Automation in the long haul: Challenges and opportunities of autonomous heavyduty trucking in the United States
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
WORKING PAPER 2018-06
Automation in the long haul: Challenges
and opportunities of autonomous heavy-
duty trucking in the United States
Authors: Peter Slowik and Ben Sharpe
Date: March 26, 2018
Keywords: Autonomous, heavy-duty vehicles
1. Introduction The implications of autonomous • increased operational efficiency
trucking are broad and extend beyond (e.g., real-time planning, reduced
Research and development in autono-
the trucking sector to include infra- truck downtime); and
mous vehicle technologies has taken
structure, urban planning, cyber
place for more than two decades, • reduced labor costs (i.e., technol-
security, privacy, and insurance. Within
with interest and investment prolif- ogy reduces or eliminates the need
the freight trucking sector, many see
erating in recent years sparked by for human drivers).
a future where the technology dra-
breakthroughs in sensing, commu- matically alters the truck driver ’s This paper explores the state of auton-
nications, and computing technolo- responsibilities and may eventually omous trucking technology and the
gies. The majority of investments and eliminate the need for a driver. Several benefits and drawbacks of its adoption
media attention to date have been industry groups envision autonomous from multiple stakeholder perspec-
concentrated in the passenger vehicle vehicle technology as an attractive tives. We are especially interested in
space (Slowik & Kamakaté, 2017), yet return on investment, with potentially how autonomous technology will affect
the technologies and their capabili- large economic benefits. However, the fuel use and emissions in the on-road
ties carry over to freight trucks and extent of these benefits is generally freight sector This paper is also a first
the commercial vehicle sector. An unknown to date, and there are also step toward better understanding the
increasing number of stakeholders risks and drawbacks to adoption of existing regulatory landscape and the
are actively involved in bringing this autonomous trucking. From a typical
types of policy measures needed to
technology to on-road commercial fleet perspective, the potential impacts
responsibly bring fuel-saving autono-
vehicles—especially, tractor-trailers. of autonomous trucking include
mous trucking technology to market.
With heavy-duty tractor-trailers
• improved on-road safety (i.e., The data and analysis presented in this
accounting for a disproportionately fewer collisions and fatalities); study focus on North America. More
high share of negative impacts—
• greater fuel efficiency and research is needed to better under-
including local air pollutants, green-
reduced emissions (i.e., higher stand the challenges and opportunities
house gas emissions, and fuel con-
miles-per-gallon); presented by autonomous trucking in
sumption (Sharpe, 2017)—the sector
other regions around the world.
is ripe for the application of autono- • ease of driving (e.g., automation
mous technology, perhaps even more technologies increasingly control Table 1 outlines the levels of automa-
so than the passenger vehicle sector. vehicle functions); tion, as defined by SAE International
Acknowledgements: This work is supported by the 11th Hour Project of the Schmidt Family Foundation. Fanta Kamakaté provided valuable input to the
development of this report. Ulises Hernandez, Nic Lutsey, and Rachel Muncrief of the International Council on Clean Transportation and Mike Roeth of the
North American Council for Freight Efficiency provided critical reviews on an earlier version of the report. Their review does not imply an endorsement, and
any errors are the authors’ own.
© INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION, 2018 WWW.THEICCT.ORGAUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
(2014) and adopted by the U.S. federal hardware such as a steering wheel or We review and summarize the relevant
government, and offers examples. brake and accelerator pedals. research literature on automated and
The remainder of the paper is orga- connected heavy-duty vehicle tech-
From Level 0 to Level 5, the auto-
nized as follows. In Section 2, we nologies, costs, deployment, and the
mated vehicle system progressively
summarize the current state of auton- implications of truck platooning and
handles additional tasks and increas-
omous trucking technology, high- other technologies on fuel economy.
ingly monitors the driving environ-
lighting the relevant technologies, Our review includes literature from
ment. Level 0 trucks may still include
costs, and demonstrations, empha- independent researchers, academia,
advanced technologies such as active
sizing those that promise to improve national laboratories, federal agencies,
safety systems (e.g., automatic emer-
fuel economy. Section 3 introduces nongovernmental organizations, and
gency braking) or warning features industry stakeholders.
(e.g., lane departure warning), but several societal acceptance con-
siderations related to autonomous Automated and connected heavy-
b e c a u s e t h e s e fe at u re s p rov i d e
trucking and discusses the potential duty vehicle technologies, costs, and
momentary intervention and are not
benefits and drawbacks of technol- deployment. Table 2 outlines several
sustained, they are considered Level
ogy adoption. Section 4 outlines examples of automated and connected
0 as defined by the J3016 standard
the current policy landscape in the vehicle technologies and technology
p u b l i s h e d by SA E I n te r n a t i o n a l
United States. Section 5 summa- applications identified in the heavy-
(2016). As shown in Table 1, heavy-
rizes the findings from 15 interviews duty trucking space. Broadly speaking,
duty trucks capable of Level 0 and
with industry experts that explore a handful of sensor, communication,
Level 1 automation are commercially technical, economic, and societal and processing software technolo-
available today, and trucks with Level barriers to higher levels of automa- gies are enabling varying degrees of
2 capabilities are rapidly nearing tion in trucking, as well as the poten- trucking autonomy by commanding
commercialization. A small number of tial ways that policy can effectively actuators such as steering and braking.
trucking demonstrations considered address these issues. In Section 6, As shown, vehicle sensor technologies
Level 3 have occurred to date. For we conclude by outlining several key include cameras, radar, LiDAR (which
autonomous vehicle systems Levels reflections from our research and stands for light detection and ranging),
3+, the technology is responsible for highlight areas for future work. and GPS units. Connected vehicle tech-
monitoring the driving environment nologies allow for communications
and is in control of vehicle functions
2. State of autonomous
across vehicles (vehicle-to-vehicle) and
and decision making. Autonomous trucking technology infrastructure (vehicle-to-infrastruc-
vehicles Level 4+ do not require This section provides an overview of ture), commonly referred to as V2V
human intervention, which means the current technology landscape for or V2I. The current leading vehicle
these vehicles theoretically could autonomous heavy-duty trucking and communications technology is dedi-
be manufactured without typical what it might mean for fuel economy. cated short-range communications
Table 1. Description of levels of automation
Technology
Level Name Description Examples status
Human performs all driving tasks, even if enhanced by Commercially
0 No automation Navistar LT, Peterbilt 579
active safety systems. available
Vehicle can perform sustained control of either steering Peloton Platooning System, Commercially
1 Driver assistance
or acceleration/deceleration. Volvo VNL available
Vehicle can perform sustained control of both steering
2 Partial automation Embark, Starsky Robotics Pre-commercial
and acceleration/deceleration.
