International Symposium on Bicycle Urbanism, Seattle, WA, June 19-21, 2013

Self-driving Tricycles
Tyler C. Folsom
Quest Integrated, Inc., Kent, WA
[email protected]
Topic: Advances in icycle Tec!nology
Key"ords: Intelligent Transportation #ystems $IT#%, &o'otics, Automated &oad (e!icles.
Abstract: Bicyclin can be promote! by proper in"rastructure !esin# $o%e&er, muc' o" t'e urban
!e&elope! %orl! is bur!ene! by !eca!es o" automobile-base! !esin# ('anin t'e in"rastructure is a
ma)or un!erta*in# A more realistic approac' may be to c'ane t'e bicycle#
+'ere are practical reasons t'at in'ibit bicyclin, suc' as inclement %eat'er, 'ills, an! clot'in t'at is
inappropriate "or e,ercise# +'ese can be mitiate! by $-. !esins t'at protect aainst rain an!
pro&i!e an electric 'elper motor "or 'ills# Status is a more important "actor/ people "eel more
important %'en t'ey replace t'eir bi*e %it' a car# We t'ere"ore propose creatin a 'i' tec'noloy
bicycle t'at can be &ie%e! as a step up "rom a car#
Sel"-!ri&in cars are e&ol&in rapi!ly# 0oole 'as !ri&en its autonomous cars more t'an 100,000 *m
in (ali"ornia tra""ic# +'e 2erce!es S-class may be capable o" !ri&in itsel" to!ay, but la%s stan! in t'e
%ay# A lare connecte! &e'icles !emonstration is in proress in Ann Arbor, 2ic'ian# Sop'isticate!
Intellient +ransportation Systems 3I+S4 are in place in Japan# 5urope 'as !emonstrate! t'e SA6+65
pro)ect, %'ere a semi-autonomous roa! train !ro&e 190 *m on Spanis' "ree%ays#
7ull sel"-!ri&in e8uipment currently costs t'ousan!s o" !ollars, an! may seem inappropriate "or a
bicycle# +'ere are t%o met'o!s t'at can interate bicycles into urban I+S systems o" t'e 21
st
century#
14 A simple !e&ice suc' as a smart p'one can broa!cast t'e cyclist9s position an! &elocity to ot'er
&e'icles# Alternati&ely, automoti&e ra!ar systems are capable o" !etectin bicycles an!
pe!estrians#
24 +'e cycle itsel" can become sel"-!ri&in an! sa"ely blen! %it' motor &e'icles#
+'is paper presents !esins "or t'e secon! option#
I" roa! &e'icle automation can ma*e tra""ic acci!ents rare, a bicycle is almost as sa"e as an SU. an!
can mi, %it' bi &e'icles on city streets# We 'a&e built a sel"-!ri&in recumbent electric tricycle, %it' a
total materials cost un!er :1000 "or &e'icle, motor, batteries, sensors, actuators, an! electronics# +'is
&e'icle is purely electric, %it' a c'oice o" manual !ri&e-by-%ire "rom a )oystic* or "ull autonomous
operation "rom microprocessors# 7uture !esins coul! be 'ybri! electric ; 'uman po%er# A 'ybri!
'uman po%ere! &e'icle 3$$-.4 coul! "orm t'e bac*bone o" an urban transportation system an! o""er
"aster an! more con&enient ser&ice t'an a car, bus or train# Urban transportation coul! be base! on
electric &e'icles ettin 0#21 <;100 *m 31000 mp4 e8ui&alent, an! run entirely on rene%able enery# A
10 * lit'ium battery %oul! pro&i!e su""icient po%er# 6ane an,iety %oul! !isappear, since a li't
battery can be s%appe! in less time t'an it ta*es to "ill a as tan*# +'e ban* o" !isc'are! batteries
pro&i!es a bu""er "or storin %in! an! solar enery#
e stand at a !istoric cusp, "!ere t!e transportation systems serving our cities may 'e on t!e
verge o) c!anging drastically. In t!e *#, it is possi'le t!at 'icycle mode s!are "ill increase
signi)icantly 'eyond t!e current +,. ut even i) ot!er cities can reac! t!e -., s!are o) Copen!agen,
/2, o) people "ill not 'e cycling. T!is paper is directed to non0'icycle users, and attempts to use
tec!nology to ma1e t!eir 2ourney as energy e))icient and non0polluting as possi'le.
W
(arious )actors in!i'iting 'icycle use !ave 'een identi)ied among people "!o 'icycle to "or1
)re3uently or occasionally 4+5. T!ese include:
Weat!er: &ain, temperature and "ind
6istance and t!e need to "or1 at multiple locations
7eed to "ear 'usiness attire and desire to avoid perspiration
8leasant "eat!er !as a positive e))ect on 'icycle usage. Women cyclists are o)ten concerned a'out 425
#a)ety
eing a'le to carry daily items
7eed to )i9 !air on arrival
T!e !ealt! advantages o) cycling out"eig! t!e ris1s 4-5. :o"ever, one mig!t con2ecture t!at people
"!o never commute 'y 'icycle perceive cycling as a ris1y activity. (arious programs !ave 'een
underta1en to encourage cycling and reduce tra))ic congestion. T!ere are at least +;; 'i1e0s!aring
programs operating in +2< cities "orld"ide 4=5.
