Skip to content Skip to navigation

Robotics Evaluation Framework - 2nd Year

Robots on construction sites

Project Team

Martin Fischer, Cynthia Brosque

Research Overview

Observed Problem:

The robotization of construction may happen on-site with single or multi-task robots that perform construction tasks. However, we must also consider the possibility of prefabricating elements off-site with the aid of robots in predictable environments, similar to industrial robotics in other manufacturing industries (Bock, 2004). These elements are then transported and assembled on site. Since some robots will prove beneficial in some circumstances and prefabrication or manual traditional methods will be preferred in others, innovation leaders in construction should be able to determine which method is better and under which conditions.

Primary Research Objective:

Building on a case study comparing on-site manual and robotic drywall installation, we propose to develop safety, quality, schedule, and cost analyses for off-site robotic prefabricated drywall systems.

Potential Value to CIFE Members and Practice:

  • An in-depth case study comparing off-site, on-site and traditional construction methods.
  • Development of metrics to determine the success of on-site and off-site robot uses.

Research provides relevant insights for:

Site managers will find relevant insights on the comparisons results and innovators on construction companies will find value in the metrics evaluated and the adoption challenges.

Research and Theoretical Contributions

  • An analysis of product, organization, and process boundaries that need to be crossed to successfully deploy robots on site versus off-site.
  • A robotics evaluation framework to consistently guide practitioners in the AEC industry.

Industry and Academic Partners:

Veidekke, Build-R, DPR Construction, DBC.


robotics; prefabrication; process; comparative case study; BIM

Research Updates & Progress Reports

Detailed Research Overview & Progress Updates - April 2020

Case Study Method

To supplement previous cases focused solely on on-site robots or off-site prefabrication, this project developed a case study comparing on-site and off-site robots to the traditional method of drywall installation. The Swedish GC, Veidekke, was interested in evaluating robotics for a high-rise residential project in Lund, looking at the drywall process and production data, schedule, cost analysis, physical effort, coordination, and non-value adding tasks. Hence, we used quantitative and qualitative data sources to compare the three scenarios.

Quantitative data sources came from measuring the drywall performance of four workers on-site over the course of a week (25-29th May 2019), the traditional project schedule, cost estimates, the robot productivity measured a laboratory setting, and off-site robot performance reported by the manufacturer and confirmed in one project using the prefabricated system.

Qualitative data sources included observations captured through site and lab visits (May 2019), pictures, videos, interviews, and email communications with relevant stakeholders: Site managers, four crewmembers, and the VDC Manager (October and November 2018, and February, March, and April 2019), two engineers from the robot start-up (October and November 2018, and February, March and April 2019), the Design Manager and Product Manager from the prefabrication manufacturer (June, October, and December 2019, and February 2020), and a Project Engineer and Project Manager in a construction site that deployed the prefabricated system (April 2020). We also gained access to the traditional project BIM, 2D plans, and drywall specifications.

The research steps consisted on:

  • Establish the key comparison metrics based on the previous on-site robots and manual work comparisons: product, organization, process, safety, quality, schedule, and cost.
  • Develop a well-documented comparison case study based on the available data and data captured in the field.
  • Revise the comparison with industry partners and determine if the approach to compare on-site robots to traditional work could also be used to evaluate off-site robots.
  • Develop recommendations for construction companies looking to implement robotic drywall placement on-site, off-site robot solutions, or a combination of both.

Independent variables (Product, Organization, and Process) analysis for off-site robots

In this section, we introduce the off-site robot Product, Organization, and Process as these are the variables that the project teams can decide and influence.

Product: three off-site robots to produce wall partitions
The Arizona prefabrication factory understudy houses five parallel production lines that can be deployed together for one project or work for different projects simultaneously (Figure 1).

Figure 1. Off-site factory layout.
Each line in the factory includes five steps: roll forming, subassembly welding (for parts that require welding on both sides), wall line welding, floor line welding (for parts that require welding on both sides), and sheathing line. These steps entail three robots each: (1) a roll forming machine, (2) a welding robotic arm, and (3) a sheathing line robot.