All tasks can be controlled by the system in some Freightliner Inspiration,
3 Conditional automation Prototype retrofit
situations. Human intervention may be required. Uber ATG / Otto
All tasks can be handled by the system without
Research and
4 High automation human intervention, but in limited environments (e.g., Not currently available
development
dedicated lanes or zones).
Automated system can handle all roadway conditions Research and
5 Full automation Not currently available
and environments. development
2 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2018-06AUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
(DSRC), but research in 5G mobile sensing information from other nearby commercial vehicle sectors, as the core
network and other technologies could vehicles, allowing them to effectively sensing, communications, and software
also facilitate vehicle connectivity in “see” the road ahead beyond the imme- technologies are generally applica-
the future (National Academies of diate surroundings that are captured by ble to both the light-duty and heavy
Science, Engineering, and Medicine, the vehicles’ own sensing technologies. trucking sectors (National Academies
2017). Software is needed to process Many of these technologies are avail- of Sciences, Engineering, and Medicine,
information gathered from sensors and able today and are being purchased by 2017). The list of industry players is far
communications and control vehicle several fleets (North American Council from exhaustive; Comet Labs mapped
functions. Autonomous and connected for Freight Efficiency [NACFE], 2016). out a chart of more than 250 companies
vehicle technologies frequently are dis- The farthest right column of the table that are pursuing autonomous vehicles
cussed together because of the syner- shows examples of companies that are (Stewart, 2017). Furthermore, we note
gies between them, yet the technolo- involved in manufacturing one or more that many companies do not disclose
gies may be deployed and adopted autonomous and connected vehicle information about which technologies
separately. Connectivity allows autono- technologies. Many of these companies they manufacture in-house and which
mous vehicles to process the additional are active in both the passenger car and they purchase from parts suppliers.
Table 2. Example automated and connected vehicle technologies in on-road heavy-duty trucking
Example
Commercially technology
Technology Components Description available? makers
Used to identify other objects using visible light. Cameras have
limitations compared to other sensor technologies and function Continental,
Cameras Yes
poorly in darkness, extremely bright light, and certain weather Mobileye, Delphi
conditions.
Bosch,
Used to identify the velocity, direction, and distance of other objects
Radar Yes Continental,
by emitting high-frequency radio waves.
Autoliv, Delphi
Sensors
Considered the most reliable and robust sensing technology. LiDAR Velodyne LiDAR,
measures the range and speed of objects using reflected light. LeddarTech,
LiDAR Range and speed are measured based on the time that laser light Yes Quanenergy,
takes to reflect. LiDAR systems can process and record images. Delphi, Strobe,
(National Highway Traffic and Safety Administration [NHTSA], 2013). Waymo
Used to identify vehicle position and velocity by communicating Linx
GPS Yes
with satellite signals. Technologies
Dedicated short range communications (DSRC) is two-way
communications in 5.9 GHz band that allows for high data
transmission over a moderate range. DSRC allows for V2V and V2I
DSRC communications which can send messages and provide alerts to Yes NXP, Qualcomm
drivers in real time. The U.S. Department of Transportation (DOT,
2017a) considers this technology the basis for intelligent vehicle
safety application integration.
Communications
The 5th generation wireless systems currently under development
will allow for higher capacity and better coverage with less latency.
5G is expected to support device-to-device communications
Not currently
5G with increased reliability. 5G is believed to be a promising No
available
communications technology with applications for connected and
autonomous vehicles within the next decade (National Academies of
Sciences, Engineering, and Medicine, 2017).
Millions of lines of software code enable autonomous and connected
trucking. Computing software systems are used to process images
Nvidia, Intel,
Algorithms, captured from sensor technologies, interpret communications
Yes, with Autoliv, Cisco,
Software artificial messages from other vehicles or infrastructure, and control
limitations Uber ATG, many
intelligence vehicle functions in real time. Software refinement and validation
others
is considered a much larger challenge than deploying sensor and
communication hardware (Tesla, 2016).
WORKING PAPER 2018-06 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 3AUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
Sensors, communication, and process- volumes and broader commercializa- commercial adoption. For example,
ing software enable a variety of vehicle tion. This information is needed to Meritor Wabco’s OnGuardActive col-
applications including but not limited better assess the value proposition for lision mitigation system is reported
to driver alerts and collision avoid- fleets to adopt various autonomous to be in 120,000 heavy-duty trucks
ance systems, automatic braking, lane trucking technologies. There are several (Meritor Wabco, 2017). Not exclu-
keeping assistance, adaptive cruise potential direct and indirect economic sive to heavy-duty trucks, Mobileye’s
control, platooning, and eco-driving benefits that autonomous trucking Advanced Driver Assistance System
optimization. Many of these technology may offer, and a deeper understanding is used in nearly 15 million vehicles
applications are described in greater of the costs will better inform the pace worldwide (Mobileye, 2017b). A study
detail in Table 3. A few examples of and scale of technology adoption. by Rodríguez, Muncrief, Delgado, and
companies that offer, or seek to offer, Baldino (2017) estimated the market
In addition to what the research lit-
automated and connected vehicle penetration of several heavy-duty
erature reveals about autonomous
technology applications in the trucking vehicle technologies in the United
trucking technology costs, rough esti-
sector are shown in the column farthest States and the EU, including auto-
mates can be compiled using infor-
to the right. mated manual transmissions (AMTs),
mation from third-party parts sup-
The market demand for autonomous predictive cruise control (PCC), and
pliers. For example, a review of the
vehicle technologies of low and high adaptive cruise control (ACC). The
parts components for Meritor Wabco’s
levels of automation is quite strong, 2015 U.S. and EU market penetration
OnGuardActive system, a radar-based
and numerous industry groups are of these technologies in new tractor-
active collision mitigation and adaptive
aggressively developing products. In trailers was 28% (U.S.) and 70% (EU)
cruise control system, on the inde-
the future, additional advancements for AMTs, 3% (U.S.) and 20% (EU) for
pendent truck parts marketplace fin-
in vehicle sensor quality, bandwidth PCC, and 10% (U.S.) and 50% (EU) for
ditparts.com website suggests system
availability for vehicle communica- ACC. More research is needed to more
costs of around $2,500 to $3,500. This
tions, and processing software and fully identify the suite of autonomous-
is roughly in line with the costs for other
algorithms are likely to enable more vehicle-related technologies that have
similar technology applications docu-
robust technology applications and been adopted to date and the fleets
mented in Table 4. There also has been
higher levels of automation. that are adopting them.
