CARPOOLNOW

Just-in-Time Carpooling without Elaborate Preplanning

Dominic W. Massaro, Benjamin Chaney, Stephanie Bigler, Jessica Lancaster, Suresh Iyer

Mrunal Gawade, Michael Eccleston, Edith Gurrola and Angelica Lopez

University of California, Santa Cruz, CA 95064, U.S.A.

Keywords: (Dynamic) Carpooling, Mobile phone, Trust, Community adhesion, Socialization, Web service, Driver

safety.

Abstract: Carpooling reduces the number of cars on the road, reduces gas consumption, and saves participants money.

In order to free carpooling from rigid schedules and preplanning, just-in-time carpooling allows a large

member base of passengers and drivers to be matched with each other automatically and instantly, allowing

for on-the-spot arrangement of rides. A mobile phone call or text message initiates an automatic process in

which drivers and passengers are matched to a shared ride wherever and whenever they need it, without the

scheduling constraints of traditional carpooling. This program faces a number of challenging barriers in

technology and behavioral science. These include the creation of a seamless interaction between mobile

phones and the internet server, voice recognition and SMS solutions, safety of mobile phone use and

driving, and motivation, safety, and trust among participating members of the carpooling community.

1 INTRODUCTION

Changes are rapidly occurring as our culture

becomes more mobile and "nomadic," able to access

information and communicate anywhere at any time.

Nomadic travel schedules, increasingly breaking

away from the 9-to-5 work schedule, have become

chaotic and greatly distributed. This may be one of

the reasons for a decline in carpooling, despite the

increasing pressure for alternative transportation

options (Ungemah et al., 2007). Our dynamic

carpooling project grew out of these observations in

a Technology Benefiting Humanity course taught by

the senior author at UCSC.

The most pervasive modern-day "nomadic

device" for our less predictable lifestyles is the cell

phone. Cell phones are small and portable, and most

people have one. In November 2007, worldwide

mobile telephone subscriptions reached 3.3 billion,

which is equivalent to over half of the global

population, and in the United States the percentage

of mobile telephone subscribers is even higher at 85

percent or 259 million subscribers (Virki, 2007).

Additionally, the most common feature used on a

cell phone other than making a phone call is the

sending or receiving of text messages. Since so

many people have cell phones and because they are

the ultimate "nomadic device", we are evaluating the

feasibility of a dynamic, just-in-time carpooling

service that allows users to utilize cell phones to

connect to potential drivers or passengers.

2 THE PROBLEM

Transportation is a major issue in our world today.

Traditional problems of transit, once thought of as a

strictly civil engineering difficulty, are increasingly

being revisited and recognized as a critical

environmental crisis. Transportation accounts for

about 29% of all greenhouse gas emissions

(Transportation and Climate, 2009). There is also a

growing awareness of the environmental impact

associated with the fabrication of cars, especially

from the nickel-metal hydride batteries in more fuel-

efficient hybrids.

The dominant form of personal transit today is

the private passenger car. Very often, these cars are

used with only a single rider. For instance, in the UK

the average car only has 1.5 people in it (Hartwig &

Buchmann, 2007). An over abundance of cars

creates many well-documented problems for urban

areas, such as increased traffic, increased pollution,

parking congestion, and the need for expensive

219

Massaro D., Chaney B., Bigler S., Lancaster J., Iyer S., Gawade M., Eccleston M., Gurrola E. and Lopez A.

CARPOOLNOW - Just-in-Time Carpooling without Elaborate Preplanning.

