J

Robotics Feasibility in the Construction Industry

By: Roozbeh Kangari

I. INTRODUCTION

At the present time, robotics in construction industry is in the early

stage of research. Before robots can practically be implemented in the

industry, major problems such as: how construction processes can be

automated; what are the sequential stages in construction automation; and

what level of automation is feasible for a given construction operation

should be investigated. Unlike the manufacturing industry, a construction

site is a dynamic and random environment, therefore, a fully automated

process requires a very intelligent control system, sophisticated sensors

for feedback, an efficient material-handling system, and an advance

mobility system. Under these conditions, it has become useful to explore

at least those problems in the range of preliminary steps for the robotics

feasibility in the construction industry.

Robot is used extensively by the manufacturing industry. However,

construction industry has unique characteristics which makes the

robotization in most cases not a feasible alternative at the present time.

It is not expected that robots enter the construction trades before the end

of next decade . The construction site is a random environment requiring a

robot of highly sophisticated intelligence combined with a large load range

and need for mobility. It seems that in the early days of robotics

application in construction industry, the awareness of construction site

hazards to labors will provide the prime motivation to design and use a

robot that would perform the tedious, repetitive, boring, dangerous and

unpleasant construction jobs.

66
 Robot technology is not new , but many industries as construction are

only just beginning to realize the impact that full automation could have

in their production. Today, construction robots are still on the stage of

research , and there are only few practical construction robots developed in

the U.S., Japan , and some other countries. However, among all these robots

only one or two may be called real construction robot, and the rest are

partially automated construction equipment.

Although today there is differences of opinion about exactly what a

construction robot is, but in general may be defined as a fully automated

mechanical device that can be programmed to perform construction tasks. In

other words, robots are the machines which are controlled by computers.

A further essential question is the determination of an economical and

practical level of automation for construction processes. There should be

an optimum level of automation for each type of constructi.on operation

since excessive application of automation to a given process may not be

economical . In certain cases, partial automation or robotization may even

lead to an increase in the unit price. One approach to this question is to

develop sequential stages in automation and perform a feasibility analysis

for each stage.

SEQUENTIAL STAGES IN AUTOMATION:

To define an optimum level of automation for a given construction

operation, the following five basic classifications as shown in Fig. 1 are

developed:

1) Pure manual labor construction operation which involves no

tools, e.g., material handling by hand, or packing.

2) Manual labor construction operation with tools, e.g., manual

67
 f

1) Pure Manual Labor Construction Operation

2) Manual Labor with Tool Construction Operation

3) Man-Machine Construction Operation

(Conventional Construction Equipment)

4) Man-Machine-Computer Construction Operation

(Partially Automated Construction Equipment)

5) Machine-Computer ( Robot ) Construction Operation

(Fully Automated Construction Equipment)

Fig. 1. - Sequential Stages in Construction Automation Process

68
 excavation with a shovel.

3) Conventional construction equipment , or man-machine operation.

These are the construction machines which are controlled by human,

e.g., drilling rock by a conventional drill, or excavation by a

conventional loader. Most of the construction equipment at the

present time are under this classification. Fig. 2 represents a

simple graphical model of this stage.

#) Partially automated construction equipment, or man-machine-

computer operation. As shown in Fig. 3, this stage of automation

improves the conventional construction equipment by adding a

partially automated control system to the actuators, e.g., laser

leveling grader, automatic gear shifting scrapers, hydraulic

excavator with bucket tilt control, or remote control construction

equipment for the construction work in dangerous places.

5) Fully automated construction equipment (robot), or machine-

computer operation, e.g., SSR-2 spray robot for fireproof spraying

on steel structures, developed by the Research Institute and

Construction Machinery Division of Shimizue Construction Co. in

Japan. In the U.S., the Civil Engineering and Construction

Robotics Laboratory at Carnegie-Mellon University is heavily

involved in research and development of the construction robots to

perform tasks in environment that are unsafe for human. These

robots require occasional human involvement as shown in Fig. u.

How does a robot operate ? Essentially the computer of robot is

provided with information representing a model of the robot, with details

of the environment, data relating to the tasks to be performed and with a

number of planning algorithms. When in operation it receives at all times

69
 1

Information Production
Human Operator
04 E---30
Machine Output
from Operating
Environment

Fig. 2. - Conventional Construction Equipment

Information
from Operating Human Operator
Environment

Display Panel

Informatio
Data Input
Sensors Controller Actuators

H Production
777 tput

Fig. 3. - Partially Automated Construction Equipment

70
 J
J

Occasional
Information Human
Involvement
rl

Information
Data input Sensors Controller
1' Actuators
Mechanical
Devices
Production
- i tput

Fig. 4. - Fully Automated Construction Equipment (Robot)

71
information concerning the robot with internally sensed information, and

the environment with externally sensed information. By using this

information in conjunction with planning algorithms, which can refer back

to past experience, the computer develops control over the robot, causing

it to move towards the correct execution of the task assigned to it.

The main difference between a construction robot and a conventional

construction equipment is that the robot is able to react with its

environment without a human intervention. However, the publicity

surrounding the introduction of robots into construction field exaggerates

the true state of the theoretical and practical knowledge of robotics. The

technical challenge is considerable because, at present, the

characteristics of robot are far from attaining the performance required in

an unstructured and dynamic construction field.

Large construction companies with an interest on equipment automation

have not given a great deal of attention to research in robotics. There

are only a few international contractors who have introduced robotics into

their field, however, these robots are not capable of detecting the complex

information directed to them from the environment.