some speculation in the media about
Return on investment (ROI) is often technology costs. As reported for the Technological barriers remain for com-
a key factor influencing the adoption American Transportation Research mercial deployment of heavy-duty
of new technologies on freight appli- Institute by Short and Murray (2016), truck platooning and higher levels
cations. Upfront technology costs additional technology costs for Uber’s of automation (Levels 3+). Despite
and per-mile operating costs are core retrofitted long-haul truck—believed to current barriers, researchers and
components that influence the value be Level 3, as identified earlier in this industry stakeholders have made pre-
proposition for fleet adoption. Table 4 report—as well as Daimler’s Freightliner dictions about the commercial avail-
shows some estimates of the upfront Inspiration, which also is believed to ability and uptake of autonomous
costs for several examples of autono- be Level 3, have been estimated at trucks. A few of these predictions are
mous heavy-duty vehicle technologies $30,000 per truck (McNabb, 2015; documented here.
and systems. Stewart, 2016). These cost estimates • NACFE (2016) reports that “it is
for a Level 3 truck are approximately extremely likely that in the near
Several of the technologies shown in
twice the estimates reported in Roland future, Class 8 tractors will be sold
Table 4 cost only a few hundred dollars.
Berger (2016). One reason for the large as platooning capable ‘right out of
These include driver assistance and
discrepancy may be the difference in the box’” (unspecified, assumed
driver alert systems, which typically are
assessing the cost of a single retrofit Level 2).
purchased for safety and collision miti-
prototype versus assuming some level
gation. While there are some estimates, • Commercial deployment of pla-
less information is known about the of market adoption and achieving
tooning applications (unspecified,
technology costs for trucking automa- economies of scale.
assumed Level 2) could occur
tion for Levels 3 and above, especially Several of the technologies and around 2020 on select U.S. roads
the projected cost reductions that systems shown in Table 4 are available (National Academies of Sciences,
would stem from increased production today, and some have had notable Engineering, and Medicine, 2017).
4 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2018-06AUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
Table 3. Examples of automated and connected vehicle technology applications in on-road heavy-duty trucking
Technology Technologies Commercially Example
applications used Description available? companies
Sensors such These systems send an audible or haptic warning to drivers when there is
Mobileye,
Lane departure as cameras, risk of the vehicle unintentionally drifting outside of the lane. This technology
Yes Meritor
warning processing is considered Level 0 because it does nothing more than alert a driver.
Wabco
software (National Academies of Sciences, Engineering, and Medicine, 2017).
Sensors such
Blind spot detection devices can detect if other vehicles are located in the
as cameras Mobileye,
Blind spot driver’s blind spots and notify the driver. The alerts can be audible, haptic,
and radar, Yes Meritor
detection or visual. Like lane departure warnings, blind spot detection alerts are
processing Wabco, Volvo
considered Level 0.
software
Sensors such Scania, DAF,
Automatic braking systems can detect the speed and distance of vehicles
as cameras Daimler,
Automatic ahead of them and automatically apply the brakes if needed. This technology
and radar, Yes Meritor
braking is considered Level 0 because the feature provides momentary intervention
processing Wabco, Volvo,
and is not sustained.
software Bendix
Automated Electronic Automated manual transmissions control the operation of the clutch and
manual control unit, gear selection automatically, based on information gathered from vehicle Eaton, Volvo,
Yes
transmissions hydraulics, sensors. AMTs are an enabling technology and are generally required on all Daimler
(AMT) software Level 1+ autonomous trucks.
On-board
diagnostics,
A system that monitors human driving and provides real-time advice and TomTom,
Eco-driving monitoring
feedback for drivers to achieve greater fuel performance, for example by Yes Ruptela,
systems and processing
moderating highway speed and by smoothing acceleration and braking. SmartDrive
software,
telematics
Sensors such
as cameras These systems monitor the vehicle placement within road lane markings. If Scania,
Automated
or radar, the vehicle is departing the lane, the system corrects the lateral direction Yes Meritor
lane keeping
processing automatically. The technology is considered Level 1. Wabco
software
Sensors such
Adaptive cruise control adjusts vehicle speed, controlling throttle and Meritor
Adaptive cruise as radar,
braking, based on the speed of the vehicle in front of it in order to maintain a Yes Wabco, DAF,
control (ACC) processing
set distance. ACC technology is considered Level 1. Volvo, Bendix
software
GPS, Predictive cruise control combines cruise control with GPS and topographical
Predictive topographical data inputs, altering vehicle speed to optimize performance over various
Kenworth,
cruise control mapping data, types of terrain. PCC technology provides maximum benefits in conditions Yes
DAF
(PCC) processing with rolling hills. The technology is considered Level 1. PCC and ACC can be
software active simultaneously or the functions could be offered separately.
Sensors such
as radar, Platooning is when groups of vehicles travel close together to minimize
processing aerodynamic drag. Truck platooning typically includes sets of two or three Yes (Level 1),
Peloton,
software, trucks paired together using sensor and communication technologies. Level 2 systems
Platooning Volvo, Uber
could also At basic levels, ACC alone (Level 1) could enable truck platooning. More are pre-
ATG, Daimler
include vehicle advanced platooning technology controls for both longitudinal (ACC) and commercial
communications lateral (automated lane keeping) movements and is considered Level 2.
using DSRC
Will likely
Highly automated trucks will be capable of operating autonomously without
include
Highly human intervention in limited environments such as dedicated areas or
cameras, radar, Daimler, Uber
automated highway lanes. Highly automated trucks (Level 4+) are not commercially No
LiDAR, DSRC, ATG
trucking available for on-road applications today, but there are a few examples of
processing
their use in mining and farming operations.
software.