Adverse cycling )actors can 'e mitigated 'y :uman 8o"ered (e!icle $:8(% designs t!at protect
against rain and provide an electric !elper motor )or !ills. #tatus is a more important )actor> people
)eel more important "!en t!ey replace t!eir 'i1e "it! a car. T!is is particularly evident in C!ina, "!ere
an increasingly prosperous middle class !as replaced 'icycles "it! cars 4<5. As t!is trend accelerates in
C!ina, India and ot!er developing countries, t!e unsustaina'le pressure on glo'al resources intensi)ies.
We t!ere)ore propose creating a !ig! tec!nology 'icycle t!at can 'e vie"ed as a step up )rom a car.
&educing one?s car'on )ootprint is o)ten presented as a sacri)ice, settling )or somet!ing t!at is less )un
and less convenient. @n t!e ot!er !and, e))icient tec!nologies suc! as electric or !y'rid cars are o)ten
depicted as too e9pensive )or t!e average person to a))ord. @ur goal is to devise a mode o) ur'an
transportation t!at is more convenient, )aster, sa)er, and less e9pensive t!an any alternative. T!is mode
"ill also 'e t!e most environmentally 'enign> people "ill c!oose to !ave a lo" car'on )ootprint )or
non0altruistic reasons.
Korea !as t!e "orst tra))ic accident rate o) any developed country. T!ey !ave set t!e goal o) 'ringing
tra))ic )atalities to Aero 'y 2;-;, "!ile dou'ling !ig!"ay capacity and !alving green!ouse gas
emissions 4/5. &oad ve!icle automation can ac!ieve t!ese outcomes. @nce suc! a system is in place,
ve!icle siAe and energy consumption can )urt!er s!rin1 to t!e point t!at automated ve!icles 'ear a
strong resem'lance to !uman po"ered ve!icles. Cars "ould see 'i1es even "!en t!e drivers don?t, and
t!e car "ould not let t!e driver !it t!e 'i1e. T!us 'icycles "ill 'e sa)e in mi9ed tra))ic, even "it!out
in)rastructure modi)ications.
T!is argument does not detract )rom t!e desira'ility o) planning 'icycle )riendly in)rastructure.
2
:o"ever, automation does not )igure in many t!irty0year transportation plans, and its advent may 'e as
disruptive as t!e Internet. Fe" people could )oresee t!e rami)ications o) t!e Internet t!irty years ago.
T!ere are a num'er o) )actors driving a transportation revolution 4B5:
!ig! price o) oil
concern )or climate c!ange
concern )or ur'an pollution
concern )or sustaina'ility
avoidance o) international con)lict over energy resources.
improved electric ve!icles
autonomous road ve!icles
congestion
T!e automo'ile does 'est on t!e open road, and t!ere is a perception t!at it is "ell matc!ed to t!e *#
'ecause o) t!e "ide open spaces. :o"ever, /<, o) *# (e!icle Ciles Travelled $(CT% are ur'an, and
most trips are s!ort> .., o) Americans drive alone to "or1, and )e" regularly use t!eir cargo0carrying
capacity 4., D5. T!us it is appropriate to concentrate on ur'an travel, "!ere tra))ic o)ten stops or slo"s.
Typical ur'an car speeds are given in Ta'le +. It is seldom possi'le to ac!ieve an average speed o) <;
1mE! $-; miE!% or greater on an ur'an trip. T!e data )rom Fa1ima s!o" t!at even in a small to"n,
travel speed is less t!an posted speed during pea1 periods 4+;5. Cost o) t!e roads in t!is study "ere
posted )or eit!er -< or -; mp!, "it! slo"er measured speeds on t!e -; mp! sections.
Ta'le +. Typical ur'an car average speeds
Location Mi/h km/h Notes
*.#. average commute 2. =< Average trip lengt! is +2 mi.
*.#. G8A city )uel rating +D -+ City driving cycle assumes =-, stopped or decelerating 4++5
Hapan city )uel rating +< 2= <2, stopped or decelerating 4++5
Cum'ai, India <0+D .0-; Cinimum and ma9imum average speed 4+25.
Fa1ima, WA, *#A 2-../ -..=. Average o) == segments posted )or -< mp! during 8C pea1.
Transit doesn?t do 'etter. Typical lig!t rail speeds are given in Ta'le 2. W!en one considers "ait times
and t!e last 1m )rom t!e station to t!e true origin or destination, t!e per)ormance is "orse t!an
ta'ulated. us lines are o)ten indirect and in)re3uent.
Ta'le 2: Typical ur'an lig!t rail average speeds
Location Mi/h km/h Notes
#eattle, *#A 22 -< 6o"nto"n to airport is +<.< mi, sc!eduled in -. minutes
plus < minute average "ait time 4+-5.
(ancouver, Canada 2. =< 6oes not include "ait times
To1yo, Hapan +/ 2/ Carunouc!i line )rom Kasumigase1i to I1e'u1uro travels
+;.. 1m in 2< minutes
7e" For1 City, *#A +B.= 2..+ Average su'"ay sc!eduled speed 4+=5
-
C!icago, *#A 2D.2 =B.+ Average sc!eduled speed 4+=5
Was!ington, 6C, *#A 2D.< =B./ Average sc!eduled speed 4+=5
City 'icycle speeds can 'e competitive "it! ot!er modes> ot!er papers at t!is symposium deal "it!