If the number of repetitive operations are very large and the output

product is fixed, then it might be economical to implement a fixed

automation plant. For example, if a prefabricated plant is planning to

build a large (infinite) number of fixed construction products (e.g.,

prestressed concrete beams) which does not require any change in size or

type of material, then a fixed automation may reach a lower unit price than

a flexible automated plant. This is due to the large volume of production

and a lower variable cost.

72
 Considering these sequential stages, the objective of this paper is to

describe the feasibility of the last stage ( robotization) in relation with

the other stages. In other words , what construction operations should be

robotized.

73
 J J J

II. FEASIBILITY ANALYSIS

A modelling procedure is needed to evaluate the feasibility of

developing robotics and justifying the implementation of robotics for

certain construction operations. Two feasibility models using seven basic

variables are proposed. The ultimate output of the models indicate whether

the robotization is appropriate for a particular construction operation or

not. In reality, robotics feasibility and justification is inter-

disciplinary since it involves the input of several professional groups,

therefore, this paper can only provide a guide for evaluation and discuss

general considerations.

The goal is to develop generalized mathematical models for evaluating

the suitability of robotics applications in the construction industry.

That is, what construction operations can be robotized? To answer this

question, a robotization index, I, must be obtained such that a given value

or range of values indicates whether robotics is feasible for a particular

construction operation at a given point in time. A linear scale of the

index can be utilized as shown in Fig. 5.

I= 0. I=I 1 I=I2 I=I3 I=100.

Manual Manual Labor Man - Machine Man-Machine and Machine-Computer

Labor With Tools Operation Computer (Robotics)

Fig. 5. - Linear Scale of Construction Operation Index

74
 To determine this index, the following seven major variables affecting

on the feasibility of the robotics are considered:

1) Cost Effectiveness and Economical Analysis

2) Level of Hazardous

3) Productivity

4) Quality Improvement

5) Standardization of Design and Level of Repetitiveness

6) Union Resistance

7) Technologically Feasible

Any construction operation, if desired to be robotized, should satisfy

a certain level of these variables. Since each construction operation is

unique in nature, therefore, each operation will have different weight

factors to the above variables depending on their level of importance in

the operation. For example, in a welding operation inside a nuclear power

plant with a high level of radiation, variables 2 and 7 will have higher

weights than variable 5.

These variables must be analyzed in order to determine whether a

particular operation should or should not be robotized. Next, these

variables are defined and discussed.

Cost Effectiveness and Economic Analysis:

Applying robotics to a particular construction operation will most

likely involve a large initial capital investment. Capital investments are

based on the evaluation of the spending requirements and the returns

75
generated over the lifetime of the equipment. Sometimes a particular

construction operation is technologically feasible but not financially.

To determine whether a robot is economically feasible, costs and

benefits should carefully be studied. There are basically two major

economic analysis techniques for evaluating the desirability of a robot:

1) payback period analysis; and 2) cash flow analysis.

The payback period estimates the length of time (e.g., how many years)

it will take to recover investment costs as shown in Eq. 1.

= Total Initial Capital Investment

Payback Period (1)
Annual Savings Resulting from the Robot

In general, a determination of the total investment required is

necessary, then the effect of the investment on operation's expenses and

profitability should be analyzed. Eq. 2 provides a simplified way to

determine a payback period:

P = C
(2)
L+D+I-M

in which P = payback period in years; C = total initial capital investment

required in robot and accessories, L = savings on annual labor costs due to

the replacement of the robot; D = annual depreciation; I = annual savings

resulting from increase in annual production and improved product quality,

the value of I should be considered negative.if there is a decrease in

annual production; and M = annual robot maintenance costs. In general, the

values of L and I can be estimated from Eqs. 3 and 4, as follow:

L = W - S (3)
I = q(L+Z) (4)

in which W = annual cost of workers before the implementation of robot; S =

annual cost of staffing after the use of robot, q = speedup (or slowdown)

76
J J J J

factor due to the increase (or decrease) of annual production when a robot

is used; and Z = annualized value of the robot, in general, it might be

assumed as annual depreciation. Current robots have payback periods of 2-3

years when compared against direct labor.

Another method of economic analysis is the cash flow analysis. In

this case, either annual rate of return (internal rate of return) on robot

investment can be estimated, or the net present value of the investment can

be calculated by applying a required or an appropriate rate of return for

the robot. This method requires that an expected net cash flow be

developed, and a discount rate must be assumed. Table 1 identifies the

major cost and benefit items,-to be considered in the cash flow analysis.

Level of Hazardous

.Hazardous construction operations are very suitable for the robotiza-

tion. The distinction between unsafe operations and hazardous operations

should be made. Unsafe operations are assumed those in which there is a

high occurrence of worker accidents. Accidents are considered to be the

fault of the worker, either through carelessness or by the misuse of

equipment. Hazardous operations are assumed those operations which expose

the worker to an unhealthy environment (e.g., dust, radiation, heat, etc.).

The worker is not considered responsible for the conditions but due to the

nature of the operation, unhealthy human exposure is required. Historical

data generally indicates the frequency of job related accidents, while

standards relating to hazardous operations are provided by OSHA.

