Safety Standards for Robots and Machines: How to Use Them

Engineering safe and compliant intelligent machines in the European Union is a discipline that sits at the intersection of technical design, legal obligation, and operational governance. For robots and smart machines, safety is not a feature to be added at the end of a development cycle; it is a property of the system that must be engineered from the first concept and continuously managed throughout the machine’s life. The regulatory framework governing these systems is a layered structure, where the European Union sets the high-level legal requirements through the CE marking directives and regulations, and standardisation bodies such as ISO, IEC, and CEN provide the detailed technical methodologies for meeting those requirements. Understanding how to use these standards in practice is therefore not merely a technical exercise; it is a core component of legal compliance and risk management.

This article addresses the practical application of safety standards for robots and smart machines within the European regulatory context. It is written for engineers, compliance officers, legal counsel, and system architects who must navigate the interplay between the EU’s legal framework and the technical specifications that make compliance achievable. We will explore how standards function as a bridge between legal principles and engineering practice, how to select and apply the correct standards for a given machine, and how to manage the compliance lifecycle in a way that satisfies both regulatory authorities and market expectations.

The European Regulatory Framework for Machine Safety

The European Union’s approach to the safety of machinery is built on the principle of the “New Legislative Framework” (NLF). This framework establishes a harmonised system for the marketing of products within the European Economic Area (EEA). For robots and machines, the primary legal instrument is the Machinery Directive (2006/42/EC), which is currently being superseded by the Machinery Regulation (2023/1230), applicable from January 14, 2027. This transition from a directive to a regulation is significant: a regulation is directly applicable law in all member states, whereas a directive must be transposed into national law. This will harmonise requirements more strictly across the EU, reducing the scope for national variations that currently exist under the transposition of the directive.

Under this framework, a manufacturer places a machine on the market by ensuring it conforms to the “essential health and safety requirements” (EHSRs) listed in Annex I of the directive or regulation. The manufacturer then compiles a technical file, draws up a Declaration of Conformity, and affixes the CE marking. Crucially, the EHSRs are often very broad principles (e.g., “The machinery must be designed and constructed so that it is safe for the intended use”). They do not specify exactly how to achieve safety. This is where harmonised standards come into play.

Harmonised Standards: The Presumption of Conformity

Harmonised standards are technical standards that the European Commission has mandated and published in the Official Journal of the EU. When a manufacturer applies a harmonised standard to their machine, they benefit from a “presumption of conformity.” This means that if the machine meets all the relevant requirements of that standard, it is presumed to meet the corresponding EHSRs of the Machinery Directive/Regulation. This is the most important legal concept for engineers and compliance professionals to understand. It provides a clear, defensible pathway to compliance.

However, it is vital to recognise that the application of harmonised standards is, in most cases, voluntary. A manufacturer can choose to design a machine in a different way, but if they do, they carry the full burden of proving to a notified body or market surveillance authority that their alternative solution meets or exceeds the safety level of the harmonised standard. In practice, for complex systems like robots, deviating from established standards is a high-risk strategy that requires extensive justification and documentation.

The Role of Non-Harmonised Standards

Not all useful standards are harmonised. Many valuable technical specifications exist at the international (ISO/IEC) or national (DIN, BS, NF) level. While these do not provide a direct presumption of conformity, they are widely recognised as representing the state of the art. In the event of an incident or a regulatory challenge, adherence to non-harmonised but widely accepted standards (like many ISO standards) is a powerful indicator of due diligence and responsible engineering. They are used to fill gaps where harmonised standards are under development or are not specific enough for a novel technology.

The Hierarchy and Structure of Safety Standards

When designing a safe robotic or automated system, one does not simply pick a single standard and follow it. Safety is a system property, and standards are structured to address different levels of this system. The most widely used classification is the “Type A, Type B, Type C” hierarchy.

Type A Standards (Fundamental Safety)

Type A standards provide the foundational concepts and terminology for machinery safety. The primary Type A standard is EN ISO 12100:2010, titled “Safety of machinery — General principles for design — Risk assessment and risk reduction.” This standard is the bedrock of machinery safety in Europe. It does not give specific instructions for a particular type of machine. Instead, it defines the overall process that a manufacturer must follow:

• Risk Analysis: Identifying potential hazards.
• Risk Assessment: Estimating and evaluating the risks.
• Risk Reduction: Implementing measures to eliminate or reduce the risks.

Following the process in EN ISO 12100 is the first step in complying with the EHSRs. It provides the logical framework for all subsequent safety decisions. If a manufacturer is challenged by a regulator, demonstrating a rigorous application of the risk assessment process according to this standard is a primary line of defence.

Type B Standards (Generic Safety)

Type B standards deal with safety aspects that can be applied to a wide range of machinery. They are more specific than Type A but not tied to one particular machine type. They are divided into B1 and B2 standards.

B1 standards cover specific safety aspects, such as:

• EN ISO 13849-1: Safety-related parts of control systems – General principles for design.
• EN ISO 13850: Emergency stop function.
• EN ISO 13857: Safety distances to prevent hazard zones being reached.

B2 standards cover safety devices, such as:

• EN ISO 13849-1 (also falls here in some classifications, but it’s a key standard for performance levels).
• EN 62061: Functional safety of safety-related electrical, electronic and programmable electronic control systems (the machinery-specific application of IEC 61508).
• EN ISO 14119: Interlocking devices.

Type C Standards (Machine-Specific Safety)

Type C standards are the most specific. They deal with the safety requirements for a particular type of machine. For robotics, the most important Type C standard is EN ISO 10218-1 (Robots and robotic devices — Safety requirements — Part 1: Industrial robots) and EN ISO 10218-2 (Part 2: Robot systems and integration). These standards specify safety requirements for the design, construction, and integration of industrial robots. They provide detailed requirements on things like speed monitoring, workspace limits, and collaborative robot operation.

When a Type C standard exists for a specific machine, it takes precedence over the more generic standards. However, the principles of risk assessment from EN ISO 12100 still apply, and the requirements of Type B standards must be met unless the Type C standard explicitly provides an alternative.

Practical Application: A Step-by-Step Approach to Compliance

Translating this framework into a practical workflow for a development team requires a structured, documented process. The following steps outline a robust methodology for using standards to achieve compliance for a new robot or smart machine.

Step 1: Define the Machine and its Intended Use

Before any risk assessment can begin, the manufacturer must clearly define the machine, its functions, its operational modes, and its foreseeable misuse. This is a foundational requirement of EN ISO 12100. For a collaborative robot, this would include defining the tasks, the speed and force limits, the types of tools to be attached, and the environment in which it will operate. This definition is critical because the entire safety analysis is based on the “reasonably foreseeable use” of the machine. A failure to properly define the scope can lead to missed hazards and an incomplete technical file.

Step 2: Conduct a Comprehensive Risk Assessment

The risk assessment is the central pillar of the safety lifecycle. It is an iterative process that runs parallel to the design process. The team must systematically identify all potential hazards associated with the machine. For a robot, this includes hazards from:

• Movement: Crushing, trapping, impact.
• Energy: Electrical, hydraulic, pneumatic, thermal.
• Environment: Slips, trips, falls, interference with other equipment.
• Human Factors: Operator error, unexpected behaviour.

For each identified hazard, the team must estimate the risk, considering the severity of potential harm and the probability of occurrence. This is often formalised using risk assessment tools like a risk graph or a risk matrix. The output of this process is a list of hazards that require risk reduction measures. This entire process must be meticulously documented in the technical file, as it forms the justification for the safety measures chosen.

Step 3: Implement the Risk Reduction Measures

Following the hierarchy of controls defined in EN ISO 12100, the team must implement measures to reduce the risk. The preferred order is:

1. Inherently safe design: Eliminate the hazard at the source (e.g., designing the robot to not have sharp edges, limiting its maximum speed).
2. Guarding and protective devices: If the hazard cannot be designed out, use physical guards (fences, cages) or protective devices (light curtains, safety mats, interlocks) to prevent access to the danger zone.
3. Information for use: Provide warnings, training instructions, and operating manuals. This is the last line of defence and is not sufficient on its own to mitigate significant risks.

This is where the specific standards come into play. For example, if a light curtain is used to protect a robot cell, the team must apply EN ISO 13855 to calculate the correct safety distance. If a safety PLC is used to control the robot’s stop function, the team must apply EN ISO 13849-1 to design and verify the safety-related parts of the control system to the required Performance Level (PL).

Step 4: Select and Apply the Relevant Standards

Based on the risk assessment and the chosen risk reduction measures, the team must identify the relevant standards. For an industrial robot, this will inevitably include:

• EN ISO 12100: For the overall risk assessment process.
• EN ISO 10218-1 and -2: For robot-specific requirements and integration.
• EN ISO 13849-1: For the safety control system’s performance level.
• EN ISO 13850: For the emergency stop function.