!o" to ma1e t!e 'icycle a more )re3uent c!oice. T!e 3uestion addressed in t!is paper is: !o" to o))er
improved transportation options to t!e non0cycling pu'lic t!at are environmentally 'enign.
A good place to start is 8ersonal &apid Transit $8&T%, an idea t!at !as 'een e9plored since t!e +D/;s.
8&T is c!aracteriAed in 4+<5 as
+. Fully automated ve!icles capa'le o) operation "it!out !uman drivers.
2. (e!icles captive to a reserved guide"ay.
-. #mall ve!icles availa'le )or e9clusive use 'y an individual or a small group, typically + to /
passengers, traveling toget!er 'y c!oice and availa'le 2= !ours a day.
=. #mall guide"ays t!at can 'e located a'ove ground, at ground level or underground.
<. (e!icles a'le to use all guide"ays and stations on a )ully coupled 8&T net"or1.
/. 6irect origin to destination service, "it!out a necessity to trans)er or stop at intervening
stations.
B. #ervice availa'le on demand rat!er t!an on )i9ed sc!edules.
W!ile t!ere are currently +-; driverless train systems
in operation "orld"ide, t!ere are only t!ree operating
8&T systems 4+/5. T!e t!ree 8&T systems include t!e
rail system t!at !as 'een operating in Corganto"n,
W( since +DB<, t!oug! it is 'ased on larger ve!icles.
T!e ot!er t"o systems !ave 'een recently
constructed and are 'ased on ve!icles running on
dedicated, paved guide"ays. T!ey are located at
Iondon?s :eat!ro" Airport and t!e environmental
city o) Casdar, *AG, s!o"n in Figure +. T!e tire
mar1s evident in t!e picture result )rom precisely
)ollo"ing t!e same pat!. 8&T ve!icles "ere o)ten envisioned as traveling on lig!t elevated rail
guide"ays. T!eir proponents argued t!at t!ese lig!t0"eig!t guide"ays "ould 'e ine9pensive to
construct, 'ut t!e cost o) constructing a city0spanning net"or1 !as "or1ed against t!e idea. T!e )ailure
o) 8&T to "in "idespread acceptance can 'e attri'uted to several )actors:
An up0)ront investment is needed to 'uild an entirely ne" in)rastructure.
Captive ve!icles are plagued "it! t!e last and )irst 1m pro'lem o) system access.
Tra))ic planners are reluctant to commit to a system t!at !as not already 'een demonstrated
else"!ere at t!e same scale.
#upporting tec!nology "as not readily availa'le in t!e 2;
t!
century.
#ome )uturists tal1 o) +;; miE! cars> suc! a goal is not desira'le and "ould only e9acer'ate spra"l.
#ustaina'le transportation re3uires t!at "e loo1 at present transportation demands, and 'uild an
improved system serving t!ose needs. Coving people )rom true origin to destination at an !onest <;
1mE! $-; miE!% average speed in t!e city is a ma2or improvement, and it is ac!ieva'le t!roug!
automation. Automation presents t"o parado9es:
+. Trip times can 'e su'stantially s!orter despite lo"er ma9imum speeds.
=

Figure +. 8&T system at Casdar.
2. Fe"er !ig!"ay lanes can carry more tra))ic.
T!e )irst e))ect is t!e result o) going )rom origin to destination in a reserved lane "it! no stops. T!e
second e))ect comes )rom increased !ig!"ay capacity "!en automated ve!icles are operated in
platoons.
T!e energy needed to move a ve!icle is t!e sum o) energy to overcome rolling resistance and energy to
overcome aerodynamic drag. Io"er ma9imum speeds lead to more e))icient energy usage in t"o "ays.
6rag energy is proportional to t!e cu'e o) speed, so dou'ling t!e speed re3uires eig!t times more
energy. &olling energy is directly dependent on speed, so "!en speed dou'les, t!is part o) energy only
dou'les> !o"ever )re3uent accelerations re3uire more energy. T!us t!e 8&T paradigm o) no stops
'et"een origin and destination uses energy e))iciently.
Gnergy to accelerate or to overcome rolling resistance is proportional to t!e mass o) ve!icle and
passengers. A +/.2 1g $-B;; l'% ve!icle carrying an ./ 1g $+D; l'% man !as +; times as muc! rolling
resistance as a D+ 1g $2;; l'% ve!icle carrying t!e same rider.
As ve!icle mass goes do"n, t!e relative importance o) overcoming "ind resistance increases. For cars,
energy to overcome rolling resistance is greater until << 1mE! $-< miE!%> at !ig!er speeds aerodynamics
dominates. For 'icycles, t!e cross0over point comes at 2; 1mE! $+2 miE!% 4+B5. :ig!ly e))icient lig!t
ve!icles must !ave small )rontal area and 'e care)ully streamlined. W!en streamlined ultra0lig!t
ve!icles are operated at reasona'le speeds, and avoid unneeded accelerations, )uel e))iciencies o) ;.2<
IE+;; 1m $+;;; miEgal% e3uivalent are possi'le. 4+.5.
Gnergy consumption at <; 1mE! $-; miE!% is s!o"n )or various ve!icles in Figure 2 4+B5. A single
occupancy automo'ile getting /./ I E +;; 1m $-. miEgal% re3uires -+ 1W $=2 !p%. W!en t!e ve!icle is
loaded "it! )ive people, e))iciency improves to B 1WEperson, rivaling a diesel commuter train?s /.<
1WEperson. Anyone capa'le o) moving a 'icycle at <; 1mE! is e9pending +.B 1W. Aerodynamics
ma1es an :8( even more e))icient, re3uiring only ;./= 1W $;../ !p%. T!is true regardless o) t!e po"er
source, 'e it !uman, gasoline, or electric. In t!e *#, a tricycle "it! a ;.B< 1W electric motor is legally a
'icycle.