Some of the construction operations are hazardous, therefore, govern-

mental and private agencies have dedicated special attention to this kind

of operations. Several studies have conducted in which permissible

77
 TABLE 1. - Economic Analysis of a Robot

Cash Flow Items to be Considered

a) Total Robot Cost

Robot
Accessories
Options
Installation

1) CASH b) Maintenance Cost:

OUT
FLOWS Spare Parts
Maintenance

c) Downtime Cost

d) Increase in Energy Cost

a) Savings on Labor Costs

b) Productivity and Quality Improvement

2) CASH
IN
c) Depreciation Saving Through Tax
FLOWS
d) Salvage Value

78
 1
J 1

exposure limits for a variety of noxious elements commonly found in

construction operations have been set. In determining if a particular

operation is hazardous, the following elements should be analyzed.:

1) Concentration of Dust. Asbestos Zinc

Berylium Cement

Uranium Fibrous Glass

2) Temperature Levels

3) Air and Water Pressures

4) Noise

5) Radiation

Some of the current permissible exposure levels that have been set by

different governmental and private agencies are:

Dust - Any job site in which atmosphere contains dust particles smaller

than 5 um in aerodynamic diameter are considered hazardous for

the human lungs.

Toxic Dust, Gases, and Fumes.

- Job sites in which Chemical particles and substances are found

in a level superior to .15 mg/m3 of air.

Noise - Any job site in which sound levels exceeds 115 dBA. is considered

hazardous for the human health.

Radiation.

- The permissible exposure levels should be < 3 REM. per calendar

quarter or < 5 REM. per year.

Unsafe operations such as: sloped excavations, tunnel excavations;

and scaffolding operations can also be analyzed in similar way.

79
Productivity

Productivity levels in a particular operation are indicators of the

effectiveness of the different resources involved in the operation.

In order to determine if a particular operation is suitable for

robotization from the point of view of productivity, it is necessary to set

a desired or expected productivity level. After having conducted a

detailed and precise study of the productivity variation according to the

type of machine being utilized and according to the expected robot

productivity variation, the decision-maker should be in the position to

decide if the operation is suitable for robotization or not.

Generally, productivity of an operation is measured by dividing the

total number of units produced by the total amount of resources utilized in

a determined period of time.

Productivity can simply be defined as the ratio of output to input,

typically given as units produced per man-hours required. A comparison

between productivities of the current system and the proposed robotic

system should be made . If historical data on productivity is not available

then a study to determine these values must be made . Cyclone modelling of

the operation's tasks and sub-tasks for both systems may be used to

determine the value of productivity. Several assumptions may be needed to

model the robotic system, especially if it is a new or unique application.

The most desirable results would indicate that the robotic system provides

greater productivity in the comparison.

If a construction operation is automated or robotized, it is expected

to have a sharp increase in the productivity. The increased productivity,

supposedly, gradually absorbs the cost incurred in the robot or automated

equipment implementation. Obviously, productivity is not the only factor

so
 J
J

that pays for the robot . In some situations, the productivity achieved by

a robotized operation remains the same, but substantial savings are

expected to occur in other cost categories such as labor , overhead, etc.,

or even cost savings achieved by a better quality of the work.

A robot might have another uses in other projects. Therefore, the

analysis must consider these possibilities , just to study whether or not

the robot cost is commensurated by the better productivity achieved.

One must remember that certain construction operations involve a lot

of risk. In this situation, the productivity plays a secondary role,

because the main objective is to avoid detrimental and hazardous conditions.

For these reasons, the project planner must weigh every factor accordingly

to the desired goals.

Quality Improvement

One major reason for the implementation of robot is to produce a

better quality compared to traditional systems. The results of quality

analysis of the SSR-2 spray robot for fireproof cover work shows that the

dispersion of the sprayed thickness decreased.- Quality of a construction

product can be measured by a numerical model which considers such

characteristics as strength, dimension, color, and etc. Only the relevant

.characteristics of an operation product should be considered. There is a

direct correlation between cost and the level of quality improved.

Standardization of Design and Level of Repetitiveness

The cyclic and repetitive operations are the most suitable operations

to be robotized or automated. A repetitive routine operation is a

81
that pays for the robot. In some situations, the productivity achieved by

a robotized operation remains the same, but substantial savings are

expected to occur in other cost categories such as labor, overhead, etc.,

or even cost savings achieved by a better quality of the work.

A robot might have another uses in other projects. Therefore, the

analysis must consider these possibilities, just to study whether or not

the robot cost is commensurated by the better productivity achieved.

One must remember that certain construction operations involve a lot

of risk. In this situation, the productivity plays a secondary role,

because the main objective is to avoid detrimental and hazardous conditions.

For these reasons, the project planner must weigh every factor accordingly

to the desired goals.

Quality Improvement

One major reason for the implementation of robot is to produce a

better quality compared to traditional systems. The results of quality

analysis of the SSR-2 spray robot for fireproof cover work shows that the

dispersion of the sprayed thickness decreased,- Quality of a construction

product can be measured by a numerical model which considers such

characteristics as strength, dimension, color, and etc. Only the relevant

.characteristics of an operation product should be considered. There is a

direct correlation between cost and the level of quality improved.

Standardization of Design and Level of Repetitiveness

The cyclic and repetitive operations are the most suitable operations

to be robotized or automated. A repetitive routine operation is a

81
desirable operation characteristic for the robotization. A construction

operation should be broken down into individual processes, tasks, and sub-

tasks. The amount and type of repetition in each of these work divisions

should be analyzed. The decision-maker determines the number of cyclic

motions required in the production of one unit.

Standardization of design also involves repetition but on a larger

scale. Here, repetition is studied on the project or activity level.

Basically, this parameter evaluates the number of production units required

for successful robot implementation. Justification depends upon whether

the number of production units fall within an optimum range. If not,

perhaps some other man/machine system is more appropriate.

There are several means by which the number of production unit in a

project may be modified to fall within the optimum range for robotization.