If the robot is a collaborative robot (cobot), an additional critical standard is ISO/TS 15066. This technical specification provides detailed information on the forces and pressures that can be safely applied to a human body part during a collaborative operation. It is not a harmonised standard under the Machinery Directive, but it is the definitive guide for designing safe cobot applications. The values in this document are essential for performing the risk assessment for collaborative applications.

Step 5: Verification and Validation

Compliance is not just about design; it is about proving the design works. The standards require a two-pronged approach:

• Verification: Confirming that the design meets the specified requirements. This is a technical check. For example, verifying that the safety PLC is configured to the correct PL(r) (required Performance Level) and that the safety functions perform as designed under all fault conditions.
• Validation: Confirming that the final machine is safe for its intended use. This is a holistic check, often involving testing the machine in its operational environment. It confirms that the risk assessment was correct and the risk reduction measures are effective.

For complex systems, this may involve a third-party inspection by a “Notified Body” if the machine falls under Annex IV of the Machinery Directive (or the equivalent list in the Machinery Regulation). This is particularly relevant for robots intended to operate in collaborative modes, as this is considered a high-risk application.

Deep Dive: Key Standards for Robotics and Smart Machines

To illustrate the practical application, let’s examine some of the most critical standards in more detail, focusing on how they are used by engineers and compliance teams.

EN ISO 13849-1: Safety-Related Parts of Control Systems

This standard is arguably one of the most important for modern machines. It provides a framework for evaluating the reliability of safety functions. It replaces the older, more simplistic “Categories” (B, 1, 2, 3, 4) with a more nuanced concept: the Performance Level (PL). A PL is a value from ‘a’ (lowest) to ‘e’ (highest). To determine the required PL for a safety function, the team uses a risk graph (provided in the standard) that considers the severity of potential injury, the frequency of exposure to the hazard, and the possibility of avoiding the hazard.

Once the required PL (PLr) is established, the team must design the safety circuit to meet or exceed it. The standard defines five key metrics that contribute to the achieved PL:

• Category: The architecture of the system (e.g., single channel, dual channel with monitoring).
• MTTFd: Mean Time to Dangerous Failure of the components.
• DC: Diagnostic Coverage to detect dangerous failures.
• CCF: Common Cause Failures (e.g., ensuring separation of redundant channels).

In practice, an engineer will select components (sensors, logic solvers, actuators) with known MTTFd values, design the architecture to a specific Category, and then perform the calculations to prove that the overall system meets the PLr. This is a core part of the technical file and is a non-negotiable requirement for any safety-related control system on a modern machine.

EN ISO 10218-1 & -2: Industrial Robots

These standards are the bible for industrial robot safety. Part 1 applies to the robot manufacturer (the “robot arm” itself), and Part 2 applies to the system integrator (the one who builds the work cell). The standards specify requirements for:

• Robot design: Requirements for joints, speed limits, and emergency stop circuits.
• Control system: Requirements for modes of operation (e.g., automatic, manual), enabling devices (hold-to-run), and safe speed limits.
• Integration (Part 2): Requirements for the safeguarded workspace, risk assessment of the entire system, and verification of the integrated system.

A key practical aspect of these standards is their treatment of collaborative operation. They define four types of collaborative operation: Safety-Rated Monitored Stop, Hand Guiding, Speed and Separation Monitoring, and Power and Force Limiting. For the latter two, the standard defers to the values in ISO/TS 15066 for the actual force and pressure limits. This demonstrates the layered nature of the standards: EN ISO 10218 sets the operational framework, while ISO/TS 15066 provides the specific safety parameters.

ISO/TS 15066: Collaborative Robots

This technical specification is critical for anyone working with cobots. It provides the scientific basis for determining if a contact between a robot and a human is safe. It defines pain thresholds for different parts of the human body (e.g., fingertips, hands, arms, head). For example, it specifies the maximum quasi-static pressure (in Newtons per square centimetre) and the maximum transient force that can be applied to a fingertip without causing injury.

In practice, a system integrator designing a collaborative application must:

1. Identify all potential contact points between the robot (including the workpiece) and the human operator.
2. Use the values from the tables in ISO/TS 15066 to determine the applicable force/pressure limits for each contact point.
3. Design the robot system (e.g., by limiting speed, using force sensors, or designing the tool shape) so that it cannot exceed these limits under any foreseeable condition, including a fault.
4. Document this analysis in the risk assessment.

Failure to adhere to these limits can result in the classification of the application as “non-collaborative,” meaning it would require traditional, more rigid safeguarding (like a physical cage), thereby defeating the purpose of a cobot.

Functional Safety and the IEC 61508 Ecosystem

For smart machines and robots with complex software and electronic systems, the principles of functional safety are paramount.