Collaborating robots

Introduction

The use of collaborative robots, (cobots), is steadily increasing across the manufacturing sector, enabling people and robots to work alongside each other. Cobots do not operate in isolation, but as part of a broader collaborative application that refers to the entire system in which humans and robots interact. By combining human adaptibility and dexterity with the precision, strength and speed of the robot, production can become more efficient  while reducing the workload of the workers. On the other hand data from EU-OSHA’s OSH Pulse survey suggest that the use of cobots is associated with heightened work intensity, reduced autonomy, increased surveillance and a higher prevalence of working alone^[1]. Further to the psychosocial risk, mechanical and ergonomic risks could also be prevalent. Therefore, ensuring the safety of human–robot collaboration requires rigorous workplace and system design in line with relevant safety standards and human-centric design principles.

This article focuses on the risks and prevention measures associated with collaborative robots that operate alongside human operators in the industrial sector. According to the 2024 European Working Conditions Survey, 10% of respondents in the industry sector reported using cobots in their workplace^[2]. Similar applications exist in the non-industrial sector; these are not addressed here. This article also does not cover the specific risks associated with specialised robot applications such as those for welding or laser processes 

What are Collaborating robots?

A collaborative robot shares a common workspace with a human operator in order to carry out, in conjunction with the operator and at the same time, a previously defined task. The International Federation of Robotics (IFR) defines collaborative industrial robots, or cobots, as robots with power- and force-limiting functions that are designed to operate alongside human workers in industrial settings^[3]. 

Cobots do not operate in isolation, but as part of a broader collaborative application. A collaborative application is defined any application (process) containing at least one collaborative task, that is, at least a portion of the robot sequence where both the robot application and operator are within the same safeguarded space^[4]. This concept encompasses the entire system in which humans and robots interact, including task design, workspace layout, control systems and organisational measures.

Onnasch et al. distinguish three main types of human–robot interaction (HRI)^{[5] [6]}:

• Coexistence: humans and robots share the same workspace but not a task. For example, a worker passing a mail delivery robot in a hallway.
• Cooperation: humans and robots pursue a shared goal, but their tasks remain independent. An example is a pick-and-place robot preparing parts for an assembly worker.
• Collaboration: humans and robots not only share a goal but also divide subtasks in time and space, creating synergies. A typical case is jointly lifting and positioning a heavy object^[6].

The IFR further recognises a fourth level, responsive collaboration^[3], in which robots react to human movements in real time, for instance, adjusting their trajectory as a worker moves within the workspace.

Across all these scenarios, cobots are characterised by their ability to sense and respond to their environment. Some rely on complex deterministic programming, while others increasingly integrate AI-based systems^[6] such as machine learning and computer vision to enhance adaptability, perception, and decision-making in dynamic environments.

According to the IFR^[3], a large share of cobots in use today are still not engaged in true human–robot collaboration. Most current applications involve shared workspaces where humans and robots work side by side but perform tasks independently or sequentially (coexistence). Common uses include lifting heavy parts, handling awkward loads, or carrying out repetitive tasks such as screw-tightening. Over time, however, a shift is underway toward deeper cooperation and real collaboration, where human skills and robot capabilities are fully integrated^[3].  

Compared to traditional industrial robots, collaborative robots offer several advantages^[7], including greater flexibility, lower cost and more straightforward installation, commissioning and reallocation processes. Collaborative robotics reduce physical effort at work and, as a result, can help to reduce work-related musculoskeletal disorders, stress and operator fatigue. However, their ability to interact with humans means that any potential risks of causing harm must be eliminated or minimised. For this reason, collaborative robots have a lower capacity to exert force or transport loads, a lower working range and a lower speed than traditional robots^[7].

Standardisation

Manufacturers placing collaborative industrial robots on the market must comply with the essential health and safety requirements for the design and construction of machinery laid down in the European Machinery Directive 2006/42/EC^[8]. Directive 2006/42/EC has been repealed by Regulation (EU) 2023/1230^[9] which will apply to machinery from 20 January 2027. 

Additional regulatory requirements may apply to collaborative robots that incorporate AI-based functionalities, as set out in Regulation 2024/1689/EU laying down harmonised rules on artificial intelligence^[10]. AI systems that are safety components of machinery or are used for safety-critical functions, for example, may be classified as high-risk. This requires risk management, data governance, transparency and human oversight measures.

Technical specifications for machinery are provided in harmonised standards. These standards translate the essential health and safety requirements into technical requirements. If a manufacturer uses a harmonised standard, the product is presumed to comply with the legal requirements. Industrial robots designed for collaborative operation must comply with the international standards set out in the EN ISO 10218 series^{[11] [12]}. The harmonised two-part EN ISO 10218 series of standards was developed for the specific hazards presented by industrial robots including those used in collaborative operations.

EN ISO 10218-1:202511 outlines the safety requirements for robots ISO 10218-2:2025 provides requirements for safeguarding operators during integration of the robots, their installation, functional testing, programming, operation, maintenance and repair. In Europe, EN ISO 10218-1/2 superseded European standard EN 775 in 2008 and detailed requirements for collaborative robots were included in ISO/TS 15066 specification^[13]. Revised versions of EN ISO 10218 were published in 2025, incorporating the requirements for collaborative robots that were set out in ISO/TS 15066.

Hazard identification and risk assessment

Since workers and collaborative robots share the same workspace, it is essential to carry out a comprehensive risk assessment before introducing cobots into the workplace. This assessment should consider not only the technical aspects of the robot (e.g. defining its spatial boundaries or installing sensors to detect human movement) but also organisational and psychosocial factors^[14].

Potential risks include^{[14] [15] [16]}:

1. Physical risks

• Impacts and collisions: unexpected robot movements, sensor failures, or programming changes can result in contact with workers. Injury severity depends on the force, speed, and contact surface. Collisions may also occur if robots fail to detect environmental factors (e.g. uneven floors), potentially causing tipping hazards.
• Entrapment and crushing: workers may become caught in or between moving parts of the robot or between the robot and fixed structures.
• Tool-related hazards: risks linked to the weight, shape, sharpness, or movement of tools used by the robot.
• Flying particles or objects: mechanical failures, defective grippers, or malfunctioning tools may release objects at high speed.
• Hydraulic and pneumatic hazards: pipe or system failures may cause sudden releases of pressure or fluids, leading to injuries, fires, or toxic exposure. Loss of pressure may also result in uncontrolled or collapsing movements.
• Electrical hazards: contact with live components or faulty electrical systems.
• Slips, trips, and falls: caused by leaked fluids, cables, hoses, or poorly organised workspaces within the collaborative application.
• Musculoskeletal disorders (MSDs): may arise from repetitive movements, awkward postures, or increased work pace when working alongside robots.

2. Organisational and system-related risks

• System failures: incorrect programming, software errors, cybersecurity vulnerabilities, electromagnetic interference, or degradation due to dust, moisture, or wear and tear.
• Poor human–robot interface design: unclear signals, lack of transparency, or poorly designed controls can lead to misuse or delayed reactions.
• Insufficient training and competence: workers may not fully understand robot behaviour, limitations, or emergency procedures.
• Maintenance and integration failures: inadequate inspection, servicing, or integration of components (including sensors and safety systems).
• Work organisation factors: increased pace of work or poorly planned workflows in human–robot collaboration.

3. Psychosocial risks

• Increased workload and work pressure: due to higher productivity demands or coordination with automated systems.
• Technostress: stress related to adapting to new technologies, particularly when systems are complex or unreliable.
• Isolation and reduced autonomy: increased monitoring, reduced decision-making possibilities, or more frequent lone working.
• Trust-related issues: too little trust (e.g. caused by false alarms or malfunctions) can lead to anxiety, distraction, and resistance to using the technology; too much trust may result in overreliance and reduced vigilance, increasing the likelihood of accidents.

Hazardous design characteristics

Where a collaborative robot is used, consideration must be given to the anticipated unobstructed accessibility by persons during collaboration with the robot. This may result in contact between the human operator and the robot. This may be intentional during performance of the work task, or be a result of foreseeable misuse, owing for example to guards having been replaced by other protective measures. It is therefore particularly important that the collaborative task be described and specific hazards identified from the description. Collaboration may be characterised by:

• Frequency and duration of presence of operators or other workers in the collaborative space whilst the robot is energised
• Frequency and duration of contact between operators and the energised robot (e.g. hand guiding, handover of the tool or workpiece)
• Restarting of the robot following violation of the minimum separation distance or following contact
• Switching of the robot between collaborative and autonomous operation when the operator enters or leaves the collaborative space
• Collaboration with several operators simultaneously
• Energise robot arm involuntary movements
• Highest incidence of contention between human and robot is when both are accessing the same object

Workplace prevention measures

Based on the results of the risk assessment, preventive measures must be put in place to ensure the health and safety of workers. Key measures include^{[14] [16] [17]  [18]}:

• integrate OSH early in planning and design: assess which tasks involve human–robot interaction, evaluate the impact on task/job design, and eliminate or minimise hazards before implementation.
• consult and involve workers: engage workers early in the process, gather feedback, and address concerns related to job security, task changes, or organisational impacts. Building confidence and trust in the technology is essential.
• ensure technical compliance: verify that cobots meet essential health and safety requirements (CE marking), and ensure that instructions for use, declarations of conformity, and technical documentation are available and accessible.
• apply relevant standards: comply with safety requirements laid down in EN ISO 10218 (see below technical requirements)
• define roles and responsibilities: establish clear agreements on who may program, operate, and maintain the cobot; clarify responsibilities for supervision and emergency procedures.
• cybersecurity measures: protect cobot systems against hacking, malware, or unauthorised access that could cause malfunctions or safety hazards.
• clearly mark and restrict access to cobot operating zones; use sensors, barriers, or light curtains where necessary to manage safe distances.
• emergency measures: provide accessible emergency stop buttons and ensure clear procedures for shutdown in case of malfunction.
• maintenance and inspections: establish regular maintenance schedules, inspection routines, and safe procedures for technicians; document all checks and repairs.
• training and information: train operators, maintenance staff, and other stakeholders on safe working practices, potential risks, emergency actions, and psychosocial aspects such as adapting to new work routines.
• monitoring and continuous improvement: regularly review incidents, near-misses, and worker feedback to adjust work processes, training, and technical safeguards.
• ergonomic considerations: ensure that cobots do no increase physical strain e.g. awkward postures or repetitive tasks, take into account cognitive ergonomics, design human-machine interface
• psychosocial support: address possible increases in workload, technostress, or changes in work pace through supportive measures and good work organisation.

Technical requirements

Requirements for collaborative applications

Robots for collaborative operation must satisfy the requirements of EN ISO 10218-1 and EN ISO 10218-2 including the requirements for functional safety performance levels for safety functions. 
This particularly applies to the safety functions described below for the various collaboration methods. These performance levels are detailed in annex C of the standards.
Hazards arising during collaboration can be avoided by reduction of interaction and of the collaborative space, to the extent reasonable in consideration of the task. Any unnecessary approaching of the robot must be avoided. 
The collaborative space must be clearly defined and demarcated. In order to provide sufficient space for the robot to move, adequate clearance must be ensured and suitable protective equipment provided in order to prevent crushing against parts of building structures or machinery. If cobots can both operate autonomously (the robot executes a programmed task without human presence in the safeguarded space) and collaborative, switching between these modes is considered a safety-critical function, and must be managed through the control system and risk assessment.

Safeguarding

During collaborative operation, operating personnel are protected by a safety function or a combination of several safety mechanisms. The operator must be able to stop the robot movement by a single action. Protective measures must be available to all operating personnel within the collaborative space. The following safety functions are typical:

• Stop functions: each collaborative workplace must be equipped with an emergency-stop and a protective-stop function for the robot.
• Enabling device: a robot movement occurs only following an enabling action by the operator (pushing a button for example during hand guiding).
• Limited collaborative speed: a maximum permissible robot speed is determined by the risk assessment. This limit value is stored in the control system, either as a fixed value or as a variable dependent upon the distance from the operator, and is monitored continually.
• Minimum separation distance: a sensor system continually detects the position of the operator and his or her distance to the robot. The required safe distance is typically determined dynamically from the relative speeds of the operator and the robot, the response times of the brake and control system, and the accuracy of measurement of the sensor system or the robot. If this safe distance is not maintained, a protective stop occurs.
• Limited force/pressure: a sensor system on or in the robot detects a contact between the operator and the robot and stops in order avoid exceeding the allowable force or pressure. A protective stop is then initiated and possibly reversal of the robot movement to enable the operator to free him or herself.

Methods of collaborative working

For safe collaborative operation, one or more of the following methods and safety functions must be selected as appropriate. Whenever a failure of the safety functions is detected, a safe stop must be initiated. Autonomous operation of the robot following a stop may be resumed only following a deliberate restart executed by the operator from outside the collaborative space.

Monitored-standstill This method does not allow for the operator and the robot to move simultaneously. Monitored standstill stops the robot's movement when a human enters the collaborative workspace. This standstill only halts motion; the robot's power is not completely removed. This is monitored by the control system to ensure that the robot remains stationary and does not restart accidentally. This means that the operator can now work directly with the robot, and for example fit a workpiece. As soon as the operator leaves the workspace, the robot can continue its autonomous movement and task. 

Hand-guided control

With this method, the operator and the robot can move simultaneously within the collaborative space, and also work closely together. This may be the method selected when performing some maintenance tasks. The robot however must not move autonomously, and must instead be guided manually by the operator. In order to position the robot manually, the operator must have a guidance device (push buttons, joystick) with an emergency-stop. The operator must have a clear view of the collaborative space and of the robot's movements. Safe limits are imposed upon the robot's speed and position. The position and posture of the operator and the guidance device itself must not give rise to additional hazards. As soon as the guidance device is released by the operator, the robot must stop again. As soon as the operator leaves the space, the robot can continue its autonomous movement.