In the project planning phases it is advantageous to orient various

building components (i.e. steel framing, doors, windows, rooms, etc.) in a

regular and predictable manner increasing the feasibility of robotization

by increasing the quantity of repetitious work cycles. Standard

dimensions, regular geometric shapes and standard size fixtures would

simplify implementation. Simplifying the construction design would in turn

simplify the robot's job, reduce the necessary 'learning period' (teaching

and reprogramming) and thereby increase robot effectiveness.

Standard design and repetitive operation are two factors that are

required for robotization or automation of any construction operation.

Several construction projects involve a determined amount of

repetitive operations but it does not mean that this operation should be

immediately robotized. Basically, the decision is based on quantity

82
J J J J j

parameters. Parameters that have been calculated based on economic

analysis of production.

Fig. 6 shows the relation between unit cost and number of units

produced, and shows if an operation should be utilized man power,

conventional machines , robot, or fixed automation based on the different

quantity parameters n1, n2, n3, and n4.

$/Unit
4

Fixed
Automation .,*O'

nl n2 Number of
n3 n4
Units Produced

Fig. 6. - Economic Stages of Automation vs. Level of Repetitiveness

83
 Because construction robot applications are in the first generation of

development and the construction products are unique items, the

standardization of design provides for an environment conducive to the

favorable application of robots. Standardization can mean the uniform use

of a particular design throughout the project or a series of uniform

activities being recognized as appropriate for robotization. Specifically,

the structural design should be simple and repetitive using standard

dimensions. Thus, the standardization of design is highly dependent upon

the future acceptance of limited design individuality in construction

projects and the future development of robots to handle the various

requirements of the construction site.

Since some activities may be repetitive but not standard , e.g., a

unique design which is being used a multitude of times on a particular

construction project. Repetitive activities will involve the cyclic

movement of or sharing of resources. A procedure to determine the

repetitive operation is to develop a number of units constructed versus

cost per unit curve. A curve developed for the construction operation

under study would yield a range such that n3<N<n4 would indicate

robotization.

Union Resistance:

Labor unions currently have few standard policies concerning the

automation or robotization of construction operations, therefore, the

reaction from organized labor can only be estimated. Unions have

traditionally viewed automation as providing improvements to working

conditions and in most cases respond in a positive manner.

84
 Union resistance is considered to be somewhat dependent upon the

following:

- number of workers being displaced

- union strength in the area

- policies of management (advance notice to union officials,

placement programs for displaced workers, etc.)

This parameter is more difficult to model because no definite

measurement scale of union resistance exists.

Generally, unions resistance vary depending upon the kind of operation

under consideration. If the robotization of a particular operation will

represent the possibility of a massive labor's displacement, unions

resistance could be so drastic that it could determine the success of the

whole operation.

In order to minimize unions resistance, the decision-maker should be

aware of the social implications of introducing new technology in the

operation. Obviously, solutions to the problem could be possible, such as

relocation of displaced workers within the industry or in other industries.

To reduce the union resistance, the following major factors must be

considered by the contractors:

minimum social disruption should be generated

2. job safety and worker satisfaction should be given large

consideration

3. the overall quality of life for workers should be enhanced.

An interview was conducted with the Business Manager of the Local 438

of the Laborers International Union of North America, located in Atlanta,

Georgia, which is affiliated with the AFL-CIO. The objective was to

ascertain the construction laborer's union's understanding of robotization

85
in the construction industry and the policy with regards to the

robotization . The following questions ( Q), and responses (R) were resulted

from the interview:

Q-1 What is your idea of a robot?

R-1 A machine that accomplishes a task in a methodical and precise

way. Not the science fiction version of humanoid-like machines

in the movie Star Wars

Q-2 Do you see a place for robotics in the construction industry?

R-2 Yes, such as in painting applications, but we will still need

human directions and involvement in site safety considerations.

Q-3 Are you familiar with the progress made in other countries such

as Japan towards robotization of the construction industry?

R-3 I have heard bits and pieces about foreign robotics development

but have no actual hands-on experience with foreign robot

advancement in construction.

Q-4 Does the union have any input with regard to non-union jobs?

R-4 None whatsoever , and union influence is weak in Atlanta and is

decreasing throughout the country.

Q-5 Does the union have a policy with regard to robots in construc-

tion? If not, is there any policy on automated equipment?

R-5 There is no current policy on r-obots with regards to construc-

tion. The policy with regards to any labor saving device is

not to hinder any increase in productivity. However, any unique

items are always negotiable points in contracts between the union

and the construction contractor.

Q-6 Would the union object to placing robots in hazardous and/or

harmful construction tasks?

86
R-6 No , but it would be a negotiable item requiring assurances that

the task be accomplished properly. I don't believe any construc-

tion job can be handled without human assistance. (This goes

beyond supervision into actual task accomplishment.)

Q-7 Is the current and future work environment in construction large

enough to absorb a small percentage of workers who could be

displaced by robots? If so, what percent would the union feel

comfortable with?

R-7 Take two of the same construction jobs over the last 30 years.

A job using 150 bricklayers in 1950 would use 40 or so today

because of improved construction techniques, materials, etc.

There is no place to absorb workers displaced by robots. At

this time, no percentage of displacement is a comfortable

prospect.

Q-8 If robots become a reality in the construction field, what in

your opinion will be the most significant changes brought about

in the workplace? To the individual worker?

R-8 As many workers would be needed to direct or maintain the

robots as are now needed to accomplish the work. Look at

computers, which equate to robots in the sense of mechanizing

human work tasks. They have created as many jobs as they have

eliminated. (It was pointed out that the new jobs created were

different and required a retraining process.)

Q-9 If robots become feasible, would the economics of dollar savings

be enough justification to displace workers? If not, what is

the union's flexibility on this issue?

R-9 The union has never stood in the way of progress. The papers

87
 have given the unions bad press. In the last 30 years the

displacement of workers has continually increased, and the union

has accepted this as part of progress . The union has never

struck against improvements.

Q-10 What stipulations would the union demand so that the least amount

of disruption would affect the worker displaced by robotization?

R-10 Retraining the displaced workers so that they could maintain

their self-respect and lead productive lives. This is the basic

concept we would negotiate for.

Q-11 Do you believe that a future situation can exist where humans

and robots work in the same environment without conflict?

R-11 Yes, robots can be utilized in construction working side-by-

side with humans . It is happening now in automobile factories.

Q-12 Since the introduction of automation did not cause a mass un-

employment crisis, do you see a parallel occurrence with the

introduction of robots?

R-12 During my lifetime, there have been many revolutionary changes

and mass unemployment did not result unless a depression was

occurring . Robots, automation , whatever will not cause mass

unemployment as long as people are retrained.

Q-13 Do you foresee a larger impact on minorities such as women and

blacks from the robotization of the construction site?

R-13 No, not in union represented work. The union is an equalizer.

Job assignments are handled fairly on a first-come-first-serve

basis.

Q-14 An extreme reaction to robotics could be construction site

sabotage. Could you foresee such incidents even if the unions

88
 deplore such tactics?

R-14 People will do most anything if pushed to the extreme, but the

union would not condone such activities. If mass unemployment

resulted, repercussions would probably occur here in Atlanta.

Q-15 How would a program for the retraining of displaced workers be

set up?

R-15 We have a training school in Atlanta which is contractually

funded through negotiated dollar amounts set aside per worker

hour. Any retraining program would use this existing system.

Q-16 Would the union consider requesting government intervention to

either delay or prohibit robotization in construction?

R-16 We belong to the AFL-CIO and they are our political arm. They

have a lobby like everybody else. So, we would lobby if it

became important to our union members.

Q-17 What about the workers who are not displaced? Will they need

some kind of training to function better in the new robotic work

environment?

R-17 Possibly, and if so we would use the existing training system

mentioned earlier.

Q-18 What job security techniques would the union utilize to assure

minimum worker disruption?

R-18 We would explore every avenue available such as more vacation

time, same pay for less time worked, etc. Of course, this would

depend on the negotiation process.

Q-19 Would the knowledge that foreign competition through the utili-

zation of robots on construction projects was placing American

companies in an unfavorable position cause the union to accept

89
 deplore such tactics?

R-14 People will do most anything if pushed to the extreme, but the

union would not condone such activities. If mass unemployment

resulted, repercussions would probably occur here in Atlanta.

Q-15 How would a program for the retraining of displaced workers be

set up?

R-15 We have a training school in Atlanta which is contractually

funded through negotiated dollar amounts set aside per worker

hour. Any retraining program would use this existing system.

Q-16 Would the union consider requesting government intervention to

either delay or prohibit robotization in construction?

R-16 We belong to the AFL-CIO and they are our political arm. They

have a lobby like everybody else. So, we would lobby if it

became important to our union members.

Q-17 What about the workers who are not displaced? Will they need

some kind of training to function better in the new robotic work

environment?

R-17 Possibly, and if so we would use the existing training system

mentioned earlier.

Q-18 What job security techniques would the union utilize to assure

minimum worker disruption?

R-18 We would explore every avenue available such as more vacation

time, same pay for less time worked, etc. Of course, this would

depend on the negotiation process.

Q-19 Would the knowledge that foreign competition through the utili-

zation of robots on construction projects was placing American

companies in an unfavorable position cause the union to accept

89
 robotization in the USA more readily?

R-19 We would prefer that anything be American made and better yet,

union made. Therefore, we would work with the contractor to

stay competitive.

Q-20 Would the union be interested in the results of the research

currently being done at Georgia Tech and help in drafting a

policy guideline with regards to robotic applications in con-

struction?

R-20 At this time no policy exists nor is one being developed, but we

would certainly be open to any help on robots when the issue

becomes pertinent to us.

It should be mentioned that the above interview is not an appropriate

sample size to draw a general conclusion with high degree of confidence,

similar questions must be asked from a broad range of union representatives

at all levels of authority and all regions of the United States.

Technologically Feasible

In spite of the technological advances achieved in the last few years,

technology does not always provide the necessary elements to develop

machines for certain kind of industrial operations. For this reason, it is

important that this factor be analyzed in the first stages of the study in

order to determine if technology provides the tools to develop the

appropriate machine for the operation in question. If the study reveals

that development of a robot is not technologically feasible, further study

of the other factors are not necessary, since the whole operation cannot be

achieved.

The various aspects to be analyzed in this area are:

90
 J

1) Type of Mobility of Robot

a. Wheels, tracks or walking devices

b. Carrieage system (traveling and standing frames)

2) Robot's Manipulators

3) Control Systems

4) Methods of the Construction Material Supply

5) Weight of Robot (within design load)

6) Size of Robot

7) Robot's Safety Functions (human life and limb protection)

It is expected that mobile robots will find increased popularity in

construction industry. A fixed robot has a limited sphere of operation and

is not appropriate for the construction sites.