Telematics systems combine telecommunications and informatics, which
GPS, DSRC, or
is the collection, classification, storage, retrieval, and dissemination of
other wireless
information. Telematics equip fleet managers with valuable real-time data
communications Zonar,
such as vehicle location, speed, service needs, weather, road conditions, and
Telematics technology, Yes Geotab,
driver performance. Telematics are expected to complement connected
asset Openmatics
and autonomous vehicles, for example by enabling the transmission and
management
processing of communications data from nearby vehicles, or by facilitating
software
identifying opportunities to link vehicles to form a platoon.
WORKING PAPER 2018-06 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 5AUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
Table 4. Estimated costs for examples of autonomous and connected truck technologies and technology applications
Technology or
Study or reference application Technology description Cost Time frame Notes
“A few
Cost estimates are per unit. Companies typically install
Considered the most robust $75,000 years ago”
one to four LiDAR units per vehicle. Waymo CEO John
Waymo (2017) LiDAR sensing technology for processing (unspecified)
Krafcik revealed the company has reduced the cost of
images.
$7,500 2017 $75,000 “top-of-the-range” LiDAR units by 90%.
$100 to
Nordrum, A. (2016) DSRC modules V2V communications hardware. Around 2016 Cost estimates are for DSRC module made by NXP.
$200
NHTSA estimates the cost of V2V equipment and
communications functions for light-duty vehicles. The
V2V V2V communications equipment $341 to
Harding et al. (2014) 2020 technologies include DSRC transmitter/receiver, DSRC
communications and functions. $350
antenna, electronic control unit, GPS, GPS antenna, wiring,
and displays.
U.S. Environmental A transmission that facilitates truck $5,100 2013
Automated EPA and NHTSA estimate the cost of automated manual
Protection Agency shifting by utilizing a computer and
manual transmissions for medium- and heavy-duty vehicles and
[EPA] and NHTSA eliminating the manual shifter and $3,750 2018
transmission report the values in 2013 dollars.
(2016a) clutch.
National Academies of A system of sensors that identifies
Blind spot $250 to Available
Sciences, Engineering, vehicles in the driver’s blind spots Cost estimates are for aftermarket system cost.
detection system $850 today
and Medicine (2017) and provides a warning.
A driver alert safety package that offers a variety of alerts
National Academies of Mobileye and driver assistance features including forward collision
Driver assistance through collision $850
Sciences, Engineering, Advanced Driver Available warning, lane departure warning, headway monitoring
avoidance intelligent vision sensor with $150
and Medicine (2017), Assistance today and warning, pedestrian and cyclist warning, intelligent
technologies. installation
Mobileye (2017a) System high beam control, turn signal reminder, and low visibility
indicator.
The collision mitigation system also includes adaptive
Radar-based sensor system
cruise control and active braking applications. More than
Meritor Wabco Meritor Wabco identifies potential collisions and Not Available
120,000 OnGuard collision mitigation systems have been
(2017, n.d.) OnGuardActive sends warning notifications to disclosed today
sold in North America and are being used by more than
drivers.
200 fleets.
Vehicle technology to dynamically $3,000 Around 2006 Cost estimates not explicit to heavy-duty vehicles.
Adaptive cruise control longitudinal movement
DOT (2014) Assumed to include sensing technologies (cameras,
control and maintain consistent following
$2,000 Around 2014 radar) and processing software.
distance.
A technology that alters vehicle
International Council on
Predictive cruise speed to optimize performance The study reports the estimated 2030 vehicle technology
Clean Transportation $760 2030
control over various types of terrain based costs and reports the values in 2015 dollars.
(ICCT, 2017)
on GPS and topographical data.
A technology that alters vehicle $953 2018 EPA and NHTSA estimate the cost of predictive cruise
Predictive cruise speed to optimize performance
EPA and NHTSA (2016) control for heavy-duty tractors and reports the values in
control over various types of terrain based
2013 dollars.
on GPS and topographical data.
$766 2027
Cost estimate indicates the advertised cost (excluding
A technology that alters vehicle
$1,300 with VAT) in Germany to purchase and install the retrofit
Predictive cruise speed to optimize performance
Daimler AG (2015) installation 2015 technology. Based on typical mileage of 81,000 miles/
control over various types of terrain based
(€1,500) year, the technology payback period from fuel savings
on GPS and topographical data.
(up to 5%) is advertised as less than 1 year.
American Trucking
Available Study not specific to heavy-duty vehicles. Together,
Associations Adaptive cruise
Vehicle technologies for today adaptive cruise control and lane keeping assist are
Technology and control and lane $3,000
longitudinal and lateral controls. (in light-duty considered Level 2 by enabling the system to control both
Maintenance Council keeping assist
vehicles) longitudinally and laterally.
(2015)
About Includes V2V communication technology and “necessary
Technology that enables vehicles
Janssen, Zwijnenberg, $11,900 additional safety measures” which are unspecified but
Platooning to travel close together to minimize 2015
Blankes, & Kruijff (2015) per truck assumed to include sensor systems such as LiDAR, radar,
aerodynamic drag.
(€10,000) and/or cameras.
Technology that enables vehicles $1,500 – Estimated cost of required technologies to enable two-
NACFE (2016) Platooning to travel close together to minimize $2,000 per 2016 truck platooning, based on industry interviews from
aerodynamic drag. truck unnamed fleet manager and technology developer.
Level 1 $1,800
Study estimated the incremental costs of adding
Level 2 $6,900 technology to enable Level 1 through Level 5 automation.
Incremental technology costs
Total incremental technology cost to reach Level 5 is
Roland Berger (2016) Level 3 (above Level 0) for Level 1 to Level $13,100 Unspecified
estimated at $23,000. Incremental technologies include
5 truck automation.
Level 4 $19,000 hardware (sensors, communications) and additional
processing software.