Speed and separation monitoring

With this method, the operator and the robot can move simultaneously and autonomously within the collaborative space but cannot work closely together. The risk is reduced by a permanent safe distance between the operator and the robot. This safe distance can be determined dynamically based on the robot’s current speed. At low speeds, the safe distance may be reduced. The maximum robot speed, the minimum distance and other parameters must be determined by the risk assessment. If the safe distance is not maintained, the robot comes to a standstill. As soon as the operator moves away from the robot, the robot can resume its movements with assurance of the minimum separation distance. The safe distance can also be ensured if the robot avoids the operator in the collaborative space by taking an alternative path. Speed and separation monitoring systems rely on devices such as safety-rated laser scanners, depth cameras, or safety-rated radar to detect human presence and measure the distance to the robot.

Power and force limiting

With this method, the operator and the robot can move simultaneously and autonomously within the collaborative space and work so closely together that they may come into contact with each other. This risk is reduced by limitation of the force and power of the robot system in the event of contact occurring. The limit values for allowable force and pressure are determined and verified by a risk assessment performed for the specific workplace in question.

This method is limited to robots of special design employing safe materials, soft surfaces, cushioning and the absence of sharp edges. The reason is that it mostly needs a contact to trigger the protective-stop function. Therefore, all robot movements should be logical and predictable for the operator in order to reduce the probability of unintentional contact.

Biomechanical limits

ISO 10218-2 also incorporates biomechanical limits, such as permissible force and pressure, for safe human-robot collaboration. These limits specify the maximum force and pressure in the event of a collision. The standard outlines the procedures for measuring and validating these limits.

Ergonomic requirements, human factors

The collaborative application should be designed in accordance with ergonomic principles^[19],  This will ensure that operation and maintenance are safe, intuitive, and efficient, while reducing the risk of human error, physical strain, and cognitive overload. Interfaces should be clear, accessible, and adapted to the diversity of users, supporting user comfort, situational awareness, and reliable task performance. The working space in which a person and a collaborative robot may come into contact should be designed so that it does not restrict the person’s physical mobility of the person. The individual's perception, attention and thought processes must not be constrained or disturbed by the working environment and the collaborative robot. The beginning of the robot's movement and its course should be visible, predictable and logical. These characteristics assist in preventing contact between the operator and the robot and possible stress upon or loss of control by the operator.

Organisational requirements 

The suitability, in terms of their health, of persons who work with a collaborative robot and are exposed to a risk of collision should be determined at suitable regular intervals. These persons must be provided with appropriate training and regular instruction on the risks and emergencies and the safety measures which must be taken. This applies in particular to installation, assembly and test work, to set-up mode and during commissioning. The particular underlying conditions for the organisation of workplaces involving collaborative robots (such as working hours, breaks, first-aid kits, log books, etc.) must be examined and defined. Following an unintended collision, the fitness for work of the affected individual and the correct setup of the workplace must be examined.

Marking and instructions

Robots for collaborative operation and points of access to collaborative areas must be marked by a suitable warning sign and symbol. The collaborative space in which the operator works directly with the robot must be clearly set out and demarcated (for example by markings on the ground, signs, etc.). The instruction handbooks for machinery must also describe the collaborative aspect:

• Design and use procedures: type of collaborative operation, detailed instructions, maintenance, abnormal situations, system limitations, cautions/warnings, personal protective equipment (PPE)
• Criteria for operator capabilities: skills, training, physical limitations
• Operators must be trained regularly of the risks, procedures in emergencies and required safety measures associated with collaborative operation

Preventing psychosocial risks

The introduction of cobots in the workplace can lead to psychosocial risks. Operators may experience increased stress due to the robot’s characteristics (e.g. speed, size) or because of unexpected movements. Cobots can also reduce opportunities for inter-human communication, contribute to feelings of isolation, or trigger fears of job loss or de-skilling^[15] [16]. 

Organisational measures to reduce these risks include^[6] [16]: 

• Early worker involvement: engage operators and other workers in the planning and design phase, ensuring their concerns, needs, ideas, experience and practical knowledge are taken into account.
• Feedback: create regular opportunities for workers to provide feedback on their experience with cobots and adjust systems or processes accordingly.
• Clear task allocation and job design: ensure cobots complement human strengths and that jobs are designed to be meaningful and varied, allowing for social interaction.
• Training and support: Organise training aimed at improving skill sets, building trust, and raising awareness of risks and prevention.
• Supervision and monitoring: ensure appropriate and permanent oversight of cobot use, addressing issues quickly and transparently.
• Communication: organise regular, transparent communication to address issues such as fears about job security and explain how the technology impacts and benefits workers.