G9perimental ve!icles modeled on t!e :8( !ave ac!ieved remar1a'le per)ormances on t!e trac1,
t!oug! t!ese are not practical ve!icles. #tudents at t!e *niversity o) ritis! Colom'ia 'uilt t!e ve!icle
s!o"n in Figure - )or t!e #ociety o) Automotive Gngineers $#AG% supermilage competition> it ac!ieved
<

Figure 2. Gnergy Consumption

Figure -. *C ve!icle at #AG competition
;.;B< IE+;; 1m $-+=< miEgal% at lo" speed 4+D5. An electric ve!icle travelled +;; 1m $/2 miles% in
one !our, e3uivalent to ;.++ IE+;; 1m $22;; miEgal% 42;5.
It !as 'een pointed out t!at t!e posted speed limit may !ave little to do "it! actual speeds 42+5. A side
e))ect o) ve!icle automation is t!at speed limits 'ecome real. T!e speeds at "!ic! automated ve!icles
are operated must 'e set to con)orm "it! driver e9pectations o) t!e proper speed. T!ere is a danger t!at
legalism and )ear o) lia'ility "ill lead to unrealistically lo" speeds, "!ic! "ill drive a"ay riders. T!e
system?s operating speed is set 'y t!e government. T!e speed s!ould 'e )ast enoug! to attract riders,
'ut slo" enoug! to optimiAe energy use.
Today, cars t!at drive t!emselves or o))er e9tensive 6river Assistance #ystems $6A#% are pic1ing up
"!ere 8&T le)t o)). Automated 'us doc1ing systems !ave 'een demonstrated. 6A# may provide
automatic par1ing, Adaptive Cruise Control $ACC%, lane departure "arnings, or collision avoidance
systems. An ACC system detects t!e distance and speed o) t!e closest in0lane ve!icle a!ead and
modi)ies )ollo"ing speed. etter tra))ic t!roug!put is ac!ieved 'y Cooperative Adaptive Cruise
Control, in "!ic! a platoon o) ve!icles s!ares in)ormation 4225.
Joogle !as driven its autonomous cars more t!an <;;,;;; 1m in Cali)ornia tra))ic and e9pects to !ave
its system on t!e mar1et "it!in )ive years 42-5. T!e Cercedes #0class may 'e capa'le o) driving itsel)
today, 'ut la"s stand in t!e "ay 42=5. Cost o) t!e "orld?s motor ve!icle codes are 'ased on t!e (ienna
Convention on &oad Tra))ic, a +D/. *7 treaty. It states KGvery moving ve!icle or com'ination o)
ve!icles s!all !ave a driverL.
A large connected ve!icles demonstration is in progress in Ann Ar'or, Cic!igan 42<5. T!e one0year
#a)ety 8ilot involves 2,.-/ communicating ve!icles, and is sc!eduled to conclude on August +., 2;+-.
#op!isticated Intelligent Transportation #ystems $IT#% are in place in Hapan 42/5, "!ic! is evaluating
"!et!er to implement !ig!"ay road trains 'y t!e early 2;2;s. Gurope !as demonstrated t!e #a)e &oad
Trains )or t!e Gnvironment $#A&T&G% pro2ect, "!ere a semi0autonomous road train drove +D; 1m on
#panis! )ree"ays 42B5. In a road train, t!e lead ve!icle is pro)essionally driven, and t!e )ollo"ing
ve!icles are under automated control. (olvo "as part o) t!e consortium in t!e #A&T&G pro2ect, and
!as set t!e goal t!at none o) its ve!icles "ill 'e involved in cras!es.
An isolated autonomous ve!icle "ill improve sa)ety, 'ut it "ill do little to relieve congestion. T!e
system0"ide improvements occur "!en automated ve!icles communicate "it! eac! ot!er, and operate
in platoons. 6riving in closely0spaced platoons can increase !ig!"ay capacity 'y a )actor o) t!ree to
eig!t 42.5. It )ollo"s t!at one or more )ree"ay lanes s!ould 'e reserved )or automated cars and 'uses>
ot!er"ise manually driven ve!icles "ould insert t!emselves and snarl t!e lane. An :@( lane could 'e
repurposed )or automatic operation, minimiAing t!e in)rastructure costs. Gven "!en t!e automated lane
operates at a )raction o) its capacity, it can divert enoug! ve!icles )rom t!e manually driven lanes to
increase t!e manual )ree"ay capacity, despite t!e lost lane.
T!us road ve!icle automation "or1s 'est "!en t!e automated ve!icles are in separate lanes. T!is in
turn simpli)ies t!e tec!nology )or automation, removing t!e need )or e9pensive radar or lidar systems
to detect un)oreseen events. I) t!e ve!icle al"ays operates on )i9ed routes, t!ese can 'e digitally
mapped accurately, and t!e odometer and compass are su))icient to determine position. An occasional
J8# )i9 or landmar1 can reset odometer dri)t. Io" cost laser or sonar range)inders can determine t!e
distances to neig!'oring ve!icles. Ad2acent ve!icles can 'e in "ireless communication. Gvery tent! o)
a second, t!ey can s!are current positions, velocities, and planned maneuvers> most une9pected
/
'e!avior goes a"ay.