Level 5 $23,400
6 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2018-06AUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
• Level 3 automation capabilities communications, and vehicle control from trial to trial, although the general
are most likely to come within a technologies are enabling much shorter magnitude of team savings—average
decade for heavy trucks (American following distances due to electronic savings of both the lead and platooned
Trucking Associations Technology systems that read and react to the vehicle—is approximately 4% to 15%.
and Maintenance Council, 2015). driving conditions several times faster Because platooning requires at least
• The International Transport Forum than a human driver. Reliability of these two vehicles and relatively fast speeds,
(ITF, 2015) predicts that Level 4 technologies is becoming increasingly there is some limitation to the percent-
trucking on highways could be important as the required reaction age of fleet-miles that can occur in
available before 2030. times for safe operation reach levels a platoon and thus the realized fuel
beyond human capabilities (NACFE, benefits. Furthermore, the literature
• A study requested by the European 2016). Relatively advanced platooning reveals that several variables have an
Parliament finds that platooning systems that control both following
technology will allow truck drivers impact on fuel savings, including fol-
distance and steering are likely needed
to legally disengage from the lowing distance, travel speeds, and
to minimize driver error and maximize
driving task within 10 to 20 years. vehicle weight. There does not appear
the fuel savings benefits that platoon-
Fully driverless trucking (Level 5) to be a consensus in the literature
ing promises.
could emerge after 2035 (Frisoni about how to ideally optimize fuel
et al., 2016). Suppliers, truck manufacturers, and savings from truck platooning when
freight operators have interest in pla- including each of the factors above.
• Early adoption of highly auto- tooning technology because of the Here are several research findings on
mated trucks (Levels 4+) may attractive fuel cost savings and return the relationships between fuel savings
occur in the form of a trailing truck on investment that the technology and the various factors:
in a platoon, following closely promises. However, freight operators
behind a driver-assisted (Level 1) are likely to chart their own unique • A review of several platoon-
truck (NACFE, 2016). paths to technology adoption. For ing tests found fuel savings
example, Auburn University (2017) generally increase as the gap
Heavy-duty truck platooning dem-
notes that perspectives on fuel savings b e t we e n ve h i c l e s d e c re a s e s
onstrations and implications for fuel
might differ for large versus small (National Academies of Sciences,
economy. A major focus of the research
fleets. For large fleets with economies Engineering, and Medicine, 2017).
literature and industry R&D efforts to
of scale, fuel savings alone can be suf- • Auburn University (2017) along
date has been on truck platooning,
ficient motivation. However, for smaller with industry partners including
driven partially by the potential fuel
fleets, low upfront costs are crucial and
savings and attractive return on invest- Peloton, Peterbilt, Meritor Wabco,
additional benefits may be needed for
ment that can result. Platooning tech- and the American Transportation
small fleets to adopt the technology.
nology combines safety and collision Research Institute (ATRI) (Short &
Key priorities for fleets to adopt driver
mitigation technologies with vehicle Murray, 2016) conclude that “trucks
assistive truck platooning technology
communications and automated vehicle should be spaced as close as safely
(Level 1) include affordability, ability
controls to tether trucks together in feasible” to optimize combined
to coexist with collision mitigation
formation (NACFE, 2016). As noted in fuel economy.
systems, and availability as a retrofit
Table 3, basic levels of platooning can
device. In the near term, platooning • Lammert et al. (2014) found 50 feet
be realized by adaptive cruise control
technology is more likely to be adopted to be the optimal distance for team
alone, while more advanced platooning
within fleets, rather than across fleets, savings, with savings decreasing
controls both longitudinal (ACC) and
until trust, assurance, and interoperabil- by about one-third with 20-foot
lateral (automated lane keeping) move-
ity is established among fleet operators following distances.
ments and is considered Level 2.
(Auburn University, 2017).
In the most basic form, platooning • Simulations by Auburn University
Numerous demonstrations and tests (2017) suggest that 2-foot lateral
could be conducted manually, which is
have occurred in recent years that offsets can increase the coef-
to say without automation, simply by
driving with short following distances; quantify the fuel savings from truck pla- ficient of drag by up to 30%,
however, this method poses significant tooning; many are outlined in Table 5. thereby squandering the aerody-
crash risk and safety considerations. As shown in the table, the realized fuel namic gains and fuel savings of
Emerging sensor technologies, vehicle savings from truck platooning varies platooning.
WORKING PAPER 2018-06 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 7AUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
Table 5. Fuel savings demonstrated in example truck platooning projects
Lead Platooned
Source vehicle vehicle(s) Team Study method Technologies used Description
Testing of one platooned and one lead
truck at following distances from 30 to
Radar, DSRC-based V2V 150 feet at 65 mph. Tests were conducted
Evaluated
Auburn communications, satellite at the Auburn test track using Peterbilt
0.4% to 8.6% to 4.5% to using “SAE
University positioning, actuation for 579 tractors with 53-foot trailers using
5.3% 10.2% 7% Type II FE test”
(2017) vehicle controls, and human- Peloton’s truck platooning system. Because
at TRC Ohio
machine interfaces longitudinal movement is automated, and
drivers were responsible for steering, the
technology is considered Level 1.
DSRC V2V communications,
Peloton
Real-world radar collision mitigation Testing of one platooned and one lead
Technology 4.5% 10% 7%
testing system, front facing camera, truck at a following distance of 36 feet.
(2017)
GPS
Testing of one platooned and one lead
Lammert, Radar, DSRC V2V
Peterbilt Class 8 tractor-trailers vehicles at
Duran, Diez, communications, vehicle
2.7% to 2.8% to 3.7% to Evaluated on the Continental Tire Proving Ground test
Burton, & braking and torque control
5.3% 9.7% 6.4% test track track in Texas. Conducted with varying
Nicholson interface, cameras, driver
speeds, following distances, and vehicle
(2014) displays
weights.
Safe Road
Testing of one platooned and one lead
Trains for the
Camera, radar, and laser Volvo FH12 rigid truck at the IDIADA test
Environment Not Evaluated on
2% to 8% 8% to 13% to support adaptive cruise track in Spain at following distances of 16
Project reported test track
control, V2V communications to 82 feet (5 to 25 meters) at 53 mph (85
(SARTRE,
km/h).
2014)
Testing of one platooned and one lead
Peterbilt 386s model year 2011 tractor
Not Real-world trailers in Salt Lake City, Utah. Conducted
NACFE (2013) 4.5% 10% Radar
reported testing on I-80 at 64 mph with 36-foot following distance
using Peloton platooning technology.
Vehicles were fully loaded.
Testing of two platooned trucks and one
Radar, laser scanner, lead truck at the AIST test track in Japan
Tsugawa 12% to Evaluated on
0% to 9% 9% to 15% adaptive cruise control, V2V traveling 50 mph (80 km/h) at distances
(2013) 22% test track
communications from about 15 to 65 feet (4.7 to 20 meters).