  1. Roll forming machine: The roll forming machine (Table 1) produces light steel drywall tracks and studs’ structure from galvanized steel sheets of 0.4 to 1.2mm. The hardware includes a cutter blade (50mm diameter), carbon steel roll shafts, leveling screws in the base with metal plates, and a cast-iron structure. The robot is fixed to the ground with the material rolled horizontally. The payload capacity is 3000kg with a speed of 20 meters per minute.
    Table 1. Roll forming machine
  2. Wall line welding with robotic arm: the second robot, a Kawasaki arc welding arm, was mounted on a fixed platform for wall line welding (Table 2). This 6 DOF robot has a rotation of 180 degrees and a real-time path modification sensor. Adaptive laser vision is used to track in real-time the joint geometry before and during welding. The robot may carry up to 10kg and weighs 150kg. Its reach is 1.4m and has a repeatability of 0.03mm.
    Table 2. Robotic arc welding
  3. Automatic sheathing line robot: The last robot fastens the sheathing to the drywall panels (Table 4). The panels are first rolled and clamped to the machine. Then, the operator uses a nail bridge to fasten the sheathing to the wall. Walls could be up to 4.9m width, between 1.2m to 3.6m tall, and 31.75cm maximum wall and sheathing thickness. The hardware consists of a sheathing open table with a movable section mounted on a track on the floor and roller conveyors to roll panels in and out. The machine includes a platform to support the operator and a nail bridge. The bridge inserts fasteners at a preset spacing, controlled by a touchscreen and joystick with an auto-return mechanism for easy access to remove the wall section.
    Table 4. Automatic sheathing machine (“CN203003437U - Lifting buckle automatic sheathing machine - Google Patents,” 2012)

The volume of partitions necessary to be worth using digital fabrication ranged from 100 panels to 40,000 sqft (best). Although the company has completed smaller projects in the past, they only fabricate and reduce their engagement in the coordination phase for these projects to be viable.

The shipment distance was another variable to determine the feasibility of prefabrication. The factory, located in Arizona, reached four main regions in the US: Northwest, Southwest, Central, and Southeast.

From the construction site perspective, we could evaluate the location, partition volume, the construction task, geometry complexity, use of standard materials, and site conditions required to use prefabrication (Table 5). We will gather production data from interior partitions with electrical rough-ins from one project to evaluate how prefabrication compared to traditional construction.

Table 5. Construction site characteristics

Organization: integration of factory and site
The organization to prefabricate wall panels integrates the prefabrication facility organization and the construction site organization. From the prefabrication facility, the organization includes in-house detailers, structural engineers, designers, a foreman running the shop, and the plant workers. From the construction side, the coordination involves the architects, MEP engineers, and the GC, especially the BIM coordinator.

Process: six phases from design to finishes
The process to use off-site robots in prefabrication involves six main phases: design, coordination, fabrication, ship to the site, installation, and finishes (Figure 2).

Figure 2. Off-site robot process.

The prefabrication team usually gets involved four to six months before the installation on site to supply input on the BIM at LOD 300 and identify which building elements can be more efficiently delivered through factory fabrication.

The coordination phase involves collaborating with the trades to model the prefabricated components in Revit so that these fit perfectly into the building. Once the BIM is locked the factory can move to production. The company employs a proprietary software to spool directly from the digital model to the automated equipment. Digital fabrication drives the whole process of prefabricating walls, including the location of punches and welds. The shop does not use paper, and instead, each product holds a bar code that was scanned to visualize the assembly information.

Fabrication involves the three discussed robots straight from digital design and bulk sheet-metal coil to a welded structural wall or interior partition. As observed from the manual process, the construction of stud walls happened in three phases: layout of the top track, installation of stud framing, and gypsum board sheathing and finishing. The prefabricated method manufactured wall panels, complete with framing for MEP openings, considering constructability issues and coordination with other building systems like HVAC ducting, plumbing, and fire protection. This differs from the traditional construction process where MEP openings were framed on site. It also requires that the installation of all systems happened as per the BIM.

The last steps of the process are shipping, installing, and finishing the walls on-site. Interior wall panels are introduced with the aid of a crane and then installed manually on the site.

The process challenges include design changes after the BIM is already frozen for fabrication, handled on a case-by-case basis according to the time in which the change occurred and the nature of the change.

Dependent variables: Safety, Quality, Schedule, Cost.

On a second step, we analyzed the safety, quality, schedule, and cost impacts of off-site robots.

From a safety and ergonomics perspective, off-site digital fabrication enabled key components of the light-gauge framing to be produced together in a safe and controlled environment. 85% of the work was performed in safe and controlled factory settings, while only 15% of the hours were spent in strenuous tasks on site because workers still need to place manually the finished interior wall panels on the right location.

Second, prefabrication could improve accuracy by precise control of the functions at an off-site facility. The welding robot has a repeatability of 0.03mm while the automatic sheathing eliminates the issues associated with screw heads and build-up at openings. The prefabrication company performed quality assurance on the equipment to ensure consistent production standards and factory workers inspected every panel as it left the factory. To adjust to on-site deviations from the BIM, the prefabrication company built-in tolerances of 3.17mm (1/8th”) into the walls’ dimensions.