An automated road ve!icle in a reserved lane satis)ies most o) t!e c!aracteristics o) 8&T. T!e only
e9ception is t!at t!e ve!icle need not 'e captive to t!e guide"ay. It may e9it t!e automated system, and
turn control over to t!e driver. #uc! a ve!icle is called Kdual modeL. It eliminates t!e )irst and last mile
pro'lems o) pu'lic transport, 'ut can 'ene)it )rom automation )or t!e middle portion o) t!e ride.
T!ere are tec!nical issues on sa)ely transitioning 'et"een manual and automated modes, and in
certi)ying t!at a ve!icle is )ully capa'le o) automated operation, and !as not 'een tampered "it!. #uc!
security can 'e designed into !ard"are systems. A dual mode ve!icle attempting to enter an automated
system "ould 'e electronically 3ueried )or credentials, esta'lis!ed at t!e ve!icle?s last inspection.
T!ese credentials can include en)orcea'le limits on ve!icle dimensions, "eig!t, emissions and )uel
e))iciency.
6river assistance systems t!at rely on operator intervention during a crisis may )ail i) a driver is
unresponsive> a)ter all, one o) t!e attractions o) sel)0drive is 'eing a'le to do ot!er t!ings, including
sleeping. It is pre)era'le to depend entirely on automation, and pro!i'it direct driver intervention.
@t!er"ise, a nervous driver mig!t panic at t!e tig!t spacing, slam on t!e 'ra1es, and collide. T!e
corollary is t!at a dual mode ve!icle must 'e drive0'y0"ire, "it! no mec!anical controls.
W!en t!e computer is operating all ve!icles in a lane, t!ere s!ould 'e no collisions. Cotorcycles "ould
'e almost as sa)e as truc1s. icycles could 2oin t!e stream i) t!ey can 1eep up and ta1e orders )rom t!e
tra))ic control computer. :uman po"ered ve!icles !ave reac!ed +-; 1mE! $.2 miE!% 42D5> "it! a
suita'le aerodynamic )airing and a t!ird "!eel )or 'alance, a 'icycle can run "it! t!e 'ig 'oys. In order
to strictly )ollo" instructions )rom t!e computer, suc! a cycle "ould need to !ave automated
propulsion, steering, and 'ra1e systems. It is possi'le to 'uild a !y'rid !uman E electric po"ered
ve!icle in "!ic! t!e rider pedals to c!arge one 'attery, "!ile a separate 'attery po"ers t!e ve!icle.
Glectronics can 'e 'uilt to s"ap t!e traction 'attery and c!arging 'attery at t!e rig!t time.
T!ere is a critical opportunity "it!in t!e ne9t several years to design automated lanes and ve!icles. I)
t!is opportunity is missed, automated ve!icles "ill 2ust 'e improved automo'iles> t!ey can 'e muc!
more. A smoot!ly operating system re3uires not 2ust t!e travel lane, 'ut a lane )or e9iting )rom t!e
middle o) platoons or merging onto t!e end o) one. &ig!t o) "ay scarcity ma1es it desira'le t!at an
automated lane 'e !al) t!e "idt! o) a standard lane, t!us allo"ing conversion o) an e9isting lane into
'ot! an automated travel lane and a transition lane. T!is in turn dictates t!at automated ve!icles s!ould
not 'e muc! "ider t!an one meter> "ider ve!icles may 'e possi'le, since computer drivers tolerate
narro" lanes 'etter t!an !uman drivers.
T!us an automated ve!icle mig!t loo1 li1e a :8( po"ered 'y a small electric motor. It mig!t carry
t"o passengers, eit!er seated in tandem or side0'y0side. T!ese ve!icles could 'e electronically or
mec!anically coupled to )orm a larger ve!icle traveling as a unit. #uc! a modular ve!icle adapts to t!e
siAe o) t!e group and t!eir cargo as needed. W!en operated on city streets as a road train, t!e ve!icles
lin1 toget!er to )orm a modular 'us 4-;5.
#el)0drive can 'e )ar more t!an an improved automo'ile. Automated road tra))ic produces an entirely
ne" transportation mode o))ering
Convenience o) t!e automo'ile
8u'lic access o) t!e 'us
B
#iAe o) t!e motorcycle
Gnergy e))iciency o) t!e 'icycle
#a)ety o) t!e train
7o congestion
A !ig!ly aerodynamic lig!t ve!icle traveling at modest speeds re3uires little energy. An electric ur'an
ve!icle could 'e po"ered 'y a +; 1g lit!ium 'attery 4-+5. Iig!t 'atteries ena'le re)ueling 'y 'attery
s"ap. T!e s"apped out 'atteries can 'e rec!arged at any time o) t!e day or nig!t, particularly t!ose
times "!en "ind or solar po"er is availa'le. T!us ur'an transportation )rom rene"a'le, non0polluting
po"er is )easi'le. #ince t!e ve!icles are automated, t!e riders need not 'e concerned "it! 'attery
c!arge> t!e ve!icle "ill stop at a re)ueling station "!en needed. Cost trips to t!e re)ueling station "ill
'e done 'y riderless ve!icles.