DOI: 10.5220/0001808902190224

In Proceedings of the Fifth International Conference on Web Information Systems and Technologies (WEBIST 2009), page

ISBN: 978-989-8111-81-4

infrastructure maintenance. In addition to these

large-scale problems, driving alone is expensive to

individuals. The rising cost of oil affected motorists

on a more personal level when gas prices reached an

all time high in July 2008 of slightly higher than

$4.00 per gallon. US households spent an average of

almost $7,500 on transportation in 2004—second in

spending behind housing. The average operating

cost (gas, maintenance, and tires) of driving a car in

2008 is 17 cents per mile (Behind the Numbers,

2009). When this expense is added to ownership

costs (full-coverage insurance, license, registration,

taxes, depreciation, and finance charge), the total

expense for driving a car 15,000 miles per year is

about $8,121 per vehicle. Given the dramatic

economic downturn in 2008, these expenses will no

longer be affordable for many US citizens.

The expenses, both environmental and fiscal, of

single occupancy vehicles can be reduced

dramatically by utilizing the empty seats in these

vehicles. Carpooling targets these empty seats: it

takes cars off the road reducing traffic and pollution,

and may provide an opportunity for social

interaction. However, traditional ways of scheduling

carpools often limit users to consistent schedules and

fixed rider groups--carpooling to the same place at

the same time with a set person or group of people.

3 DYNAMIC CARPOOLING

Dynamic carpooling overcomes some of the

traditional carpooling restrictions by allowing a

large membership base of passengers and drivers to

be matched with each other automatically in real

time, allowing for on-the-spot arrangement of rides.

Dynamic carpooling thus opens the domain of

carpooling to not only traditional employment trips

but also to irregular trips, including so called "last

mile" trips, which are also the most difficult to

service by most public transit systems.

Because dynamic carpooling relies heavily on

frequent access to the ride matching service, it is

essential that accessing the service be as easy as

possible for users. Access should be quick, simple,

and convenient. Providing this access to users could

occur via a mobile phone service. Mobile phones are

ubiquitous in today's society; they travel with most

people everywhere they go. Because of this, and the

other benefits cell phones provide such as easy

interfacing with the internet and possible location

awareness, mobile phones are an excellent platform

for access to a dynamic carpooling service.

If the primary mode of interaction is by mobile

phone, access would be available at any time and

any place. Passengers could arrange a ride wherever

and whenever they needed it, without the rigid

planning of traditional carpooling. Additionally, this

would allow drivers to designate any trip available

for carpooling at the time of the trip.

4 SYSTEM IMPLEMENTATION

A dynamic carpooling system relies on two

underlying sources of information: quick and easy

route announcement by the members (member end)

and flexible and intelligent route matching by the

system (server end). Multiple methods of route

announcement, including email, text messages, and

voice telephone calls, would be similarly handled by

the server. The message would announce the

member's name (which would be detected

automatically by his or her telephone number), an

indication of the starting location, and whether the

member is a driver, a passenger, or either depending

on the quality of the match for the trip. If it is a

voice call, automatic speech recognition would be

used to recognize these indices. Given the simplicity

of the dialog and the system’s protocol to check the

accuracy of the recognized information, this part of

the process could be essentially error free.

4.1 Related Work

Existing solutions could be broadly classified into

two types, based on their approach. First is the ride

share bulletin board (e.g., PickupPal, 2009) in which

the drivers and riders post their carpooling plans.

The users manually search the existing posts to find

a match, which is time intensive and requires

internet browsing access.

The second type uses automated route-matching

solutions to match up a rider and a driver. For

example, iCarpool (2009) uses high precision trip

matching to find the best carpool match. Although

arrived at independently, our system shares a similar

conceptual framework as Aktalita (2009), which

combines the Web, a geo spatially enabled database,

and a Java enabled cell phone to provide real-time

dynamic carpooling between drivers and passengers.

Both of these types of solutions require the

member’s access to a computer, a browser enabled

phone, or a java enabled cell phone. In contrast, we

expect that the more ad hoc and easy the carpooling

solution the more popular it would be. A simple call

or a text message from a phone would simply trigger

WEBIST 2009 - 5th International Conference on Web Information Systems and Technologies

220

a precision route-matching algorithm.

4.2 Prototype

We developed a prototype to check the feasibility of

this system.