Vehicles were unloaded.
Testing of one platooned and one lead
Browand, Freightliner 2001 Century Class tractor-
McArthur, 10% to Evaluated on Electronic longitudinal control trailers at the Crows Landing runway in
5% to 10% 8% to 11%
and Radovich 12% test track system California. Conducted with varying speeds
(2004) and following distances, and the trucks
were empty.
Testing of one platooned and one lead
Mercedes-Benz ACTROS semi-trailer
trucks at the Papenburg test track in
Not Evaluated on Germany. Conducted at 37 mph and
3% to 9% 9% to 21% Electronic tow bar
reported test track 50 mph (60 km/h and 80 km/h) with
Bonnet and following distances from about 15 to 53
Fritz (2000) feet (4.5 to 16 meters). Vehicles were
partially loaded.
Simulation to extrapolate potential fuel
Not
2% to 6% 13% to 17% Simulation Simulation savings at 50 mph (80 km/h) when trucks
reported
are fully loaded and weigh up to 40 tons.
Summary findings based on desk research,
Compilation
events, and industry interviews with fleets,
of literature
NACFE (2016) 3% to 5% 8% to 19% 4% Not applicable manufacturers, and platooning technology
review and
developers in North America. Values based
interviews
on following distance of 40 to 50 feet.
8 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2018-06AUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
• Ellis, Gargoloff, and Sengupta speed, gap distance, weight, number addition to platooning that can improve
(2015) found that platooning dis- of platooning vehicles, and the types fuel performance of heavy-duty
tances of 16 feet (5 meters) or less of trucks and their aerodynamic trucking, including predictive cruise
can have significant costs by reduc- profiles to optimize for fuel economy. control, adaptive cruise control, auto-
ing engine cooling air flow and Beyond optimizing for fuel savings, mated eco-driving, driver feedback
therefore requiring a cooling fan more research is needed to identify systems to promote eco-driving, and
and mitigating the potential fuel the implications of each of the factors automated manual transmissions.
savings that platooning promised. on safety, road infrastructure, public Adaptive and predictive cruise control
acceptance, and logistics. are considered Level 1. Automated eco-
• Lammert et al. (2014) found higher
driving was unspecified but assumed
average fuel savings at lower Testing of truck platooning typically
to be Level 1 by enhancing vehicle
speeds (55 mph versus 65 mph or has been conducted in relatively
acceleration and deceleration profiles.
75 mph). limited real-world on-road applica-
Automated manual transmissions and
tions. An important industry con-
• Lammert et al. (2014) found fuel driver eco-driving feedback systems
sideration for technology adoption
savings from platooning were do not automate either longitudinal
is the number of freight miles that
reduced with higher gross vehicle or lateral movement and therefore are
are suitable for platooning. National
weight. Similarly, Bonnet and considered Level 0. As shown in the
Renewable Energy Laboratory (NREL)
Fritz (2000) found through pla- table, the fuel benefit of eco-driving
researchers examined real-world truck
tooning simulations that greater feedback systems can be significant,
usage data in the United States to
truck weight reduced fuel savings at approximately 10%. However, the
statistically analyze the percentage of
benefits. materialization of this fuel consump-
miles suitable for platooning (Muratori
tion benefit will depend on the extent
• Increasing the number of trucks et al., 2017). By using recent data on
to which drivers actually use feedback
in a platoon could accrue highway vehicle usage and velocity,
and react appropriately, and therefore
more benefits as more vehicles they found that approximately 65%
may require monitoring verification and
realize the slipstream benefits of truck miles could be driven in a
incentives to maximize fuel benefits.
(National Academies of Sciences, platoon, at 50mph or greater, trans-
Engineering, and Medicine, 2017). lating to about 4% reduction in Less information is available on the
overall trucking fuel consumption in fuel-savings potential for technology
• For truck platooning systems
the United States based on a team applications such as blind spot detec-
using trucks with different aero-
fuel improvement of 6.4% as found in tion, lane departure warning, forward
dynamics, it is most favorable for
Lammert et al. (2014). More research collision warning, and other collision
the least aerodynamic truck to be
is needed to identify the percentage mitigation systems. Furthermore,
in a following position (Auburn
of fleet platoonable miles that can tip improving trucking fuel performance
University, 2017).
the fleet value proposition in favor of is not the intent of these types of
adopting platooning technology. technology applications. The safety
In addition to these factors, atmospheric
benefit and potential payback period
conditions like temperature, wind, and Fuel efficiency benefits of nonpla-
of collision mitigation systems are
humidity can have an impact on aero- tooning technologies. Platooning has
discussed in Section 3. Nevertheless,
dynamics and therefore on the ability received significant attention in the
collision avoidance systems may
of tractor-trailers to realize fuel benefits research literature and industry R&D
have an indirect relationship with
from platooning. Similarly, traffic con- efforts, but several other autonomous-
fuel economy. For example, Meritor
gestion, terrain, road construction, and vehicle-related technology applications
Wabco’s OnGuardActive system, which
other real-world factors can reduce the are poised to offer some fuel savings
includes collision mitigation as well as
feasibility and benefits of platooning. benefits as well. In this section, we
adaptive cruise control functions, could
NACFE (2016), for example, recom- highlight other autonomous trucking
help smooth acceleration and decelera-
mends fleet managers should expect technology applications and discuss
tion profiles and therefore enable more
smaller fuel savings than reported in the their implications on fuel performance.
efficient driving. The extent to which
demonstration projects, which are not Table 6 summarizes several research
these types of technology applications
representative of road traffic congestion. and demonstration projects.
result in real-world fuel economy gains,
More research is needed to identify As shown, our research identifies however, has yet to be quantified and is
the ideal combinations of travel several of technology applications in largely unknown.
WORKING PAPER 2018-06 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 9AUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
Table 6. Potential fuel savings of nonplatooning trucking applications
Technology Fuel efficiency
Source application improvement Notes
The real-world CO2 emissions and fuel consumption reduction benefit from
EPA & NHTSA Predictive cruise
2% predictive cruise control as estimated by the U.S. federal agencies in the
(2016a) control
Phase 2 medium- and heavy-duty vehicle regulatory impact assessment.
