Collaborative robot implementation can help manufacturers automate selected tasks while retaining the flexibility and judgment of production staff. But a cobot is not a productivity improvement simply because it can share a workspace with people. The work must be analyzed, tasks allocated appropriately, and the effect on the entire production flow assessed. A case study at Ericssonâs BorĂ„s facility illustrates a practical way to evaluate these questions before committing to a production installation.
Start with the process, not the robot
Traditional industrial robots are often associated with dedicated, highly automated cells, while collaborative robots are designed for applications in which people and robots can work near one another. Human-robot collaboration (HRC) can take several forms: people and robots may occupy the same area at different times, work side by side on separate tasks, or work simultaneously on the same component. The appropriate arrangement depends on the operation and its risks.
The potential value lies in combining different strengths. Robots can perform repeatable handling or assembly tasks, while people can contribute flexibility and dexterity. That combination is particularly relevant where products or production requirements vary. However, a collaborative application still requires task-specific risk assessment and suitable safeguards; the âcollaborativeâ label alone does not establish that a workstation is safe.
Planning also needs to span several organizational levels. Production staff can identify operational needs and constraints, management must define scope and resources, and line- and workstation-level teams need to determine the layout and division of work. Including operators in planning can also support understanding and acceptance of a changed work process.
A practical framework for collaborative robot implementation
The Ericsson study combined workplace observation, work-element analysis, time studies, task assessment and robot simulation. For manufacturers considering a similar project, these activities can be organized into a connected pre-study:
- Observe and document the work. Visit the production area and break each stationâs work into clear, ordered tasks. A process map can show how the product moves through operations, inspections, handling and storage.
- Measure task and station times. Record repeatable work elements across multiple cycles and, where possible, different operators. Use the results to estimate standard times and understand how much of each cycle is manual work. Measurements should reflect actual production conditions.
- Assess task suitability. Review factors such as component size and weight, sensitivity, gripping requirements, assembly motion, presentation to the robot, fastening method and safety considerations. A structured scoring method can help compare tasks, but judgments about factors such as component sensitivity may still vary between assessors.
- Allocate work between people and robot. Use the task analysis to identify what a robot could perform and what should remain with an operator. Consider the sequence and interaction between tasks, not only whether an individual component can be picked up.
- Validate the proposed work with simulation. Test the allocation using computer simulation or a controlled trial with a robot. This can expose practical constraintsâsuch as an unsuitable gripperâthat a desk-based assessment may miss.
- Check cycle time, takt and bottlenecks. Compare the proposed station cycle with production requirements and identify the constraint that governs flow output. A slower robot cycle may reduce the lineâs output if it creates a new bottleneck, even when it reduces manual effort.
- Review the whole cell and prepare for operation. Map the revised process, assess workstation safety, organize tools and materials, define standard work, and plan replenishment. These implementation activities are necessary beyond the productivity-focused pre-study.
Why the bottleneck changes the productivity calculation
Productivity relates output to input. Total productivity accounts for multiple inputs, while a partial measure focuses on one resource. The Ericsson analysis used labor productivity: units produced relative to manual labor. This measure suited the studyâs focus, but it does not capture every resource or cost. If equipment, energy, materials or other inputs change, those effects need separate consideration.
In a production flow, output is constrained by the bottleneck. Automating a station that is not the bottleneck may reduce labor without increasing the number of units the line can produce. Conversely, if a robot slows a station enough to make it the new bottleneck, overall output could fall. The robotâs cycle time therefore has to be evaluated in the context of the entire process, not in isolation.
What the Ericsson case indicated
The case examined a radio-unit production flow that included board testing, assembly, calibration and other tests. The study measured station cycles and manual work, then estimated the effect of assigning selected tasks to a collaborative robot. Temperature testing was identified as the flow constraint, and the assessment did not find that station suitable for the robot under the conditions studied.
Assembly offered the largest modeled labor-productivity opportunity. The analysis estimated that a robot could perform most assembly tasks, with labeling remaining manual. Under the studyâs assumptions, placing a robot at that station could reduce manual work per product and increase labor productivity by 85.7%, while line output remained unchanged because the bottleneck was elsewhere. This was a theoretical estimate based on time studies and simulated tasksânot a measured result from a completed production installation.
The estimate also depended on how released labor would be used. If an operator still had to attend the station for every product to apply a label, the labor benefit could differ from a scenario in which labeling was organized in batches or the operator performed other productive work. The studyâs time scope was limited to selected station activities and did not include factors such as material replenishment or transport between stations. These boundaries matter when interpreting the result.
Turn a promising estimate into an implementation decision
A sound cobot business case should distinguish labor savings from output gains and make its assumptions explicit. Confirm that the task allocation works with the intended end effector and component presentation; verify the revised cycle against takt and bottleneck capacity; and assess safety, operator training, material supply and standard work before deployment. Select productivity measures that match the business objective, and avoid treating a single labor metric as a complete measure of operational performance.
The practical lesson is straightforward: collaborative robot implementation is a process-improvement exercise as much as an automation project. Systematic observation, measurement and validation can reveal where a cobot may reduce manual effort, while process-level analysis helps prevent local gains from creating new constraints. The strongest decision is based not on the robotâs theoretical capability, but on a verified fit between task, workstation, workforce and production flow.
Access Official Technical PDF