1. A passenger submits a request for a ride or a

driver offers a ride. If a member has the flexibility to

be either a passenger or a driver, then they will also

be given the option to select both.

2. Once submitted, the system tries matching

the request with a corresponding valid match.

3. If a match is found, the result is sent back to

the two matching members.

The prototype is implemented in the Python

programming language. MySQL is used as the

database for storing the entries. Members who

request a ride and members who want to offer a ride

send a Short Message Service (SMS) message

containing information about the ride to a mobile

number. The SMS message containing the

information is forwarded to an email address. A

Python script uses Post Office Protocol (POP3) to

fetch the email message and then parses the email to

extract the data required for the database. The

primary database entries that are important are:

driver, passenger, or either, source/destination

locations, and the time window in which the user

needs or could offer a ride. Once we have the

matching entries we send back SMS messages to

both the member who requested the ride and the

member who offered the ride.

5 PSYCHOLOGICAL FACTORS

Psychological research on multitasking in human

performance, establishing trust, motivation to

participate, and community adhesion contribute to

the possible influences on people's decision to

participate in a carpooling service. We begin with a

short history.

5.1 Carpooling and its Decline

Historically, carpooling has had varying points of

popularity. It first appeared in U.S. policy during

World War II because of oil and rubber shortages

and did not reappear again until the mid 70s because

of an oil crisis (Ferguson, 1997). Carpooling became

a research topic in the late 1970s as the federal

government began to provide carpool demonstration

projects. This was a time when carpooling seemed to

have great promise. In terms of demographic

characteristics, studies at the time revealed that

carpoolers were not different from people who drove

alone (Oppenheim, 1979).

However, hope for the marketing of carpooling

came to a decline in the mid 1980s as baby boomers

moved to the suburbs in great numbers, creating

road congestion and fewer carpools despite the

marketing for it (Pisarski, 1987). Increasing

household vehicle availability, lowering of real

marginal fuel cost, and higher average educational

attainments among commuters were found to be

significant influences on the decline of carpooling

(iCarpool, 2009). In 1980, 19.7% of commuters

carpooled to work where only 10.7% of the US

population carpooled in 2005. Use of public

transportation was also low with less than 5% of

workers using buses, trains, or subways (Sharpe,

2007).

5.2 Safety of Mobile Phone Use

Experimental studies of driver performance and case

reports of crashes inform us about the risks involved

in cell phone use while driving.

Experiments on driver performance in simulated

driving studies point to the decreased safety of

operating a motor vehicle while talking on a cell

phone. It has been shown that both handheld and

hands-free cell phone use increases mental

workload, which interferes with an individual’s

information processing abilities (Törnros & Bolling,

2005). The speed of the vehicle is also slowed by

cell phone use, but only in a handheld condition,

possibly to compensate for the increased workload.

Initiating a call on a cell phone poses an even

greater traffic safety hazard (Törnros & Bolling,

2005). There was more lateral deviation in the lane

when dialing a number on a cell phone than when

just conversing on the phone. This was true for both

handheld and hands-free conditions.

Treffner & Barrett (2004) found that driving

while talking on a hands-free cell phone detracts

from the driver’s ability to control the car, compared

to driving in silence. There was no effect of

conversation difficulty, indicating that it is the mere

act of talking on the cell phone that impairs

performance, not the degree of complexity for the

conversation.

Event-related brain potentials (ERPs) and

reaction times were examined in hands-free and

handheld conditions (Garcia-Larrea et al., 2001).

ERPs were collected that reflected the speed of

processing, allocation of attentional resources, and

preparedness to respond to a visual task. The ERP

CARPOOLNOW - Just-in-Time Carpooling without Elaborate Preplanning

221

data from this study suggest that using a cell phone

at the same time one is performing a visuo-motor

task impacts two physiological mechanisms. The

first is a general decrease in attention allotted to

sensory inputs. This is likely due to the effects of the

situation being a “dual-task,” meaning that the

addition of a conversation divides the resources that

would otherwise be focused on only the task of

driving. This decrease in attention is seen in both the

hands-free and handheld conditions. The second

finding is that there is a weakening of preparedness

to respond with a motor action, but this was limited

only to the use of a handheld phone.