It is not enoug! to invent a superior transportation mode> "e must consider implementation. Any
system re3uiring an entirely ne" in)rastructure "ill 'e still'orn. :o" does t!e )irst sel)0drive ve!icle
get onto t!e roadM 7evada !as 'een touted as t!e )irst state to legaliAe autonomous ve!icles. In )act,
"!at t!ey !ave done is to re3uire t!at an e9perimental autonomous ve!icle !ave t"o people on 'oard
monitoring it> don?t e9pect to get driverless piAAa delivery in 7evada. :o" muc! testing "ill 'e
re3uired to convince aut!orities t!at automated ve!icles are sa)eM Is t!ere any test program t!at can
anticipate all possi'ilitiesM Gven i) t!ere is, i) t!e ve!icle !as evolved over time, !o" can "e 'e certain
t!at t!e current version is still sa)e on all testsM :o" can politicians 'e motivated to c!ange motor
ve!icle codesM W!o "ill 'e lia'le i) somet!ing goes "rongM
As mentioned a'ove, t!e )irst automated ve!icles on t!e !ig!"ay are li1ely to )orm a road train,
)ollo"ing a ve!icle "it! a driver. T!is tec!nology could 'e 'roug!t to city streets to )orm a modular
'us 4-;5. T!e 'us "ould consist o) small pods. T!e main 'us "ould not stop at 'us stations> pods
"ould accelerate as t!e 'us approac!ed, and 2oin onto its rear. Any pod could e9it t!e train )rom t!e
middle, "it! t!e ot!er pods )orming up to )ill t!e gap. T!e pu'lic pods "ould 'elong to t!e transit
agency and initially only operate "it! t!e modular 'us.
@nce a city 'egins to operate modular 'uses, it is a small step to route t!em on dedicated 'us"ays or
:@( lanes. I) t!ere are no ot!er ve!icles in t!ese lanes, t!ere may 'e no need )or t!e !uman0driven
lead ve!icle. T!e system "ould evolve into a 8&T0li1e system operating on converted lanes. T!e
paradigm is similar to lig!t rail in t!at t!e city provides 'ot! t!e ve!icles and t!e lanes.
#uc! a pod0'ased transit system mig!t add elements o) ve!icle0s!are. T!e city could purc!ase dual
mode pods, and allo" people to drive t!en a)ter leaving t!e automated system. @nce suc! a system !as
proven itsel), it could admit privately0o"ned dual mode ve!icles.
T!e Glcano 8ro2ect !as 'uilt a prototype )or an ultra0lig!t dual mode autonomous ve!icle, s!o"n in
Figures = and < 4-2, --5. It is 'uilt on a standard tadpole recum'ent tricycle, modi)ied to 'e an electric
'i1e. Total cost o) components is under N<,;;;, including t!e mec!anical c!assis, electronics, sensors,
and actuators. 6erailleurs, sproc1ets, and c!ain are replaced 'y a !u' motor po"ered 'y lit!ium
'atteries. A standard e0'i1e controller responds to t!e t!rottle. T!e operator uses a 2oystic1 to control
t!e ve!icle. A !eavy duty servo steers t!e )ront "!eels. A second servo pus!es t!e dual ca'le 'ra1e
lever to activate dis1 'ra1es on t!e )ront "!eels. A similar system could 'e used to control an
automo'ile> t!e tricycle steering is Ac1erman, and t!e servos provide enoug! t!rust and ro'ustness. It
is possi'le to use regenerative 'ra1ing )rom t!e electric motor to c!arge t!e 'attery.
.
Figure =. Glcano unloaded> seat removed Figure <. Glcano "it! rider
T!e mac!ine "as originally 'uilt to compete in t!e #eattle &o'otics #ociety?s &o'o0Cagellan contest
)or outdoor autonomous ro'ots 4-=5. T!e contest re3uires eac! ro'ot to navigate to latitude and
longitude points t!at are revealed only at t!e start o) t!e race. T!e pro2ect is named a)ter Huan #e'astian
Glcano, t!e )irst Guropean to sail around t!e "orld. Ferdinand Cagellan never returned )rom !is
2ourney> !e "as 1illed in t!e 8!ilippines. Glcano managed to 'ring +. survivors 'ac1 to #pain. T!e
open source code and electronics o) t!e Glcano 8ro2ect is designed to ma1e completion o) &o'o0
Cagellan easy. T!e )irst prototype ve!icle satis)ies t!e <; pound $2- 1g% "eig!t limit and )our )oot
$+.2 m% dimensional limits o) t!e contest "!en t!e 'oom is removed.
A micro0computer controls t!e prototype ve!icle. An operator uses a 2oystic1: up to accelerate, do"n to
'ra1e, le)t or rig!t to turn. T!e computer interprets t!ese motions, and sends appropriate signals to
actuator controllers. T!e ne9t p!ase is to replace t!e operator "it! a set o) small computers. Custom
electronics !ave 'een designed and )a'ricated to ma1e t!e systems ro'ust. A second prototype ve!icle
is 'eing con)igured 'y t!e Computer #o)t"are and #ystems 6epartment o) t!e *niversity o)
Was!ington, ot!ell. 6r. Folsom "ill direct !is Gm'edded #ystems class to complete t!e so)t"are )or
autonomous operation starting in Fall, 2;+-. T!e Glcano 8ro2ect is cooperating "it! ot!er groups
interested in ma1ing sel)0driving ve!icles. A suita'le simulator is 'eing developed at &utgers
*niversity. T!e Glcano design is also compati'le "it! t!e *#A&sim simulator, "!ic! is designed )or a
"ide variety o) ro'ots, including automated road ve!icles. *se o) a simulator ena'les researc! 'y
groups not !aving access to a p!ysical ve!icle.