Investigated the fuel economy improvement realized through predictive
cruise control technology compared to conventional cruise control. Unlike
Lattemann, Neiss, 2.5% to 5.2%
Predictive cruise conventional cruise control, predictive cruise control takes road elevation
Terwen, and over conventional
control information into account using GPS and 3-D mapping information.
Connolly (2004) cruise control
Predictive cruise control was found to be more effective with greater
vehicle weights and on roads with more rolling hills.
Enabled by GPS and mapping technology, predictive cruise control saves
Predictive cruise
DAF (2017) 1.5% to 4% fuel by anticipating the road ahead and adjusting vehicle speeds to
control
optimize fuel consumption.
Lutsey, Langer, Predictive cruise
0% to 5% Estimates based on industry communication and stakeholder workshop.
and Khan (2014) control
The press release indicates that 5% fuel savings is the high end of what
Daimler AG Predictive cruise
5% the predictive cruise control technology can offer. Estimates are based on
(2015) control
typical annual mileage of 81,000 miles (130,000 kilometers).
Adaptive cruise
Faber et al. Estimates from the euroFOT project. ACC offers fuel consumption benefits
control, forward 1.9%
(2012) from smoother truck speed profiles.
collision warning
The authors report estimates of fuel savings for automated eco-driving
Automated
of 4% to 10%. Automated eco-driving elements are assumed to include
ITF (2017) eco-driving 4% to 10%
smoother acceleration and deceleration profiles which could stem from
(unspecified)
technologies like adaptive cruise control.
National Center
Results of a truck driving simulator. Eco-driving feedback technology
for Sustainable Eco-driving
11% provides driver feedback in real time, recommending optimal speed and
Transportation feedback system
alerting drivers in instances of aggressive acceleration and speed.
(2016)
1% to 3% Automated manual transmissions offer fuel savings from more efficient
Automated manual
NACFE (2014) reduction in fuel shifting. Estimates based on compilation of existing sources on the
transmissions
consumption technology performance and data as well as input from industry.
Lutsey, Langer, & Automated manual Estimates based on review of literature, industry communication, and
2% to 3%
Khan (2014) transmissions stakeholder workshop.
3. Societal acceptance and 2016; National Academies of Sciences, considerations related to automated
Engineering, and Medicine, 2017; ITF, trucking. It is likely that nearly all of
the benefits and drawbacks
2017). From the industry perspective, these acceptance issues will need to
of autonomous trucking new technologies and the potential be resolved or minimized before wide-
This section introduces several societal for reduced costs from fuel savings spread adoption of the technology.
acceptance considerations related to and collision avoidance can be appeal- A recurring theme among industry,
autonomous trucking and discusses ing. For drivers, platooning and higher drivers, and the public is the concern
the potential benefits and drawbacks of levels of automation could ease the over safety and system reliability.
technology adoption. tediousness of long shifts and allow Heavy-duty trucks by their very nature
drivers to engage in other tasks. Yet could cause significantly more harm
SOCIETAL ACCEPTANCE the research indicates that there are and damage compared to a passenger
several major challenges and barriers car. Research investigating the poten-
Several studies have outlined the
to widespread acceptance. Table 7 tial adoption of platooning technology
societal acceptance considerations
captures many industry, driver, and found that fleets want proof that the
of automated trucking, identify-
public acceptance concerns related to technology works and the ability to
ing both positive and negative per-
autonomous trucking. pilot and test the technology before
ceptions (Tsugawa, 2013; American
Trucking Associations Technology and As shown in the table, there are several investing (Auburn University, 2017).
Maintenance Council, 2015; NACFE, negative perceptions and acceptance There are also uncertainties related
10 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2018-06AUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
to the impact of automation on the driver comfort, stress, and vigilance • g re a t e r f u e l e f f i c i e n c y a n d
operating conditions for drivers. For (American Trucking Associations reduced emissions (i.e., more
example, as tasks are increasingly Technology and Maintenance Council, miles-per-gallon);
handled by the computer system, there 2015). Safety testing, reporting, • ease of driving (e.g., technology
is the risk of drivers becoming compla- and demonstrations of automated increasingly performs control of
cent, overly disengaged and unable to heavy trucks are needed and should vehicle functions);
reengage in a timely and safe manner, be available to the public (Short &
• increased operational efficiency
and experiencing high stress from Murray, 2016).
(e.g., real-time planning, reduced
close following distances. There is also
truck downtime);
the possibility of future job loss due to BENEFITS AND DRAWBACKS OF
high levels of automation. From indus- • additional road capacity (i.e., less
AUTONOMOUS TRUCKING
try’s perspective, studies suggest that gap between vehicles, better
fleet managers are unlikely to make Autonomous and connected trucking physical road usage);
major operational and logistical altera- has the potential to offer several
• reduced labor costs (i.e., tech-
tions to their freight schedules to take benefits, yet the extent of these
nology eliminates or reduces the
advantage of platooning. benefits is generally unknown to date.
need for human drivers).
Furthermore, there are several risks
Many of the considerations in Table 7 Some of the potential impacts of
and drawbacks to adoption of the
are somewhat uncertain and need to autonomous trucking are likely to be
technology. Multiple research studies
be investigated further. More research realized with no or low levels of auto-
(e.g., Short & Murray, 2016; NACFE,
is needed to more fully understand the mation (Levels 0–2), whereas others
2016; National Academies of Sciences,
impacts of automated trucks—from emerge with high or full automation.
Engineering, and Medicine, 2017; ITF,
lower levels of automation to highly For example, driver alert and collision
2017) outline the potential impacts of
automated trucking—on industry, avoidance systems (Level 0) can offer
autonomous trucking from the fleet
drivers, and the general public. For significant safety benefits, and truck
perspective, which include
example, studies could inform how platooning (Levels 1 or 2+) can offer
obstructed views from platooning • improved on-road safety (i.e., fewer considerable fuel savings. At higher
at close following distances affect collisions and fatalities); levels of automation (Levels 3+), the
Table 7. Industry, driver, and public acceptance concerns about autonomous trucking
Industry Driver General public
• Some fleet managers are unlikely to • Potential boredom and complacency • Safety, system security, and reliability.
make operational and logistical changes when the system is operating the vehicle. • Risk of hacking and hijacking a long-haul
or reroute trucks to take advantage of • Monotonous yet high stress when freight truck poses much greater danger
platooning. platooning at close following distances. than a passenger car.