Rework with off-site robots could be reduced by avoiding manually measuring and cutting panels. According to the prefabricated company, the interpretation of 2D details, measuring, and cutting are the source of more than 95% of rework in construction. In digital fabrication, the robots used lasers for accurate placement straight from the coordinated BIM. Finally, sheetrock waste in prefabrication was reported as null by digitally optimizing stud spacing. The company tracked its exact stud waste when punching elements and reported virtually no waste of studs.

The main schedule benefit of prefabrication is starting the wall construction in the factory before having a finished concrete slab on site. We will quantify non-adding value tasks such as the impact of coordination and transportation for a specific project to determine how this influences the project schedule.

Off-site robots have estimated overall similar costs or slightly less than existing construction methods as they required more coordination, modeling at a higher level of detail, and transportation, but reduced schedule days, eliminated interpreting shop drawings, measuring, and cutting in the field, where labor costs are highest. We still need to obtain further data on the specific costs for the project with interior wall partitions and electrical rough-ins.

Anticipated comparison results
So far, the comparison of traditional methods with on and off-site robots used for drywall installation showed that each solution could be preferred depending on the circumstances. For example, in a residential project in Sweden, the on-site robot would have been too big for the apartments’ clearances. However, a commercial building with full-size boards could benefit from having the on-site robot to extend the working hours.

The prefabricated method showed potential safety, quality, and schedule improvements but would still require a solution to place the interior partitions on site. Exterior prefabricated walls, different from interior walls, could be placed on site by crane, providing safety benefits for the workers with a superior quality product. The prefabrication process requires detailed coordination in BIM which takes approximately four to six months before construction. The prefabricated company has already gotten repeat clients with high-rise residential projects and complex healthcare industry buildings which speak well of the process developed.

Additionally, robots both on and off-site supply progress information useful to the project beyond the task. Much of the manual labor time associated with installing drywall was spent looking for materials and information, interpreting paper drawings, and taking measurements. The direct nature of communication between a digital model and a robot allows for a seamless information integration that provides a feedback loop to the progress report.

Finally, the manual work of the four workers under the course of a week proved to be highly variable. Reducing site uncertainty with predictable production rates, especially in a key task like drywall installation, could be a potential benefit beyond the safety, quality, schedule, and cost impacts for on and off-site robots

Detailed Research Overview & Progress Updates - December 2019

Overview & Observed Problem

Robots have been used in several fields to ensure workers safety while increasing productivity and quality. Current advances in sensing and computing technology are now enabling the application of robotics on construction sites, outside of the controlled environment of factories. One example is a drywall placing robot developed by Build-R to reduce the manual strain of lifting heavy boards. To understand robotic feasibility, we can compare its performance to the traditional manual work (Kumar & Leena, 2008). However, we should also consider the possibility of prefabricating elements off-site with the aid of robots in predictable environments.

Prefabrication can help deal with the uncertainties of introducing robotics in construction by creating an off-site production environment, similar to industrial robotics in other manufacturing industries (Bock, 2004). The prefabricated elements are then transported and assembled on site. According to Bock (2004), it is likely that in the future automatic production in factories and mobile robots on-site will be integrated. However, the author adds that it is unclear yet where the division line between both should be and whether one method will be preferred over the other.

Since some robots will prove beneficial in some circumstances and prefabrication or manual traditional methods will be preferred in others, innovation leaders in construction should be able to consistently evaluate which method is better and under which conditions.

Theoretical & Practical Points of Departure

Several authors have compared on-site robotics to manual work focusing on variables such as productivity, cost, quality, and safety (Warszawski & Rosenfeld, 1994; Balaguer et al., 1995; Kangari & Halpin, 1989; Sousa et al., 2015; Usmanov et al., 2017).

Literature has also addressed the use of robots in prefabricated off-site facilities and its effects on resource-allocation, schedule, and quality (Kangari & Halpin, 1989; Pan et al., 2018; Saidi et al., 2016; Warszawski & Sangrey, 1985). According to Kangari and Halpin (1989), the similarities of construction prefabrication to industrial production make it a logical first area to focus the efforts of automation and robotization. Koskela (2000) states that prefabrication can alleviate site problems by reducing the number of activities carried out on-site. Moreover, robotics combined with CAD-CAM have enabled extensive prefabrication, flexible modularization, new standard building materials and components, and quality improvements (Saidi et al., 2016; Warszawski & Sangrey,1985). Other benefits of prefabrication are waste reduction and flexible working conditions (Pan et al., 2018).

Despite this existing literature on prefabrication benefits, we have not found a comprehensive case that compared off-site robotics in prefabrication, on-site robotics, and traditional manual construction for a specific construction task.

