When a collaborative robot shares a workspace with a human worker, the stakes for getting safety right could not be higher. Unlike traditional industrial robots locked behind fences, cobots operate in the open — reaching across workbenches, handing off parts, and moving through shared aisles. The standard that governs this new reality is ISO/TS 15066, the first international technical specification to define precise safety requirements for human-robot collaboration, including the widely adopted power-and-force-limiting (PFL) mode.

For engineers, safety managers, and operations leaders investing in industrial automation, ISO/TS 15066 is not just a compliance checkbox. It is a practical engineering framework that determines how fast a robot can move, how much force it can apply, and what happens the moment a person steps into its path. Getting these parameters right separates a genuinely safe collaborative system from one that merely looks safe on paper.

This article breaks down what ISO/TS 15066 actually requires, how power-and-force-limiting works at a technical and operational level, what the biomechanical thresholds mean in a real factory setting, and how these principles extend beyond robot arms to the broader world of autonomous mobile robots and smart warehouse automation.

What Is ISO/TS 15066 and Why Does It Matter?

Published by the International Organization for Standardization, ISO/TS 15066 is a technical specification that supplements the foundational robot safety standard ISO 10218. While ISO 10218 covers the general design and safeguarding of industrial robots, ISO/TS 15066 specifically addresses collaborative robot systems — situations where robots and humans intentionally share the same workspace at the same time without a physical barrier between them.

The specification matters for one fundamental reason: traditional robot safety relied on separation. Keep the human out, and you never have to worry about the robot injuring them. Collaborative robotics eliminates that physical separation, which means safety must be engineered directly into the robot’s behavior, speed, force output, and situational awareness. ISO/TS 15066 provides the technical language and measurement criteria that make this engineering rigorous rather than guesswork.

Compliance with ISO/TS 15066 is also increasingly a commercial requirement. Many large manufacturers, automotive OEMs, and logistics operators now specify ISO/TS 15066 conformance when procuring collaborative automation. Understanding the standard is therefore not just a safety obligation but a business prerequisite for any organization deploying cobots in shared workspaces.

The Four Types of Collaborative Operation

ISO/TS 15066 defines four distinct modes of collaborative operation, and it is important to understand that most cobot deployments use a combination of these modes rather than a single approach. Each mode has different implications for workspace design and risk assessment.

• Safety-Rated Monitored Stop (SRMS): The robot pauses whenever a person enters the collaborative workspace and resumes only after the human has left. This is the most restrictive mode and is often used for operations requiring occasional human access to a largely automated cell.
• Hand Guiding: A human operator directly guides the robot’s end effector through a task, with the robot providing force feedback and following the operator’s lead. This mode is common in teaching or setup scenarios.
• Speed and Separation Monitoring (SSM): The robot monitors the distance between itself and detected humans in real time, reducing speed as the human approaches and stopping if the minimum protective distance is breached. This requires reliable presence-detection technology such as laser scanners or vision systems.
• Power and Force Limiting (PFL): The robot is designed and configured so that any contact with a human — whether intentional or accidental — does not cause injury. This is currently the most widely implemented collaborative mode and the primary focus of ISO/TS 15066’s biomechanical data.

In many real deployments, SSM and PFL are used together: the robot slows as a person approaches under SSM, and if contact does occur, PFL ensures the force is within safe limits. Understanding which mode applies to which phase of an operation is a core part of the risk assessment process.

Power-and-Force-Limiting Explained

Power-and-force-limiting is a design philosophy as much as a technical specification. At its core, PFL means that a robot’s mechanical design, control system, and operating parameters are configured so that the forces generated during a collision with a human body remain below levels that would cause injury. This can be achieved through inherently compliant mechanical design, torque-sensing at each joint, real-time force monitoring, or a combination of all three.

Modern PFL cobots continuously measure the forces and torques at every joint during operation. If the measured values approach or exceed a preset threshold, the control system immediately reduces motor torque or brings the robot to a stop. The response time is typically in the range of milliseconds, fast enough to limit the energy transferred in a transient contact event before tissue damage can occur.

It is important to recognize what PFL does and does not cover. PFL is designed to handle transient contact — brief, incidental collisions where the human and robot briefly touch and then separate. It is not designed to make a robot safe if a human is trapped between the robot and a fixed surface, a scenario known as quasi-static contact. In quasi-static situations, the robot continues to press against the human because there is no path for the limb to move away, and the pressure can build to harmful levels even at low force settings. Identifying and mitigating quasi-static contact risks is one of the most critical — and most frequently underestimated — parts of a PFL risk assessment.

Biomechanical Thresholds: The Numbers Behind Safe Contact

The most technically detailed contribution of ISO/TS 15066 is its Annex A, which provides biomechanical limit values for different body regions. These values define the maximum allowable force and pressure that each part of the human body can tolerate during transient and quasi-static contact without injury. The data was derived from extensive biomechanical research into pain thresholds, tissue stiffness, and injury onset across different body regions.

The specification identifies 29 body regions and provides values for both force (in Newtons) and pressure (in N/cm²) for each. To illustrate the range of these thresholds:

• The skull and forehead are among the most tolerant regions, with higher allowable force values because the underlying bone provides significant protection.
• The hands and fingers have relatively lower thresholds, particularly for quasi-static contact, because the fine structures of the hand are vulnerable to compression injuries.
• The chest and abdomen have moderate thresholds for transient contact but lower limits for quasi-static scenarios due to the risk of internal compression.

In practice, system integrators use these values to set the force and torque limits on each robot joint, taking into account the geometry of the end effector, the payload being carried, and the specific motion paths the robot follows. A robot carrying a sharp or rigid tool will have a very different effective contact pressure than one carrying a soft, padded gripper — and the risk assessment must account for this. Selecting appropriate end effectors and tooling is therefore not just a performance decision but a direct safety engineering decision under ISO/TS 15066.

Risk Assessment Under ISO/TS 15066

ISO/TS 15066 does not replace the general risk assessment requirements of ISO 12100 and ISO 10218-2; it builds on them. The risk assessment process for a collaborative robot application must identify every plausible human-robot contact scenario, classify each contact as transient or quasi-static, identify the body regions at risk, and verify that the robot’s operating parameters keep forces and pressures within the Annex A limits for those regions.

A structured risk assessment for a PFL deployment typically follows these key steps:

1. Define the collaborative workspace – Map exactly where the robot and humans will operate simultaneously, including all approach paths and work positions.
2. Identify contact scenarios – List every way a human could come into contact with the moving robot, including intentional interaction and accidental collision from any direction.
3. Classify contact type – Determine whether each scenario results in transient or quasi-static contact, paying particular attention to pinch points near fixed surfaces, fixtures, or other equipment.
4. Map body regions at risk – For each contact scenario, identify which body parts are realistically in the contact zone.
5. Verify force and pressure limits – Use measurement tools (force plates, pressure-sensitive mats, or validated simulation) to confirm the robot’s actual contact forces and pressures fall within ISO/TS 15066 Annex A values for the identified body regions.
6. Document and validate – Record all findings, parameter settings, and validation measurements in a format that supports audit and re-assessment when the application changes.

One critical point that experienced safety engineers emphasize: the risk assessment must be repeated whenever the application changes. A robot that was validated safe at a certain speed and payload configuration is not automatically safe if the tool is swapped, the payload increases, or the workspace layout changes. ISO/TS 15066 compliance is a continuous process, not a one-time certification.

Power-and-Force-Limiting in Real Industrial Environments

Understanding the theory of PFL is one thing; deploying it effectively in a busy production environment is another. Several practical challenges consistently arise when implementing power-and-force-limiting in real facilities.

Speed versus safety trade-offs are the most common operational friction point. PFL inherently requires the robot to operate at speeds and force levels below what an uncollaborative system could achieve. Production engineers often push to increase robot speed to meet cycle time targets, but higher speeds mean higher contact forces in a collision event. Every speed increase must be validated against the biomechanical limits, and there is a ceiling beyond which a robot simply cannot operate collaboratively regardless of how good its force sensing is.

Tooling and end effector design is frequently underestimated. A robot arm that passes all force limit tests with a compliant gripper may generate dangerous pressure concentrations when fitted with a rigid metal fixture or a pointed component. Integrators must evaluate the actual geometry of the contact surface — not just the peak force — because pressure (force per unit area) is what determines injury risk. Rounded edges, padded surfaces, and large contact areas all reduce effective pressure and improve the margin of safety.

Operator acceptance and behavior change is a human factors challenge that no technical standard fully addresses. Workers who are accustomed to traditional fenced robots may either be overly cautious around cobots (reducing productivity) or, conversely, overly familiar (taking risks that erode safety margins). Training, clear visual cues, and well-designed collaborative workspace boundaries all contribute to maintaining the intended safety culture alongside the technical safeguards.

How AMR Safety Connects to Collaborative Robot Standards

While ISO/TS 15066 specifically addresses stationary or quasi-stationary collaborative robot arms, the safety principles it embodies — particularly the emphasis on limiting contact forces, continuous environment sensing, and structured risk assessment — are directly relevant to autonomous mobile robots (AMRs) operating in shared human spaces. Mobile robots that move through factory floors and warehouses alongside workers face many of the same fundamental collision-avoidance and force-limitation challenges as cobots at a fixed workstation.

Reeman’s autonomous mobile robots, including platforms such as the Big Dog Delivery Robot and the Fly Boat Delivery Robot, are designed with human-shared environments in mind. These platforms use laser navigation and SLAM mapping to maintain real-time awareness of their surroundings, enabling autonomous obstacle avoidance that mirrors the speed-and-separation monitoring principles described in ISO/TS 15066. The underlying goal is the same: ensure that the robot can coexist with humans in an unstructured environment without creating injury risk.

For operations looking at a more modular approach to mobile autonomy, Reeman’s robot chassis platforms — including the Big Dog Robot Chassis, the Fly Boat Robot Chassis, and the Moon Knight Robot Chassis — provide the foundational mobility layer on which collaborative and safety-aware automation applications can be built. The full range of mobile chassis platforms is designed to support industry-specific deployment with built-in safety-oriented navigation.

In warehouse and logistics environments where autonomous forklifts share aisles with pedestrians, the force-limitation and detection principles become even more critical given the much larger mass and energy involved. Reeman’s autonomous forklift lineup — including the Ironhide Autonomous Forklift, the Stackman 1200, and the Rhinoceros Autonomous Forklift — incorporates multi-layer safety sensing to detect and respond to human presence in the vehicle’s operating zone, applying the same philosophical framework that ISO/TS 15066 establishes for collaborative robot arms. The IronBov Latent Transport Robot further extends these principles to latent-type AMR operations in dense warehouse environments.

Common Misconceptions About Cobot Safety

Several persistent misconceptions can undermine safety in collaborative robot deployments, and it is worth addressing them directly.

“Cobots are inherently safe, so no risk assessment is needed.” This is perhaps the most dangerous misconception in the industry. ISO/TS 15066 is explicit that even robots with built-in force-limiting capabilities require a rigorous, application-specific risk assessment. A cobot is designed to be capable of safe collaboration — but whether it actually achieves that in a given installation depends entirely on how it is configured, what tooling it carries, and how the workspace is designed. The robot manufacturer’s safety certification covers the robot itself; it does not cover the complete integrated system in your specific application.

“If the robot stops when it touches something, it is safe.” A robot that stops on contact has reduced the duration of contact, but if the initial contact force exceeds the biomechanical limits for the affected body region, injury can still occur in the brief moment before the stop. This is why PFL systems aim to limit the force of contact itself, not just its duration. Stopping behavior is a secondary safeguard, not a substitute for proper force limitation.

“ISO/TS 15066 compliance means the same thing everywhere.” ISO/TS 15066 is a technical specification, not a certification scheme. There is no single global body that certifies robots or installations as “ISO/TS 15066 compliant.” Conformance is demonstrated through the risk assessment process and validated through measurement, and it is the responsibility of the system integrator and the end user — not just the robot manufacturer — to demonstrate that the complete installation meets the specification’s requirements.

Conclusion

ISO/TS 15066 and power-and-force-limiting represent a genuine step forward in making human-robot collaboration both productive and safe. By grounding safety requirements in biomechanical data rather than general engineering judgment, the specification gives system integrators and safety managers a concrete, measurable framework for validating collaborative robot applications. The key takeaways are straightforward: PFL is powerful but not unlimited, risk assessment is mandatory and must be repeated when anything changes, quasi-static contact scenarios deserve special attention, and end effector design has a direct and significant impact on actual contact pressure.

As industrial automation continues to evolve beyond stationary robot arms toward fully mobile autonomous systems, the principles embedded in ISO/TS 15066 — sensing-based awareness, force limitation, structured risk management, and continuous validation — will remain the foundation of safe human-robot coexistence, whether the robot in question is a collaborative arm on a workbench or an autonomous mobile robot navigating a busy warehouse floor. Building safety in from the ground up, rather than adding it as an afterthought, is what separates automation that truly transforms operations from automation that creates new risks in the process of solving old ones.