A construction wheeled vehicle robot, such as a motor car, with firmly

inflated tires represents an ideal system with minimum eneregy to operate

on smooth surfaces which have sufficient friction to the wheels to propel

and steer the robot without slipping. Wheeled systems can only operate

over relatively smooth surfaces. The track systems are the known

alternatives to wheels for rough ground mobility.

91
 I

III. FEASIBLITY MODELS

Two basic feasibility models are presented for justifying the

implementation of robotics in certain construction operations: 1) a

simplified management decision model; and 2) utility decision model. The

first step in the formation of the models is to identify the relevant

variables. The second step is to develop a criteria for estimating a

management decision index.

1) Simplified Management Decision Model:

This model allows the management to make a quick decision about the

automation or robotization of a particular construction operation. The

seven major variables discussed previously are considered as major

management decision variables as shown in Table 2. Different weight

factors should be assigned by the management to each variable. These

weight factors indicate the level of importance of the variables, and they

may vary one operation to the other. Columns 3 and 4 show the actual and

acceptable standard levels of the variables. For example, as shown in

Table 2, for a given construction operation the level of dust is measured

as 3 pm, however based on the N.S.C. safety and health standards, the

acceptable level of dust is 5 pm. Since the measured particles of dust are

smaller than the standard size, therefore, this construction operation is

unsafe, and it is appropriate for the robotization.

The index for other variables are evaluated as described in previous

sections. Some indices may not be possible to evaluate, such as union

resistance, in this case, a 'Yes' or 'No' answer is sufficient. Column 5

shows the necessary relation between columns 3 and 4 in order to robotize.

If the relation on column 5 holds, then a 'Yes' answer with the given

92
 J

U)
aJ U) ,^
M a H
a)
4.1 H 0
0 0 0 a
0
xi v.c
^ I I I I „ i I
O
°
z z
z°+ a
ci u) C0 U) U)
4-1 u +^ 1J 1J 1.J ^+
H C0 H H H H H H
0 0 0 0 0
o H 0. 0.. 0. °
a
0 ° 0 cn
N
O O In If., cn
Ri r. M H H H 01
I
1 to '- II 11 II 11 II II II
U) U) to V) to U) to
w aJ a) a) a) a) a) a)

O NN
4-1 H
N Lf)
IC .+ H
CYJI 0
U) -H 0 L ^ ^ v v V V n
U) 1-1 P4
a) co H N M .t in I>
U H Q) 0 H H H H H H H
a) a) ,= H
zxL.u
v ^4 4-1

C v q
H >+ W
4-1
"0 a) Lr) O H

•aXro
0 4J H
- N

II II u
LLB
11
I I
ca
iJ W ^4
CO H N M Cn
C/) 0> H H H H

a) U) H
10

H a) O
U) 0 LE
a) H M
O C ca
,.
M
NII 11
N II
00 0
II
I I
U) 41 H
N
ca w
a7 W 0 H H H H
O
W U) U) N U) U) U)
41 JJ iJ 41 1.J 1J 1J

0 0 0 0 0 0
w U a. 0. . H
° O Cn O O Cn Cr) If p .,
H 3.1 001.)
o M N H H H
a) O H U N O
CL (1) W `-^ II 11 II II II II II C
U

H v 3 3 3 3 3 3 3

U) U
0 E I a)
w Q) > 41
r-I co > 0 (1) cd H

0) ca ^4 41
0 W H H E R'. H -Hl
H a) U > H U) 00
44 0 `0 CO a 0 Q)
H a U 0 a) w 1
U U) rj co r4 JJ U L ° U)
a) a) .. O .D H Cd O H H C C. CH a7
A --I H C a) 11 H b -I al U O U) 6
+-^ o ca > U) -o u 0 CT) >> H u ca E-+
11 al U 94 a) co 4J 4 O aJ H C a) a) o
0H Wv a AfJ4 a) w O' a 11 Hw E-4
•r 41
Cro co .. ,-, .. .. r. .. ..
H N Cl) d Cr) a0 N-

93
weight factor should be entered in column 6, otherwise, a 'No' answer with

the same weight factor will be assigned to the column 7. Total of columns

6, and 7 are shown on the last line. It is the management decision to find

the cut-off points of the 'Yes' answers. It is suggested that those

operations with 75 points or more 'Yes' answers should be carefully

analyzed for the robotization.

2) Detailed Feasibility Analysis by Utility Model:

The management model described is a simple 'Yes' or 'No' analysis

which can be used for the preliminary feasibility study of the robotization.

A utility model is developed in this section to allow the contractors a

detailed study of the operation.

This model assumes that it is technologically feasible to build a

robot, and those operations which are not technologically feasible cannot

be implemented, therefore, they do not apply to the modeling process. To

avoid the double counting of the variables, those variables with high

correlation will be combined. Such as productivity and payback period will

be assumed as one variable due to their high correlation.

This model implements the utility theory to establish a utility

function for each of the variables, then these utility functions are

combined to estimate a single index of robotization. The following

variables with low correlations have been considered in this model: a)

payback period, b) level of hazardous, c) quality, d) level of repeti-

tiveness, e) union resistance.

The next step in the modeling process is to assign acceptable upper

and lower limits to the possible range of values in each variable. Both

desirable and undesirable magnitudes should be included in the range. As

94
shown in Fig. 7, these values are plotted along a horizontal axis, and the

utility scale is the vertical axis. Two points on the utility scale are

defined: 1.00 is assigned to the most preferred value, and zero is

assigned to the point lying directly between the desirable values and

undesirable values, a neutral point. A utility function can be developed

between the two points on the scale.

As shown in Fig. 7a, a payback period of zero year is assumed to have

the highest utility, and an accepted period of 2.5 years by the industry as

the indifference point. In Fig. 7b, a dust particle size of 5 um is

assumed as indifference point, in other words, if the level of dust

particles are less than 5 pm, then the operation is suitable for the

robotization, therefore, it has a positive utility (satisfaction). Fig. 7c

shows two levels of quality a1 and a2 with zero and one utility value.

These values must be estimated by the management. If the quality of the

work performed by robot has a value less than a1, then the implementation

of robot based on this variable is not desirable. Fig. 7d is developed

based on Fig. 6. If the level of repetitiveness is between n3 and ny, then

the robots should be used. Level of n4 is an optimum repetition when the

robotization is compared with the fixed automation. Fig. 7e divides the

level of the union resistance into three levels: low, medium, and high.

An average medium level is considered as zero utility. The assigned values

should be adjusted by the decision maker under different conditions.

To consider the impact of different variables and their individual

contribution to the overall index, each of the measurement scales must be

converted to one common scale. A set of scaling factors (weight factors)

is used in the conversion. Each variable receives a scaling factor, the

magnitude of which is based on the estimated importance of the variable to

95
 J

Utility of First Variable Utility of Second Variable

u(V2)
IL
1.00 t-----------

a) Payback Period vs . Utility b) Size of Dust Particles vs. Utilit

Utility of Third Variable Utility of Fourth Variable

u(v3) u(v4)

1.00 1.00

0.75

0.50

0.25

0.00 v3 0.00 v4
a2
(Quality Level) (Total
-0.25 t' -0.25 Number
Repet i-
tions)
c) Quality vs. Utility d) Level of Repetitiveness vs. Utili
 J

Utility of Fifth Variable

u(v5)

1.00

0.75

0.50

0.25

0.00

-0.25

e) Union Resistance Level vs. Utility

Fig. 7.- Utility Function of Each Variable

97
I 1

the index. Each utility is multiplied by its scaling factor as shown in

Eq. 5. The products are then summed to yield a total relative index for

the operation. The alternatives with the highest global utility (index)

should be selected. In this model the alternatives are a combination of

conventional systems versus a proposed robotic system.

I = w1 u(v1) + w2 u(v2) + w3 u(v3) + w4 u(v4) + w5 u(v5) (5)

in which I = robotization index or global utility; wi = weight factor of

ith variable, vi = value of ith variable ; and u = utility function. A

positive value of I indicates that the operation should be carefully

considered for the robotization.

The success of this model is directly related to the proper selection

of the variables , accuracy of the utility curves and weight factors. If

the decision criterion fall within separate disciplines of study then a

professional in that area should be assigned to evaluate the measurement

scale. In the assignment of weight factors the user must have a thorough

knowledge of not only the operation tasks, but also the goals of the

management . Overall the modeling of robotics for construction operations

is necessary for the justification of implementation. The model enables

the management to reduce the risks involved and consistently estimate the

results of implementation. In the long run, this would aid in the

increased use of robots throughout the industry.

98
 J

IV. SUMMARY AND CONCLUSIONS

Seven major variables affecting on the feasibility of the robotics in

construction industry were identified as: 1) cost effectiveness; 2) level

of hazardous; 3) productivity; 4) quality improvement; 5) standardization

of design and level of repetitiveness; 6) union resistance; and 7) techno-

logically feasible.

Two models were developed for the robotics feasibility in the

construction industry: 1) simplified management decision model; and 2)

utility decision model. The ultimate output of these models provide an

index which indicates the level of automation.

Hazardous construction operations are the prime motivation in the U.S.

to implement robotics in the construction domain. However, the problem of

lower productivity in construction industry is expected to be an incentive

for future use of robotics. Developing new design techniques based on

standard elements and repetitive operations must be further investigated.

This can result in developing entirely new techniques of construction,

feasible for the robotization.

99
 I

V. ACKNOWLEDGEMENTS

The efforts of graduate students in the Construction Management

Program at Georgia Institute of Technology are the basis of this report.

Appreciation is extended to H. Jones , S. Chawla, F. Nakad, A. Gutierrez, P.

Fernandez , and C. Obetts.

100
 t J J 1

VI. REFERENCES

1. Ayres, R.U., and Miller, S.M., "Robotics", Ballinger Publishing Co.,

Cambridge, Mass., 1983.

2. Berger, P., "Robot Catalog", Dodd, Mead and Co., New York, NY, 1984.

3. Carr, Robert I., "Simulation of Construction Project Duration",

Journal of the Construction Division, ASCE, Vol. 105, No. C02, June
1979, pp. 117-128.

4. Carr, Robert I., and Meyer, Walter L., "Planning Construction of

Repetitive Building Units", Journal of the Construction Division, Vol.
100, No. C03, September 1974.

5. Crawley, J.L., "Dynamic World Modeling and Navigation for an Intelli-

gent Mobile Platform", Paper presented at the Workshop Conference on
Robotics in Construction, Carnegie-Mellon University, Pittsburgh,
PA, June 1984.

6. Dabbas, Majed, A.A., and Daniel W. Halpin, "Integrated Project and

Process Management ", Journal of the Construction Division, ASCE, Vol.
108, No. C03, September 1982, pp. 361-374.

7. Engelberger, Joseph F., Robotics in Practice, Anacom, New York, 1980.

8. Fenves, S.J., and Rehak, D.R., "Role of Expert Systems in Construction

Robotics", Paper presented at the Workshop Conference on Robotics in
Construction, Carnegie-Mellon University, Pittsburgh, PA, June 1984.

9. Fondahl, J.W., "Photographic Analysis for Construction Operations",

Journal of the Construction Division, ASCE, Vol. 86, No. C02, Proc.
Paper 2483, May 1960, pp. 9-25.

10. Furukawa, Tsuned, "Introduction of Robots to the Construction Work

Site", Robot, March 1983, pp. 14-19.

11. Halpin, Daniel W., "CYCLONE: Method for Modeling of Job Site
Processes", Journal of the Construction Division, ASCE, Vol. 103,
No. C03, Proc. Paper 13234, September 1977, pp. 489-499.

12. Halpin, D.W., and Woodhead, R.W., "Design of Construction and Process
Operations", John Wiley and Sons, New York, NY, 1976.

13. Hasegawa, Yukio, "Robotization of Construction Work", Robot, March

1983, pp. 41-46.

14. Holland, J.M., "Basic Robotic Concepts", Howard W. Sams and Co., Inc.,
Indianapolis, IN, 1983.

15. Japan Mechatronics Letter, "Ohbayashigumi to Develop Construction

Robotics", May 20, 1983, p. 6.

101
 J

16. "Japan Takes Early Lead in Robotics", Engineering News Record, July
21, 1983.

17. Manninen , M., and Karkkainen , P., "Supervising Control of Large-Scaled

Manipulators in Severe Environments ", Paper presented at the Workshop
Conference on Robotics in Construction , Carnegie - Mellon University,
Pittsburgh , PA, June 1984.

18. "Mechatronics ", Product News , Sedno 120A00 , Komatsu America Corp.,
Norcross, GA, 1983.

19. Moore, W.C., "Application of Small Computers in Construction", ASCE

Proceedings, ASCE National Convention, San Francisco, CA, October
1984_

20. Olishifski, J.B., "Fundamentals of Industrial Hygiene", 2nd Ed.,

National Safety Council, Chicago, IL, 1979.

21. Paulson, B.C., "Automated Control and Robotics for Heavy Construc-
tion," Paper presented at the Workshop Conference on Robotics in
Construction, Carnegie-Mellon University, Pittsburgh, PA, June 1984.

22. Paulson, B.C., "Automated Control and Robotics for Construction",

Paper presented to the Journal of Construction Engineering and Manage-
ment, ASCE.

23. Rehak, D . R., and Fenves , S.J., "Expert Systems in Civil Engineering,
Construction and Construction Robotics ", 1984 Annual Research Review,
Robotics Institute , Carnegie - Mellon University , Pittsburgh, PA.

24. "Robots : A Manager ' s Guide", Cincinnati Milacron , Cincinnati, OH,

1982.

25. "Robots Coming to Jobsites", Engineering News Record , February 10,

1983.

26. Safford, E.L., "Advanced Robotics", Tab Books Inc., Blue Ridge Summit,
PA, 1982.

27. Sangrey, D.A., "Constraints on the Development of Robots for Con-

struction", Civil Engineering Department, Carnegie-Mellon University,
1984.

28. Shimomura, Y., and Sonoda, T., "Tunneling by Robots: Shield Driving
Automatic Control System ", Hazama-Gumi , Ltd., Paper presented at the
Workshop Conference on Robotics in Construction, Carnegie-Mellon
University, Pittsburgh, PA., June 1984.

29. Sullivan, B.J., "Industrialization in the Building Industry", Van

Nostrand Reinhold Co., 1980.

30. Susnjara , K.J., "Industrial Robots ", Prentice-Hall, Inc., Englewood

Cliffs, NJ, 1982.

102
31. Tamura , Y., and Kano , N., "A New Management Tool for Robotized
Construction Projects", Paper presented at the Workshop Conference on
Robotics in Construction, Carnegie-Mellon University, Pittsburgh,
PA, June 1984.

32. Tarandi, V., and Birke, H., "A Pre-Study of Robotics in Construction",
Paper presented at the Workshop Conference on Robotics in Construc-
tion, Carnegie - Mellon University , Pittsburgh , PA, June 1984.

33. Tucker, Richard L., et al., "Implementation of Foreman-Delay Surveys",

Journal of the Construction Division, ASCE, Vol. 108, No. C04,
December, 1982, pp. 577-591.

34. Unchida, Seiichi, "Execution of Civil Engineering Work by Using

Robots", Robot, March 1983, pp. 47-52.

35. Vallings, H.G., "Mechanism in Building", Second Edition, Applied

Science Publishers, Ltd., 1975.

36. Vend, Takatoshi, "Robotization of Spraying Work for Fire Proofing

Steel Structure", Robot, March 1983, pp. 53-61.

37. Warring, R.H., "Robots and Robotology", Tab Books Inc., Blue Ridge
Summit, PA, 1983.

38. Warszawaki, A., "Application of Robotics to Building Construction",

Paper presented at the Workshop Conference on Robotics in Construc-
tion, Carnegie-Mellon University, Pittsburgh, PA, June 1984.

39. Whittaker, W., "Remote Reconnaissance Vehicles", Technical Report,

Carnegie-Mellon University, The Robotics Institute, Pittsburgh, PA.

40. Yoshida, T., Veno, T., Nonaka, M., and Yamazaki, S., "Development
of Spray Robot for Fireproof Cover Work ", Paper presented at the
Workshop Conference on Robotics in Construction , Carnegie - Mellon
University, Pittsburgh, PA, June 1984.

41. Zeldman, M.I., "Robots", Marcel Dekker, Inc., New York, NY, 1984.

103