A drywall system is one example we can study and compare for the three scenarios. Digitally fabricated interior and exterior panels made robotically are commercially available with examples from Digital Building Components, Katerra, Blokable, Entekra, Boklok, and Williams Robotics. Skanska has also developed near-site factories to assemble wall panels (FRAMBE  project, 2017). Likewise, on-site drywall placing robots are now available, with examples Sweden (Build-R) and Japan (Shimizu).

Cynthia Brosque (the graduate research assistant proposed for this research project) is carrying out ongoing research to determine the impact of three single-task robots compared to the traditional manual methods for three construction tasks to formalize an evaluation framework for robots. One of these case studies compared the schedule, cost, quality, and safety impacts of a drywall-placing robot to the manual drywall installation on-site. The project selected for the manual study, a high-rise residential project in Lund built by Veidekke, combined traditional drywall with prefabricated bathroom units delivered to the site. As observed, GCs can combine different solutions based on cost, time, and quality goals.

The anticipated conclusions of this case study led us to question how does the introduction of on-site robotics compares to off-site prefabrication? For example, we need to know how does prefabrication affect the product, process, and organization? How could it pay off and under which conditions? What is the potential schedule, cost, and quality impact? What are the implications for the workforce, particularly regarding safety? How does it affect decision making, for example during the design phase?

Research Methods & Work Plan

We propose to answer these questions with a quantitative and qualitative study comparing the three scenarios: prefabrication with off-site robotics, traditional construction, and construction robotics on-site for drywall systems. We will assess schedule, cost, safety, and quality impacts and address product, organization, and process boundaries faced for each of the three scenarios.

The proposed research will develop as follows:

  • Conduct interviews and site visits with relevant stakeholders: GCs, manufacturers, designers, and key subcontractors.
  • Gather prefabricated project data through direct observation, 3D/4D models, schedules, and cost estimates.
  • Contrast the data to traditional and on-site robotic construction. Revise comparison metrics with industry partners.
  • Study product, organization, and process boundaries for the successful implementation of the prefabricated system.
  • Determine schedule, cost, quality, and safety impacts.
  • Develop recommendations for AEC practitioners looking to implement robotic drywall placement on-site, prefabricated systems, or a combination of both.

Expected Contributions to Practice

  • An in-depth case study comparing off-site, on-site and traditional construction methods.
  • A comprehensive analysis of the possible challenges of robotic adoption on and off site and the metrics that determine its feasibility versus traditional methods.

Expected Contributions to Theory

  • An analysis of product, organization, and process boundaries that need to be crossed to successfully deploy robots on are off-site.
  • A robotics evaluation framework to consistently guide practitioners in the AEC industry.



Abraham Warszawski, B., & Rosenfeld, Y. (1994). Robot for interior-finishing works in building: feasibility analysis. Retrieved from

Balaguer, C., Pastor, J. M., Garcia, A., Pe, L. F., Rodriguez, F. J., & Barrientos, A. (1995). Evaluation and Comparative Study of Robotics vs . Manual Spraying of GRC Panels . Automation and Robotics in Construction, XII, 489–497.

Kangari, R., & Halpin, D. W. (1989). Potential Robotics Utilization in Construction. Journal of Construction Engineering and Management, 115(1), 126–143.

Kumar, V. S. S., & Leena, A. (2008). Robotics and Automation in Construction Industry by. AEI 2008: Building Integration Solutions, 1–9. Retrieved from

Pan, M., Linner, T., Pan, W., Cheng, H., & Bock, T. (2018). A framework of indicators for assessing construction automation and robotics in the sustainability context. Journal of Cleaner Production.

Saidi, K. S., Bock, T., & Georgoulas, C. (2016). Robotics in Construction. In Siciliano & Khatib (Eds.), Springer Handbook of Robotics (2nd ed., pp. 1493–1519). Springer.

Sousa, J. P., Azambuja Varela, P., Martins, P. F., & Faup, /. (2015). Between Manual and Robotic Approaches to Brick Construction in Architecture Expanding the Craft of Manual Bricklaying with the Help of Video Projection Techniques. Retrieved from

Usmanov, V., Bruzl, M., Svoboda, P., & Šulc, R. (2017). Modelling of industrial robotic brick system. In 34th International Symposium on Automation and Robotics in Construction (pp. 1013–1020). Taipei: International Association for Automation and Robotics in Construction.

Warszawski, A., & Sangrey, D. A. (1985). Robotics in Building Construction. Journal of Construction Engineering and Management, 111(3), 260–280.

Funding Year: 
Stakeholder Categories: