint torque and velocity. If contact forces approach predefined limits, the system triggers an immediate protective stop. This differs from speed and separation monitoring, which aims to prevent contact altogether. PFL is essential for applications like hand guiding and collaborative assembly, where intermittent physical interaction is part of the workflow. Its effectiveness is validated against standardized biomechanical data for different body regions.

ISO/TS 15066 DEFINED MODES

PFL vs. Other Collaborative Operation Modes

A comparison of the four primary collaborative operation modes as defined by the ISO/TS 15066 safety standard, highlighting their operational characteristics and safety mechanisms.

Feature / MechanismPower and Force Limiting (PFL)Safety-Rated Monitored StopHand GuidingSpeed and Separation Monitoring (SSM)

Core Safety Principle

Inherently limited contact force

Stop motion on intrusion

Direct human guidance

Maintain protective separation distance

Human-Robot Contact During Operation

Robot Motion During Collaboration

Autonomous, within force limits

Stopped

Compliant, human-directed

Autonomous, speed-modulated

Primary Safety Sensor

Joint torque sensors / current monitoring

Area scanners / light curtains

Force-torque sensor / teach button

Area scanners / vision systems

Requires External Safeguarding (e.g., fences)

Typical Application

Assembly, finishing, co-manipulation

Machine tending, loading/unloading

Programming, precise positioning

Material handling, kitting

Dynamic Response to Human Proximity

Continues task with limited force

Stops; resumes automatically on exit

Follows human input

Reduces speed; stops if minimum distance breached

ISO/TS 15066 Biomechanical Limits Applied

Transient & quasi-static contact

Separation distance formulas

INDUSTRIAL & RESEARCH DOMAINS

Common Applications of PFL

Power and Force Limiting (PFL) enables safe physical collaboration by design. Its applications span industries where robots must work in close proximity to people without traditional safety barriers.

02

Machine Tending

This involves loading and unloading parts from CNC machines, injection molders, or presses. A PFL robot can safely operate in the same space as a human operator who performs quality checks, deburring, or tool changes. The safety mode allows for:

  • Close-quarters operation when the human needs to access the machine.
  • Safe recovery from collisions if the human enters the workspace unexpectedly.
  • Seamless hand-off of raw materials and finished parts, increasing line flexibility and utilization.
03

Quality Inspection & Testing

PFL robots are used to manipulate products or sensors for inspection. A human can physically guide the robot's sensor (e.g., a camera or probe) to a specific area of interest (kinesthetic teaching), and the robot can then autonomously repeat the scan. In material testing, the robot can apply controlled, limited forces for destructive or non-destructive testing (e.g., button press testing, flex testing) with the operator present to load samples and evaluate results in real-time.

04

Packaging & Palletizing

In end-of-line operations, PFL cobots collaborate with humans to handle irregular, fragile, or high-mix items. The human might orient complex items or build mixed pallet patterns, while the robot handles the heavy, repetitive lifting and placing. The force-limiting capability is critical for handling deformable packages (e.g., bags, boxes) without crushing them and for safe operation in spaces where people are manually loading trucks or applying labels.

05

Laboratory Automation

In research labs (pharmaceutical, life sciences, chemistry), PFL robots safely work alongside scientists. They automate repetitive tasks like:

  • Liquid handling and pipetting.
  • Microplate manipulation.
  • Weighing and dispensing powders. The safe contact force allows a researcher to interrupt a protocol, manually adjust a sample, or correct a misaligned vial without triggering an emergency stop, maintaining workflow continuity and protecting expensive samples and equipment.
06

Rehabilitation & Physical Therapy

PFL is a foundational technology for robotic rehabilitation devices. Exoskeletons and therapy robots use force-limited actuation to provide assist-as-needed support for patients relearning motor skills. The system can safely apply therapeutic forces for gait training or upper-limb movement while ensuring it never exerts force beyond the patient's pain or injury threshold. This enables close physical interaction essential for effective, adaptive therapy.

POWER AND FORCE LIMITING (PFL)

Frequently Asked Questions

Power and Force Limiting (PFL) is a foundational safety standard for collaborative robots. These questions address its technical implementation, standards, and role in modern human-robot collaboration.

Power and Force Limiting (PFL) is a collaborative robot safety mode defined in ISO/TS 15066 where the robot's inherent design and control software actively restrict its output power and applied force to levels deemed non-injurious in the event of unexpected contact with a human. It works by integrating multiple safety features: inherently force-limited robot arms with back-drivable joints and low inertia, torque sensors in each axis to monitor applied force in real-time, and a safety-rated monitored speed control that dynamically reduces velocity when contact forces approach predefined biomechanical thresholds. The system continuously compares measured forces against transient and quasi-static contact limits (e.g., 150 N for the abdomen) and will trigger a protective stop if thresholds are exceeded.

Prasad Kumkar

About the author

Prasad Kumkar

CEO & MD, Inference Systems

Prasad Kumkar is the CEO & MD of Inference Systems and writes about AI systems architecture, LLM infrastructure, model serving, evaluation, and production deployment. Over 5+ years, he has worked across computer vision models, L5 autonomous vehicle systems, and LLM research, with a focus on taking complex AI ideas into real-world engineering systems.

His work and writing cover AI systems, large language models, AI agents, multimodal systems, autonomous systems, inference optimization, RAG, evaluation, and production AI engineering.