Professional Documents
Culture Documents
net/publication/269111475
CITATIONS READS
15 753
3 authors:
Bahram Ravani
University of California, Davis
223 PUBLICATIONS 4,468 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Ty A. Lasky on 12 March 2021.
Abstract: This paper presents a cost-benefit analysis of mobile terrestrial laser scanning (MTLS), specifically in highway infrastructure (HI)
Downloaded from ascelibrary.org by University of California, Davis on 12/02/14. Copyright ASCE. For personal use only; all rights reserved.
applications using data and requirements from the western United States. MTLS can be used for rapid data collection in a digital point-cloud
format. This data can then be analyzed for mapping and feature inventory in several applications including asset management, preservation,
maintenance, and operational planning performed by State Department of Transportation agencies. In this paper, data and requirements related
to some of the operations of Washington State Department of Transportation and the California Department of Transportation are used as the
basis for this cost-benefit analysis. DOI: 10.1061/(ASCE)IS.1943-555X.0000192. © 2014 American Society of Civil Engineers.
Author keywords: Light detection and ranging (LiDAR); Cost benefit; Highway maintenance; Highway construction; Surveys.
are presented. degree field of view (FOV) by design. New generations of 2D Li-
DAR scanner designs enable 360° FOV. The LiDAR range error
directly adds to the overall MTLS-system absolute error. Therefore,
MTLS Systems Overview only a few LiDAR scanners are suitable for high-accuracy pave-
ment survey work. Currently, most survey-grade MTLS systems
MTLS systems are similar to airborne light detection and ranging use the Optech Lynx V100, V200, or M1, or the Riegl VQ-250
[LiDAR (laser scanning)] systems. Fig. 1 illustrates the basic sys- or VQ-450 LiDAR time-of-flight (TOF) scanners. Both Optech
tem-architecture concept. The system consists of a dual-frequency (Vaughan, ON, Canada) and Riegl (Orlando, FL) scanners have
real-time kinematic (RTK) global navigation-satellite system long-range and high-range accuracy. A few custom-engineered
(GNSS) receiver(s), a six degree-of-freedom (DOF) inertial meas-
MTLS systems (Burchi, presented at SPAR, 2008; Graefe, pre-
urement unit (IMU) (typically three accelerometers and three gyros
sented at SPAR, 2010) employ Z&F (Bridgeville, PA) or Faro
orthogonally mounted), a distance-measuring indicator (DMI),
(Lake Mary, FL) phase-based laser scanners which have higher
LiDAR scanner(s), data synchronization electronics, data logging
range accuracy. However, their practical range is much shorter than
computer(s), and digital camera(s) (Mrstik and Newby 2005;
the TOF scanner. Consequently, the operator may have to drive
Newby and Mrstik 2005; Glennie 2007; Graefe, presented at
multiple passes on different lanes of the multi-lane freeway to
SPAR, 2010; Mandli et al. presented at SPAR, 2010). The com-
gather all the required data. Nevertheless, MTLS systems with
puter(s) collects the synchronized data of GNSS carrier-phase
phase-based laser scanners would yield higher relative-accuracy
measurements, IMU and DMI outputs, digital photographs, and
point-cloud data and are better suited for urban areas.
LiDAR-scanner data for the postprocessing software. Combining
MTLS systems have been developed to collect geospatial data
GNSS base-station raw data collected at the same time, based
(Frecks, presented at SPAR, 2009; Adams and Dunham, presented
on the GNSS measurement, IMU, and DMI data of the rover,
at SPAR, 2010). Their high mobility and capturing ability of large
the software will provide position and orientation solutions (at up-
highway areas (roadways, roadsides, structures) enables DOT users
date frequencies of 100–1,000 Hz) of the sensor platform contain-
to collect massive amounts of data for highway surveying, asset
ing the LiDAR scanners and digital cameras. To achieve the highest
management (Howerton and Sideris 2010), as-built documentation,
possible accuracy, the raw GNSS/IMU data are postprocessed with
and maintenance operations (Frecks, presented at SPAR, 2008;
GNSS base station(s) raw data with high-accuracy satellite orbital
Harrison and Frecks, presented at SPAR, 2010; Mandli et al. pre-
measurements. Furthermore, the system may have multiple LiDAR
sented at SPAR, 2010; McNease and Perry, presented at SPAR,
scanning sensors with digital cameras in visible light, near infrared
2010). They have also been used in railway-tunnel scanning
(NIR), or ultraviolet (UV) wavelengths. An NIR or UV camera may
(Sandrone and Wissler 2011).The LiDAR scanner’s resultant point
be used to better determine health of certain plant species or to
cloud, shown in Fig. 2, is then postprocessed to extract mapping
compensate for poor lighting conditions.
data, roadside asset–feature data, and various measurements. If
The LiDAR scanner produces distance and angular measure-
such a database were developed, the extracted data could then
ments to objects as well as the amplitude of the light-return signal.
be readily imported into a database for analysis and be accessible
by all DOT business areas, thus reducing duplicate data collection
GNSS and storage in separate silos. Of course features like culvert ends,
Infrastructure dry wells, regulatory outfalls, and anything underground are not
visible in the point cloud; therefore, they cannot be extracted.
However, these types of features are relatively safe to collect
Extrinsic Sensors Data
Mobile Proc. Point
Digital
Platform GNSS Camera Sensor Cloud
Fusion
LiDAR
Human
Intrinsic Sensors Interface
IMU DMI
World
Highway Model
Infrastructure
Bridge-Clearance Requirements
Statement of the Problem
State DOTs have programs that are responsible for collecting
The mission of state DOTs is to keep people and business moving bridge vertical and horizontal clearance as well as other bridge
by operating, maintaining, and improving the state’s transportation as-built structure information. The absolute accuracy requirement
systems vital to the public, businesses, and communities. State for bridge location is relatively low [ca. 2 ft (0.6 m)]. The following
DOTs require geospatial data for planning, maintenance, operation, are practical recommendations for the relative accuracy require-
and project delivery. Despite the high initial costs of MTLS systems ments for bridge clearance data:
and data-extraction software, the solution may be cost-effective • Bridge columns/piers (horizontal clearance): within 4 in.
compared to existing methods. The costs and benefits of MTLS (10 cm) optimum;
solutions are examined in detail in the remainder of this paper, • Bridge girders (vertical clearance): within 1 in. (2.5 cm);
sorted by functions. Not all functions that would benefit from • Overhead restrictions (including truss members) (vertical/
LiDAR are included, just those with tangible data. horizontal clearance): within 1 in. (2.5 cm) vertical; within
3 in. (7.6 cm) horizontal;
• Bridge rails/guard rails (horizontal clearance): 3 in. (7.6 cm);
Application Areas • Edge of pavement (horizontal clearance): 1.5–6 in. (3.8–15 cm);
and
There are at least three existing programs [roadside feature inven- • Sign bridges (vertical clearance): within 1 ft. (0.3 m).
tory program (RFIP), bridge clearance measurement, and ADA fea- All of these measurements are achievable with a survey-grade
ture inventory], performed by some DOTs, that would directly MTLS system. Beyond the clearance information needed for post
benefit from employing MTLS systems. These programs’ historical heights, bridge strikes by over-height trucks can occur several times
and current expenditures in the western United States are used in per year. Proving damage by having existing condition data would
this cost-benefit analysis. Their data and functional requirements be extremely valuable in cases resulting in litigation. In addition,
are well-defined. There are other state-DOT functional groups that existing condition would be valuable in evaluating the effects of
can greatly benefit from the data produced by these mobile systems environmental hazards.
(Chang et al. 2006; Glennie 2009; Mendenhall 2011). However,
data on their program expenditures for geospatial-data collection
ADA Feature–Inventory Requirements
were not fully available (or varied greatly from year to year) to do
proper comparisons. State DOTs are responsible for collecting data on ADA ramps
at intersections to ensure their compliance to ADA standards and
regulations. The types of ADA features that are usually captured in
RFIP Requirements
this type of inventory are
The RFIP used by the Washington State Department of Transpor- • All curb-ramp types, including detectable warning-surface
tation (WSDOT) involves collecting, storing, and reporting roadside placement;
features such as guardrails, culverts, signs, objects in clear zones, • Bridge-end sidewalk ramps;
and other features from different regions within the state. This data • Sidewalks;
can be used for asset management, project and system design, and • Pedestrian access routes (PAR) with protruding objects and
overall system analysis. In some cases, some individual businesses other obstructions;
in a state have collected similar information (e.g., utility poles, signs, • Crosswalks;
guardrails, tree groupings, and slope information) independently of • Driveways;
one another. This causes duplication of efforts and expenses, which • Island and median cut-throughs;
can result in the data not being consistently stored in a statewide, • Accessible pedestrian signals (APS), push-button types, and
standard format that would allow shared use and proper mainte- locations;
nance of the data. Because of the advancements in technology, • Shared-use paths;
GIS applications, and the creation of the RFIP, it will be desired • Independent walkways;
by state DOTs to combine this information resulting in a single • Facility ramps;
source for data retrieval. In this fashion, the stored data will be more • Edge protections;
accurate and, therefore, the ability to analyze and maintain the data • Handrails;
increases exponentially, saving the agency time and money. • Railroad crossings;
that it costs $300 per structure including labor, lodging, per diem, capture highway interchanges. In order to reach a baseline target
and basic vehicle costs. The BPO also estimated that it spends for the cost-benefit analysis, it is assumed here that there are ap-
$80,000=year to collect data on the 270 structures/records per year. proximately 15,000 mi (24,000 km) that would need to be driven
Consequently, it takes the BPO about 10 years to complete a full by the data-collection vehicle in order to pick up the entire state’s
cycle of data collection for all WSDOT bridge structures. In order network of highways handled by the state DOT. There can also be
to achieve a 10-year cycle for data collection, this agency needs a limitations on the number of days per year that data can be collected
data-collection rate exceeding 270 structures per year. due to weather and other conditions such as construction and main-
tenance. For the state of Washington, for example, due to its typical
ADA-Feature Inventory Data–Collection Methods and weather conditions, it is assumed here that data collection can only
Costs be conducted for 6 months of the year. In such a condition, a single
vehicle can collect data for the entire WSDOT highway network
As part of WSDOT’s ADA-transition plan, WSDOT has started
in 2 years.
collecting data for an ADA-feature database that includes gathering
The total cost can be broken down to several cost categories as
data on what features exist or are missing in the field. A survey
follows:
crew collects data on a wide array of ADA features and their
• Information-technology (IT) costs. This includes the cost of
elements. In these efforts, the precise measurements of the ADA
data-extraction software, storage, server, backup, and worksta-
elements are the primary objectives of the crew.
tions for data extractions. The current WSDOT-IT charges are
In the case of a curb ramp, for example, the ADA data–
approximately $7.3 per gigabyte (GB) for storage and asso-
collection crew will locate the ramp by placing a mapping-grade
ciated backup costs. In addition the cost of a virtual server is
GPS unit at the center of the ramp cut along the back of the curb.
approximately $1,000. Our pilot study indicated that the esti-
The mapping grade GPS has an absolute positional accuracy of
mated data requirement is about 3 GB=mi [1.86 GB/km (Yen
5 ft. (1.5 m). Once a GPS location tag is generated for the ramp
et al. 2011)]. Thus, the total data storage requirement is 45
location, the ramp attributes are measured by hand tapes and digital
terabytes (TB) and costs $328,500 for a 2-year data-collection
levels [Stabila (Stabila, Inc., South Elgin, IL) or Smart Level
period. Five workstations ($15,000) are needed for data postpro-
(SmartTool Technologies, Oklahoma City, OK)] and entered into
cessing. In addition, data-extraction software licenses and yearly
a digital-data collector using a data dictionary created specifically
renewal costs are approximately $80,000 and $20,000 per year,
for the ADA-collection program. Digital range-finders are also used
respectively;
to measure the horizontal/vertical differences of remote features. The
• Data-collection costs. The total costs associated with data
field crew shares a pickup truck, a digital camera, a laptop computer, a
collection is composed of the equipment cost, equipment-
digital level, and tape measures. The crew size typically consists of
maintenance cost, personnel cost, and vehicle cost. The
two people. A three-person crew is sometimes needed in heavily
vehicle costs are estimated at $700=month or $0.56=mi
urbanized areas primarily to provide a traffic spotter.
ð$0.35=kmÞ; and
This process at a curb ramp takes a well-trained crew about
• Data-extraction cost. This is the total cost of the personnel to
10 min to complete. The time required for measuring all the ramps
extract the required geospatial data for various WSDOT pro-
in the entire intersection depends on the number of ramps in the
grams. The data-extraction cost depends highly on the number
intersection. Usually, there are only ADA ramps and crosswalks
of features required to be extracted.
in each quadrant. However, there may be a need to collect more
In order to simplify the calculation of personnel costs, an aver-
data in some intersections that have traffic islands, pedestrian fa-
age cost of $50=h is used for labor rate for all of the personnel
cilities, or obstructions. The estimated time to complete an entire
needed for data collection and processing in mapping-grade appli-
intersection with minimal attributes is 1–2 h. However, an intersec-
cations. For survey-grade applications, a cost of $135=h is used.
tion with extensive collection points and heavy traffic may take 4 h
Deployment of MTLS technology can be achieved in many
for the crew to complete all the measurements. In-office postpro-
cessing time is usually 1 h per intersection per data collector used. different ways. Seven of the most common approaches are dis-
The estimated program cost from July 2009 to June 30, 2010 cussed below, where their costs are estimated and compared. These
was $118,414, and the cost from July 2010 to March 2011 was options are
$237,321. Itemized ADA feature-inventory data collection costs 1. Contract for mapping-grade MTLS services;
are provided in Table 3 2. Contract for bridge-clearance measurement services;
3. Rent and operate a mapping-grade MTLS system;
4. Purchase and operate a mapping-grade MTLS system;
MTLS-Deployment Option Details and Costs 5. Rent and operate a survey-grade MTLS system;
6. Purchase and operate a survey-grade MTLS system; and
To collect data for a state-DOT roadway network, a MTLS-system 7. Purchase fractional ownership of a survey-grade MTLS
vehicle would need to be driven in both directions on each roadway. system.
time, and the DOT would not have a mobile-LiDAR system equipment maintenance. The disadvantage is that DOT would not
available at a moment’s notice. This option provides data with have a mobile-LiDAR system available at a moment’s notice. The
mapping-grade accuracy, and the service provider also provides cost is approximately $100–$150 per structure. The service pro-
feature-extraction services. Most State DOTs contract MTLS-data vides only the vertical and horizontal bridge clearances and clear-
collection from service providers for their asset management. Some ance diagrams as deliverable. Point-cloud data are usually not part
state DOTs perform their own data extraction as needed, while of the deliverables. The cost depends on the number of structures to
other DOTs rely on contractors to extract their data. For example, be surveyed and their geographic separation. For example, Caltrans
Tennessee DOT (TnDOT) had a contract with Mandli Communi- had a contract with a service provider for such services, Terrame-
cations (Madison, WI) to perform an MTLS scan of their entire trix, which processed 589 bridges for Caltrans and is currently
state highway network. The total size of resulting point-cloud documenting 7,250 additional bridges for horizontal and vertical
and digital image data was 15 TB. The operation took 2 years clearance. The bridge office also needs accurate as-built plans
to complete at a cost of approximately $70=mi ð$43=kmÞ. The data of existing bridge-structure conditions to address litigation con-
extraction cost was an extra $62=mi ð$38=kmÞ. It should be noted cerning ongoing accidental bridge strikes, which average about
that data extraction costs depend on the number of features to be one a month. Bridge-repair and court costs include proving damage
extracted and can vary greatly. (change detection, based on existing scans). Not all structures
Feature extractions typically take about 2–5 human-hours=mi. would require the same level of detail. For example, over/under
Table 4 lists the cost breakdown for contract services with each data crossings would be different than river crossings. The deliverable
collection having a period of 2 years. The data-extraction costs are of this option does not provide accurate as-built data. WSDOT
usually higher for the first cycle in data collection since most assets has a total of 2,700 structures. Thus, the total cost of surveying
industry where asset costs are high. Fractional ownership reduces data. The plan provider will supply the personnel and vehicle
the cost of entry and risk of technological obsolescence. The ad- for data collection (100 days=year for 50% ownership) and raw-
vantage of this option is that the contractor will supply the equip- data postprocessing. The customer (a state DOT) will have to per-
ment, personnel, and GNSS/IMU postprocessing. The state DOT is form its own data extraction using in-house personnel and software.
not burdened with training and equipment maintenance. On the Similar to previous survey-grade options, the data extraction is
other hand, the contracting process could be complicated and estimated to take about 3 h=mi ð1.9 h=kmÞ at $50=mi ð$31=kmÞ.
can take a long time. Furthermore, the state DOT would not have Table 5 lists the Caltrans partial-ownership cost breakdown for
an MTLS system available at a moment’s notice. The equipment pavement survey, while Table 6 lists the Caltrans partial-ownership
availability is lower than the rental or the purchase option. Under cost breakdown for bridge-clearance measurement. Partial owner-
the fractional-ownership plan, customers purchase a fraction (20, ship is not cost-effective for bridge-clearance measurement.
30, 40, or 50%) of the system. The customer (for example, a state
DOT) pays an initial fractional-equipment cost based on the per-
centage of the fractional ownership. The plan also requires a Caltrans 50% Fractional Ownership of a Survey-Grade
MTLS
monthly management and maintenance fee ($18,000/month for
50% ownership). The customer also pays a fixed per mile usage Under the 50% fractional ownership plan, a customer may have
and data-processing fee [$75=mi ð$47=kmÞ] for the point-cloud up to 100 days of use and 25 times access per year. Fractional
$1,387,500
$5,633,000
$5,047,500
$5,602,500
$5,615,100
$6,140,100
$3,375,412
Total 6 yrs
(3 cycles)
ownership can be a cost-effective solution for pavement survey
savings
with careful planning and time coordination to complete a lot
K of projects in the 100 days available. In addition, the DOT
does not require trained personnel to operate the MTLS system.
Subsequent-
However, availability of the MTLS system is limited. Fractional
$717,000
$462,500
$711,304
$1,179,500
$1,151,000
$1,441,000
$1,340,200
$1,770,200
savings
cycle ownership is not a cost-effective solution for bridge-clearance
J
$462,500
$2,811,500
$3,274,000
$2,745,500
$2,720,500
$2,934,700
$2,599,700
$1,952,804
First-cycle
savings
for Caltrans. These tables indicate that in all the options, approxi-
savings
(tons)
CO2
23
23
23
29
29
29
H
4
3
4
or urban area. The feature-extraction time estimate provided in
Table 4 is based on discussions with contractors and feature extrac-
savings
8
0
purchase options are very close. However, the subsequent data col-
E
0
1
lection–cycle costs for the purchase options are much lower. Thus, in
the long run, the purchasing options seem to be a better alternative
when cost is the only consideration. On the other hand, if the data-
postprocessing
Annual FTE
0
9
11
11
11
options may be better, again considering the cost as the only con-
D
collection
MTLS system would be shared 50% of the time so the work output
required
for data
would be 50% of all the other options. In the near future, data-
1.5
1.5
C
0
1
and processing
data-collection
$2,466,000
$1,694,500
$1,404,500
$2,413,300
$1,983,300
$3,042,196
$337,500
$3,217,000
$2,945,500
$2,970,500
$3,664,300
$3,999,300
$4,646,196
$337,500
First cycle
Table 7. Itemized Savings for Each WSDOT-MTLS Option
cost
Utilization rate is the key factor in reducing the average cost per
A
mobile-LiDAR system
mobile-LiDAR system
mobile-LiDAR system
mobile-LiDAR system
(mapping-grade)
services
for use in other state DOT processes. Not all options may satisfy
$14,000
all of a state DOT’s program requirements. Table 8 summarizes
compatibility of each option for WSDOT programs. The check-
$12,000 mark indicates compatibility.
$120,000
y = 4125.4x + 74571 feature data-collection cost, and the total feature data–collection
$100,000 cost for the entire Washington state is estimated at $5,691,000
$80,000 [$813=mi × 7,000 mi ð$504=km × 11; 300 kmÞ] for the first cycle
$60,000 of data collection. The cost for subsequent cycle of data collections
will be less because most data has already been collected in the first
$40,000
cycle. This analysis assumes that data collection for the subsequent
$20,000
cycles of data collections would be half of the first cycle or approx-
$0 imately $2,845,500.
0 2 4 6 8 10 12 14 16
Highway Miles
Bridge-Clearance Measurement
Fig. 5. Historical MTLS pavement survey–project costs (with feature Based on the BPO’s estimate, the WSDOT BPO spends
data extraction and ground-control setup) versus project length (high- $80,000=year to collect bridge-clearance data, and the data-
way miles) collection cycle for the entire WSDOT’s bridge clearances takes
approximately 10 years to complete. Therefore, if a survey-grade
The authors gratefully acknowledge the Washington State Depart- Institute and State Univ., Blacksburg, VA.
ment of Transportation (WSDOT) for the support of this work Johnson, W. H., and Johnson, A. M. (2012). “Operational considerations
through contract GCA6059 and California Department of Trans- for terrestrial laser scanner use in highway construction applica-
portation (Caltrans) through contract 65A0368 to the Advanced tions.” J. Survey. Eng., 10.1061/(ASCE)SU.1943-5428.0000084,
214–222.
Highway Maintenance and Construction Technology (AHMCT)
Lasky, T. A., Swanston, T., and Ravani, B. (2006). “STRUCTVIEW:
Research Center at the University of California, Davis. The authors
A vehicle-based laser scanning system for roadway structure
thank the WSDOT for their support; in particular the guidance and
profile assessment.” IEEE Int. ITS Conf., IEEE, Piscataway, NJ,
review provided by WSDOT Technical Advisory Group (TAG) 331–336.
members: Mark Finch, Heath Bright, Lori Beebe, Eric Jackson, Ro- Mendenhall, S. (2011). “Moble laser scanning—Caltrans evaluates the
ger Caddell, Rhonda Brooks, Scott Campbell, Tom Clay, Marc technology’s costs and benefits.” CE News, 23(5), 42–43.
Faucher, Kurt Iverson, David Luhr, Rick Mowlds, Steve Palmen, Mrstik, P., and Newby, S. (2005). “Highway mapping in Afghanistan.”
John Tevis, and Brent Schiller for their valuable participation and Geospatial Solut., 58–59.
contributions. The authors would also like to thank Larry Orcutt, Newby, S., and Mrstik, P. (2005). “LIDAR on the Level in Afghanistan.”
Mark Turner, Robert McMillan, Nelson Aguilar, and Kevin Akin of GPS World, 16(7), 16–22.
Caltrans for their input and support. Moreover, the authors express Olsen, M. J., et al. (2013). “Guidelines for the use of mobile LiDAR in
gratitude to the mobile-LiDAR vendors—Eagle Mapping (Port transportation applications.” NCHRP 15–44 Rep. 748, National
Coquitlam, BC, Canada), Earthmine (Berkeley, California), ESM Cooperative Highway Research Program, Transportation Research
Consulting (Everett, Washington), Mandli Communications Board, Washington, DC.
(Madison, WI), Optech (Vaughan, ON, Canada), The PPI Group Sandrone, F., and Wissler, R. (2011). “Laser scanning images analysis for
(Portland, OR), and Topcon (Livermore, CA)—for their pilot-study tunnel inspection.” Proc., 12th Int. Congress on Rock Mechanics—
participation, efforts, and generous support; this study would not Harmonising Rock Engineering and the Environment, CRC, Boca
have been feasible without their gracious assistance. We also thank Raton, FL, 454–456.
Chris Harris of Tennessee DOT (TnDOT) and Ron Singh of Oregon Steering Committee on Space Applications, and Commercialization.
DOT (ODOT) for their valuable input. The authors also acknowl- (2003). Using remote sensing in state and local government: Informa-
edge the dedicated efforts of the AHMCT research team members tion for management and decision making, National Research
Council of the National Academies, Washington, DC.
who have made this work possible.
Terrametrix. (2013). “Home page.” 〈http://www.terrametrix3d.com/〉 (Sep.
11, 2013).
References WSDOT. (2013). “WSDOT GeoMetrix.” 〈http://www.wsdot.wa.gov/
mapsdata/geometrix/default.htm〉 (Sep. 11, 2013).
Chang, J. R., Chang, K. T., and Chen, D. H. (2006). “Application of 3D Yen, K. S., Ravani, B., and Lasky, T. A. (2011). “LiDAR for data
laser scanning on measuring pavement roughness.” J. Test. Eval., 34(2), efficiency.” Rep. No. WA-RD 778.1, Washington State Dept. of
83–91. Transportation, Olympia, WA.