• Privacy and access to key data • Risk that drivers get pushed to operate • Lack of awareness and familiarity.
and tracking by competitors and longer hours if disengaging means that
governments. • Trust over system reliability when driving
drivers are considered off the clock. next to a computer-controlled tractor-
• One accident could eliminate the • Risk of passenger cars breaking into the trailer.
monetary gain from platooning fuel platoon unsafely.
savings. • Ability to merge on and off highways
• Fuel savings could be outweighed by between a series of trucks in platoon
• May need to pay a premium for the driver negative impacts on drivers and driver formation.
of the trailing truck in a platoon due to health.
high stress from close following distance. • Long-term employment security and
• Big Brother and constant monitoring. potential job loss.
• Costs of driver education, training, and
technology maintenance. • Long-term employment security and
potential job loss.
• Ability for drivers to safely operate
with limited situational awareness and • System security and reliability; drivers
restricted views due to platooning at close must believe the system is safe and
distances. appropriate.
• System security and reliability.
• Truck platooning with other companies
could increase liability and insurance
pressure.
WORKING PAPER 2018-06 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION 11AUTOMATION IN THE LONG HAUL: CHALLENGES AND OPPORTUNITIES OF AUTONOMOUS HEAVY-DUTY TRUCKING IN THE U.S.
autonomous system handles multiple their incident mitigation potential. The levels of vehicle autonomy (Level
elements of driving execution and agency found that no rear-end colli- 4+), safety risks posed by hacking
monitors the driving environment. In sions occurred in more than 3 million and remote hijacking could become
theory, this would enable drivers to miles of data (NHTSA, 2016a, 2016b). a s i g n i f i c a n t i ss u e a n d d e s e r ve
disengage, opening up the possibility f u r t h e r st u d y ( U. S . G ove r n m e n t
Slightly more advanced technolo-
of handling alternative tasks, resting, Accountability Office, 2016).
gies are expected to further improve
or not being present in the vehicle at
on-road safety by handling a greater Greater fuel efficiency and reduced
all. In the following subsections, we
number of driving tasks and miti- emissions. In North America and
describe and discuss each of the poten-
gating human error. For example, Europe, fuel consumption typically
tial benefits of autonomous trucking in
Peloton’s platooning system (Level 1) accounts for 25% to 40% of total costs
more detail while highlighting relevant
in theory is expected to prevent col- in long-haul trucking (Sharpe, 2017).
risks, limitations, and drawbacks.
lisions through more reliable, precise, Freight trucks contribute a dispropor-
I m p r ove d o n - r o a d s a f e t y. T h e and instantaneous braking (Peloton, tionately large share of overall fuel con-
potential for autonomous vehicles 2016). Yet experts express the need sumption and environmental pollution
to improve on-road safety is attrac- for additional testing to validate the from on-road vehicles. Improvements
tive to industry groups and govern- safety potential both of the system in long-haul fuel economy directly lead
ment stakeholders and is one of the itself and more holistically across all to economic benefits and emission
most frequently cited benefits of the highway transportation under all road reductions. In the autonomous trucking
technology. The elimination of driver conditions and environments (National space, the most frequently discussed
error could save the trucking industry Academies of Sciences, Engineering, fuel savings technology application is
billions of dollars each year from colli- and Medicine, 2017). platooning. According to Peloton, a
sion avoidance as the system increas- technology company developing Level
No real-world data are available on
ingly handles driving tasks (Short & 1 platooning systems, the fuel savings
the safety impacts of vehicles with
Murray, 2016). Although there is some from truck platooning is significant,
higher levels of automation (Levels
early evidence of safety improvements estimated at $3,000 to $11,000 per
3+). Improved safety frequently is
from the technology, government reg- truck annually, ranging from $0.02 to
cited as a key benefit to vehicle
ulators currently lack the data needed $0.042 per mile per truck (Peloton,
automation, and more information is
to validate safety impacts (National 2016). A review of several truck pla-
needed to validate this claim. Sivak
Academies of Sciences, Engineering, tooning demonstrations finds that
and Schoettle (2015) argue that the
and Medicine, 2017). platooning can improve average fuel
assumption that autonomous vehicle
Several collision mitigation systems technologies will result in zero fatali- savings by 4% to 15% (see Table 5).
are available today and are consid- ties is unrealistic. Furthermore, some A series of industry interviews con-
ered Level 0 or 1. These systems typi- experts argue that Level 3, which ducted by Auburn University (2017)
cally alert drivers of potential safety re q u i re s h u m a n i n te r ve n t i o n , i s show that the benefits and drawbacks of
risks through blind spot detection and unsafe and could even increase traffic platooning are likely to differ based on
forward collision warnings, or actively collisions (Naughton, 2017). Some company operations and fleet size. For
support drivers with automatic emer- industry groups find that humans are example, although large fleets realize
gency braking systems. Forward col- too quick to fully trust Level 3 tech- substantial savings from economies
lision warning systems are estimated nology, and that they are less likely of scale, upfront costs are very impor-
to reduce rear-end collisions by 10% to successfully reengage with the tant to smaller fleets, and they may
(National Academies of Sciences, vehicle. As a result, some compa- require benefits (e.g., safety) beyond
Engineering, and Medicine, 2017). nies plan to skip Level 3 automation fuel savings to invest in the technol-
Meritor Wabco reports that its colli- (Naughton, 2017; Volvo, 2017; Google, ogy. Freight logistics could be another
sion mitigation system has reduced 2015). This discussion has been con- factor, and some companies reported
accidents by 87% and accident costs centrated mostly in the passenger that their trip distances are not long
by 89%, paying for itself in approxi- vehicle sector, but the implications enough for platooning, and that their
mately 12 months (Meritor Wabco, extend to heavy trucks. Further study trucks travel alone. Inducing additional
2014, 2017). The National Highway is needed to identify and assess the miles traveled by going out of the way
Traffic Safety Administration (NHTSA) safety potential of connected and to create a truck platoon could under-
conducted a field study of heavy-vehi- autonomous vehicle technologies at mine the fuel savings and environmen-
cle crash avoidance systems to assess all levels of automation. With higher tal benefits. Key considerations and
12 INTERNATIONAL COUNCIL ON CLEAN TRANSPORTATION WORKING PAPER 2018-06You can also read