Automotive radar and lidar systems cost t!ousands o) dollars> t!e Glcano sel)0driving cycle omits t!em.
In a group o) ve!icles, a less "ell0e3uipped ve!icle can increase its 1no"ledge o) its location and
surroundings 'y communicating "it! ve!icles "it! 'etter sensors 4-<5. T!e present design is 'ased on
an ine9pensive sensor suite:
Jlo'al 8ositioning #ystem $J8#% receiver.
Inertial 7avigation *nit $I7*% "!ic! provides a digital compass and tilt sensor.
Cyclometer, measuring odometry )rom t!e "!eel.
Codi)ied optical mouse, "!ic! gives t"o0dimensional optical odometry.
#onar range )inders, "!ic! can detect o'stacles.
D
6igital camera to )ind lane mar1ings, ve!icles or landmar1s.
6igital map o) intended route.
#ystem organiAation is given in Figure /. T!ere is a 'rea1 'et"een t!e lo" level micro0controller
$designated C2% t!at spins t!e "!eels and points t!e steering, and t!e !ig!er level processors t!at )ind
route and position. T!e prototype demonstrates lo" level control> an alternate micro0controller and
actuators could !andle any ot!er ve!icle ranging )rom a radio controlled toy car to a )ull siAed car. T!e
!ig!er level )unctions are independent o) t!e ve!icle.
T!e !ig!er level navigation is per)ormed 'y t!ree
separate processors, operating continuously:
+. A localiAation processor $C/% reads
sensors and )orms t!e 'est estimate o)
current position and !eading.
2. A route )inder $C=% reads a digital map o)
t!e area, inputs t!e current location, and
selects t!e 'est pat! to t!e destination.
-. A pilot $C-% directs t!e ve!icle to )ollo"
t!e ne9t segment o) t!e route, deviating
around any o'stacles t!at it )inds.
Additional processors $CB% may 'e assigned to
!ig! in)ormation sensors, suc! as a camera.
Communications "it! ot!er ve!icles "ould need
an additional processor.
Alternately, all t!ese tas1s could 'e implemented
on a single computer or smart p!one. We !ave
instead c!osen to distri'ute t!e tas1s among
small processors to assure real0time availa'ility
"it!out depending on an operating system. T!e
!ard"are separation o) computing tas1s en)orces
modular design.
T!e production ve!icle is envisioned to add an
enclosed s!ell, providing superior aerodynamics
and protection )rom t!e "eat!er. T!e s!ell needs
to 'e designed )or easy entry and egress. It must
provide proper ventilation, all "eat!er visi'ility,
and avoid solar over!eating.
T!e goals o) t!e Glcano 8ro2ect are:
Ca1e autonomy availa'le to non0specialists.
8roduce an e9perimental ve!icle and electronics costing less t!an N<,;;; total. A )ully
enclosed road0"ort!y production ve!icle s!ould cost less t!an N+;,;;;.
Jenerate pu'lic demand )or road automation.
Gncourage !ig! )uel e))iciency $+;;; mpg E ;.2< IE+;; 1m% t!roug! ultra0lig!t automated
ve!icles.
#et standards )or cooperative automation using a scala'le distri'uted Tra))ic Canagement
+;
Figure /. Arc!itecture
#ystem.
T!e long term vision is not a single sel)0driving ve!icle, 'ut a collection o) suc! ve!icles t!at
communicates "it! eac! ot!er. Automatic ve!icles can ta1e instructions )rom a roadside Tra))ic
Canagement Computer t!at manages a section o) road"ays. T!ese computers "ould lin1 toget!er to
)orm a distri'uted, scala'le Tra))ic Canagement #ystem> suc! a system can reduce congestion.
Conclusion
Autonomy c!anges t!e very nature o) ur'an transportation. Its potential )or sa)ety is "idely recogniAed.
Iess evident is its potential to 'rea1 transportationOs automo'ile )i9ation. W!en accidents 'ecome rare,
ve!icles o) all siAes can sa)ely mi9. (e!icles can 'ecome modular and miniature, "it! additional
carrying capacity added only "!en needed. #mall ve!icles re3uire an order o) magnitude less energy,
'rea1ing oil dependency, removing range an9iety )rom electric ve!icles, and ena'ling sustaina'ly
po"ered transportation.
References
+. :einen, G., K Caat, and van Wee $2;++%, T!e day0to0day c!oice $not% to commute 'y 'icycle,
Transportation &esearc! oard, Was!ington, 6.C.
2. T"addle, :., F. :all, and . racic $2;+;%, Iatent icycle Commuting 6emand and G))ects o) Jender
on Commuter Cycling and Accident &ates. In Transportation &esearc! &ecord: Hournal o) t!e
Transportation &esearc! oard, 7o. 2+D;, Transportation &esearc! oard o) t!e 7ational Academies,
Was!ington, 6.C. pp. 2.0-/.
-. :artog, H.H. 6e, oogaard, :., 7i2land, :., :oe1, J. $2;+;% 6o t!e :ealt! ene)its o) Cycling @ut"eig!
t!e &is1sM Gnvironment :ealt! 8erspectives. Aug>++.$.%:++;D0+/. Gpu' 2;+; Hun ++
=. #!a!een, #.A., #. JuAman and :. P!ang, $2;+;% i1es!aring in Gurope, t!e Americas and Asia, In
Transportation &esearc! &ecord: Hournal o) t!e Transportation &esearc! oard, 7o. 2+=-, pp.+<D0+/B.
<. Iin Iee, T!e Ioss o) t!e Kingdom o) icycles, icycle *r'anism #ymposium, Hune +D022, 2;+-,
#eattle, WA.
/. Foung0Hun Coon, Korean Intelligent Transportation #ystems, Transportation &esearc! oard $T&%
D2
nd
annual meeting, , Han +<, 2;+-
B. Jil'ert and 8erl, Transport &evolutions, 2;+;.
.. 6@T, Tra))ic (olume Trends, Was!ington, 2;;..
D. Federal :ig!"ay Administration $2;;D% #ummary o) Travel Trends Q 2;;D 7ational :ouse!old Travel
#urvey.
+;. City o) Fa1ima, Travel #peed #tudy o) *r'an #treets *sing J8# and JI#, 2;;2.
!ttp:EE""".ya1ima"a.govEservicesEgisE)ilesE2;+2E;<Ereportman.pd)
++. #. C. 6avis, #. W. 6iegel and &. J. oundy, Transportation Gnergy 6ata oo1: Gdition 2D. Huly 2;+;
ta'le =.-2. online cta.ornl.govEdata
+2. !ttp:EEarticles.timeso)india.indiatimes.comE2;+20;+0+.Emum'aiE-;/-.==BR+R1mp!0coastal0
road0travel0speed
+-. !ttp:EE""".soundtransit.orgE&ider0JuideEIin10lig!t0rail
+=. Catt Ho!nson, Average sc!eduled speed: :o" does Cetro compareM 2;+2.
!ttp:EEgreatergreater"as!ington.orgEpostE<+.-Eaverage0sc!edule0speed0!o"0does0metro0compareE
+<. Anderson, H.G, KFundamentals o) personal rapid transitL, Con)erence on Automated 8eople Covers, Ias
Collinas, TS, Carc! +.02+, +DD-.
+/. I. Fa'ian, A8C Juide, 2;++. @nline: !ttp:EE""".air)ront.usEapmguide2;;.EdataEapmstats.!tml
+B. 6. J. Wilson, icycling #cience, -rd ed, Cam'ridge, CA, 2;;=
+.. Tyler Folsom, Gnergy and Autonomous *r'an Iand (e!icles, IGGG Tec!nology and #ociety CagaAine,
#ummer 2;+2, pp. 2.0-.. !ttp:EE""".3i2.comEinde9.p!pEtransportation
+D. Time CagaAine, est inventions o) 2;;/ 0 Gasy rider.
2;. !ttp:EE""".electrat!onamerica.orgE7e"sR)ilesECic!igan Glectrat!on press release+.pd)
++
2+. Art!ur #la'os1y, (e!icle #peed Iimit is a Concept> #peed is t!e &eality, icycle *r'anism #ymposium,
Hune +D022, 2;+-, #eattle, WA.
22. #. G. #!ladover, Cooperative Adaptive Cruise Control to #ta'iliAe Car Follo"ing, Transportation
&esearc! oard $T&% D2
nd
annual meeting, session .;-, Han +/, 2;+-.
2-. Ant!ony Ievandos1i, Joogle?s #el)06riving Car, Transportation &esearc! oard $T&% D2
nd
annual
meeting, session -//, Han +=, 2;+-
2=. C!ristian WTst, UAuto &evolution: A 8romising Future )or #el)06riving CarsU 6er #piegel, Han 2.,
2;+-.
2<. 6avid I. Ielanc, 6ata )rom #a)ety 8ilot Codel 6eployment, T& D2
nd
annual meeting, session .;-,
Han +/, 2;+-
2/. Fasu!iro @1umura, Auto 8ilot Initiative o) Hapan, T& D2
nd
annual meeting, session .;-, Han +/, 2;+-
2B. Tom &o'inson, Gric C!an, Gri1 Coeling!, @perating 8latoons on 8u'lic Cotor"ays: An Introduction to
t!e #A&T&G 8latooning 8rogramme, +B
t!
World Congress on IT#, 2< @ct, 2;+;
2.. #. G. #!ladover, &easons )or operating A:# ve!icles in platoons, in 8. A. Ioannou Automated !ig!"ay
systems, $#pringer, 7e" For1, +DD/%.
2D. !ttp:EE""".i!pva.orgEland.!tm
-;. T.C. Folsom, KT!e Codular usL, su'mitted to Tra))ic Tec!nology International, 2;+-.
-+. W. C. Corc!in and :. @man, Glectric icycles, IGGG 8ress, 2;;/
-2. !ttps:EEgit!u'.comEelcanoE
--. """.elcanopro2ect.com
-=. !ttp:EE""".ro'ot!on.orgEro'ot!onEro'o0magellan.p!p
-<. :ao Ii and Fa"Ai 7a!as!i'i, KCooperative Culti0(e!icle IocaliAation *sing #plit Covariance
Intersection FilterL IGGG Intelligent Transportation #ystems CagaAine, (ol <, 7o. 2, #ummer 2;+-,
pp.--0==.
+2