Collaborative Robot Application Guide: Deployment and Safety

Collaborative Robot Application Guide

This guide explains how to plan, design, validate, and deploy collaborative robot applications (cobots) with a focus on safety, standards compliance, and practical engineering controls. It covers the four collaborative operation modes defined in ISO/TS 15066, the implications of the updated ISO 10218-1:2025 terminology (which emphasizes collaborative applications), end-of-arm tooling (EOAT) design guidance, risk assessment methodology, validation and test procedures, and regulatory expectations for documentation and training. Throughout, we reference the principal technical literature and standards that automation engineers and integrators should use as the basis for decisions and acceptance testing.

Key Concepts

Successful collaborative robot projects rely on a rigorous combination of standards-based risk assessment, inherently safe design choices, and controlled operational modes. Two standards form the backbone of collaborative safety assurance:

• ISO/TS 15066:2016 — Provides guidance and quantitative data (force, pressure, and pain thresholds by body region) that integrators use to evaluate Power and Force Limiting (PFL) and to derive acceptable contact limits for quasi-static and transient impacts. It defines four collaborative modes and gives measurement/assessment methods for contact forces and pressures. According to the specification, PFL is the only mode that permits intentional physical contact and therefore requires detailed evaluation of contact characteristics and mitigation measures (ISA InTech 2016; Chemweno et al., 2020).
• ISO 10218-1:2025 — Incorporates the collaborative guidance from ISO/TS 15066 and replaces the term "collaborative robot" with "collaborative robot application," clarifying that safety functions may be implemented in the robot controller, protective devices, or both. The 2025 edition sets requirements and harmonizes language for manufacturers and integrators to document and verify collaborative tasks (ISO OBP).

The industry aligns with these ISO documents through national adoptions such as ANSI/RIA R15.06 in the United States; OSHA enforces workplace safety regulations at the end-user level while standards bodies provide the technical requirements and harmonized practices (Six Degrees of Robotics).

Four Collaborative Operation Modes

ISO/TS 15066 defines four primary operation modes used to structure safety strategy. Each mode has different technical and verification requirements:

• Safety-Monitored Stop (SMS) — The robot stops when a human enters the protected area. Re-start conditions must be safety-verified and monitored. There is no permitted contact during SMS.
• Hand Guiding (HG) — An operator intentionally guides the robot by direct manipulation for tasks such as programming or teaching. The system must ensure that hazards from unintended motions are minimized and that the hand-guiding controls are fail-safe and ergonomic.
• Speed and Separation Monitoring (SSM) — The robot reduces speed or stops to maintain a safe distance between the human and hazardous robot movements; this requires sensors (e.g., light curtains, area scanners) with safety-rated performance to detect person presence and calculate stopping distances.
• Power and Force Limiting (PFL) — Allows direct contact between human and robot. PFL uses a combination of passive means (compliant EOAT, rounded edges, padding) and active means (torque/force limiting, soft-axis limits, monitored stops) to keep contact forces and pressures below the body-part-specific limits stated in ISO/TS 15066. PFL is the only mode that permits intentional physical contact, and it requires documented verification against the ISO/TS 15066 charts (ISA InTech 2016).

Implementation Guide

Implementing a collaborative robot application is a series of engineering activities: site and task assessment, design for inherent safety and controls, EOAT engineering, validation testing, and documentation for compliance. Below we provide a step-by-step procedure that maps to ISO 10218 and ISO/TS 15066 methodologies.

1. Application Assessment and Hazard Analysis

Begin with a full hazard and risk assessment as required by ISO 10218 and reinforced in ISO/TS 15066. Document the task purpose, human presence patterns, frequencies of interaction, maximum reachable speeds, mass/inertia of the robot and tooling, and expected worst-case contact scenarios. Use the following checklist:

• Define workspace layout and delineate collaborative vs. non-collaborative zones.
• List tasks with potential human interaction and identify the body parts that may contact the robot.
• Estimate or measure potential contact conditions (velocity components at impact, mass of contacting parts, angle, duration).
• Rank severity using ISO/TS 15066 body-part charts (quasi-static vs. transient impacts) and record whether the task will fall under PFL or the other three modes.
• Identify possible safeguards and means of reducing risk by design (passive/active).

According to the ISO framework, a thorough analysis must show why chosen safeguards reduce risks to acceptable levels and should be recorded in the machine technical file for the Machinery Directive in the EU or equivalent national documentation elsewhere (Chemweno et al., 2020).

2. Select Collaborative Mode and Safety Functions

Match the task to a collaborative mode. Use SSM when you can maintain reliable separation with sensors; use SMS for low-frequency human entry; use HG for teaching; use PFL only when the task requires direct, predictable contact. For PFL applications, select both passive and active mitigations. Passive mitigations include rounded edges, compliant materials, and reducing tool mass. Active mitigations include software torque limits, safety-rated soft-axis and speed limits, and monitored stops (ISA InTech 2016).

3. End-of-Arm Tooling (EOAT) Design

EOAT design frequently determines whether an application can meet PFL limits. Design objectives include reducing effective mass and impact momentum, eliminating sharp edges, and increasing contact area to reduce pressure. Practical measures:

• Use compliant covers, elastomeric bumpers, or soft padding around contact areas to spread force over a larger surface and prolong contact time (reduces transient peak).
• Minimize mass and inertia of EOAT and attachments; locate high-mass components away from expected contact surfaces.
• Use shear pins, magnetic couplings, or mechanical breakaways where appropriate to prevent high-energy impacts.
• Prefer blunt, rounded geometries and avoid thin edges or protrusions that concentrate pressure.
• Include mounting points for force/torque sensors and for installing safety-rated quick-disconnects to expedite maintenance and reduce downtime.

Design changes must be re-verified with the same assessment and test procedures used for the original EOAT. Many integrators maintain a pre-validated EOAT library with documented mass, center of gravity, and contact surface characteristics to speed re-certification.

4. Controls, Sensors, and Redundancy

SSM requires safety-rated presence-detection sensors and a validated stopping performance (stopping time/distance) given the robot's speed and path. Implement redundant safety functions where failure would lead to hazardous motion. Use safety-rated controllers and certified components for safety functions. Document performance levels and safety integrity requirements per the applicable standard (ISO 13849 / IEC 61508 considerations where applicable).

5. Validation and Testing

Validation confirms that the implemented measures meet the risk acceptance criteria. For PFL, validation includes experimentally measuring contact forces and pressures and comparing them to the ISO/TS 15066 thresholds for the relevant body parts. Test protocols typically include:

• Instrumented trials with calibrated force and pressure transducers to measure quasi-static and transient contact events in worst-case conditions (highest speed, heaviest tooling, most unfavorable contact geometry).
• Repeatable test fixtures to reproduce impact scenarios and to build a traceable test record.
• Verification that active limits (torque, speed) function under fault and recovery conditions and that monitored stops transition to safe states reliably.
• Documentation of test procedures, measurement equipment calibration certificates, and pass/fail criteria, retained in the cell technical file.

ISO/TS 15066 includes guidelines for how to measure and evaluate contact forces and pressures; integrators should follow those procedures and keep measurement records for compliance and liability reduction (ISA InTech 2016; Chemweno et al., 2020).

6. Commissioning and Safe Deployment

Commissioning is a controlled set of steps that transition the system from factory setup to production. Combine engineering acceptance tests with operator training. Key tasks include:

• Functional tests of safety functions and re-check of sensor coverage.
• Emergency stop verification and restart procedures.
• Operator and maintenance staff training on permitted interactions, emergency procedures, and change control rules.
• Installation of fixed signage and markings that define collaborative zones and prohibited actions.
• Finalization of the machine technical file and handover of all safety documentation to the end-user (required by the EU Machinery Directive Annex VII for CE files and recommended elsewhere).

7. Change Management and Maintenance

Any hardware or software changes that affect speed, mass, or contact geometry require a re-assessment and likely re-validation. Institute a change control process that prohibits field modifications without authorized engineering review. Routine maintenance should include inspection of compliant materials, EOAT wear parts, and verification of safety sensor calibration.

Best Practices

Integrators and end-users achieve safer and more reliable cobot installations by following these proven practices:

• Document everything: Keep traceable records of risk assessments, test measurements (including calibration certificates), decision rationales for mode selection, and training logs. Documentation supports conformity assessment and reduces liability exposure (Chemweno et al., 2020).
• Prioritize inherently safe design: Passive measures (geometry, padding, minimizing energy) reduce dependence on sensors and software and improve robustness.
• Use a layered safety approach: Combine physical barriers, presence detection, and PFL or speed limiting so a single failure does not create a hazardous condition.
• Standardize EOATs: Maintain an approved library of tooling with documented mass, CG, and validated contact test results to avoid repeated validation work.
• Train operators and maintainers: Training should cover allowed interactions, emergency stop procedures, and the implications of changing tooling or workpieces.
• Plan for worst-case scenarios: Validate using the most conservative contact assumptions (maximum speed, maximum tool mass, smallest contact area).
• Follow national/adopted standards: Use ISO 10218 and ISO/TS 15066 as primary technical references and follow national implementations (e.g., ANSI/RIA R15.06) and regulatory enforcement (OSHA in the U.S.) (Six Degrees of Robotics).

Comparison Table: Standards and Key Requirements

Standard	Key Requirement for Cobot Deployment	Status / Notes (as of 2026)
ISO/TS 15066:2016	Body-part force/pressure limits; charts for quasi-static and transient impacts; defines SMS, HG, SSM, PFL	Guidance document; limits and test methods widely used; incorporated into ISO 10218-1:2025
ISO 10218-1:2025	Safety requirements for industrial robots; treats collaborative functions as part of the application; requires documented risk assessment	Current edition; harmonizes TS 15066 language and expectations
ANSI/RIA R15.06	U.S. adoption aligning with ISO 10218, requires risk assessment and documentation	Updates pending to fully align with ISO 10218-1:2025; end-user guidance available via TR R15.706 (2019)

Operational Modes