With regard to case reports of crashes, a study

done in Japan revealed 129 crashes to be cell phone

related in the month of June, 1996 (Yomiuri, 1999).

At the time of the crash, 42% of people where

answering the phone, 32% were dialing on the

phone, 16% were talking, 5% were in the process of

hanging up the phone, and the status of the

remaining 5% was unknown.

Hospital emergency room cases were

documented in Perth from April 2002 to July 2004

(Yomiuri, 2009). The increased risk was estimated

by comparing cell phone use in the 10 minutes prior

to the crash to how a person was driving in the week

before to the crash. The risk of crashing increased

fourfold when drivers were talking on a cell phone.

This was true for both handheld and hands-free

phones.

In sum, it is a common misconception that a

hands-free cell phone would be a better choice for

drivers than a handheld phone. Cell phones should

not be used while driving, whether or not it is hands-

free or permitted by law.

5.3 Establishing Trust

One of the greatest barriers to carpooling is trust,

which is obviously multidimensional (Paine et al.,

2008) and fits with the theory that behavior has

multiple influences (Massaro, 1998). Ability,

integrity, and benevolence are three dimensions of

trust identified by Bhattacherjee (2002). We have

identified a number of features to instill trust to

attract new members and to maintain a comfortable

membership. Carpooling membership can build on

current internet practices. Buyers and sellers know

each other’s history on eBay, Amazon.com readers

have access to summaries and reviews written by

other users of the site, and social networking sites

have various protective devices such as age

requirements or credit card enrollment.

As much evidence as possible about members

should be mutually available without compromising

their privacy. Members could produce a simple

personal profile page that would primarily aggregate

other web presences. Members could easily link to

their professional or personal pages, which would

allow a member's pre-existing web presence to be

utilized to bring the information necessary for trust

building. It is also conceivable to extend the

information on a profile to include driving data not

available elsewhere on the web, such as a traffic

record from the DMV and their insurance status.

To further promote trust members would have

the option of creating "favorites" and "blocked" lists

of members. The service would give preference to

users on a member's "favorites" list and exclude

those on the "blocked" list. Manually created lists

could prioritize carpools with acquaintances and

avoid particularly awkward matches. Automatically

created and updated smart lists could let members

set parameters for their matches, allowing for a user-

customized matching routine. For instance, a

member could block smokers and prefer dog lovers.

There would also be a way for users to decline a

match made by the service. Having this degree of

control over the matches may increase the trust in

any service.

5.4 Motivation to Participate

Although a high-degree of trust is important, it may

be insufficient to compel individuals to participate in

carpooling. Individuals with a pro-social attitude

who also have a high degree of trust—as opposed to

individuals with a pro-self attitude and any level of

trust—are more likely to participate in carpools

(Flannerlly & McLeod, 1989). For those with pro-

self attitudes who believe that driving alone is

preferable, marketing will have to focus on both

attitude change and trust.

Trust is also critical in terms of reliability,

referred to in the literature as “cybertrust,” or “trust

in … information and communication technologies”

(Flannerlly & McLeod, 1989). Users must trust that

the technology is reliable or else they are unlikely to

use it or recommend it to other users, and could even

discourage its use. This technological trust is

important for the development of our user-interface,

data security, ease of use, and numerous other

features of the design and implementation.

5.5 Community Adhesion

To ensure the success of and the need for dynamic

carpooling, community involvement and adhesion is

WEBIST 2009 - 5th International Conference on Web Information Systems and Technologies

222

necessary. In other words, its success is dependent

on the participation of the community and the

willingness of the community to work together.

Community involvement and adhesion is facilitated

through the acknowledgement and definition of a

problem. The public must express dissatisfaction

with the current transportation options and recognize

that others also struggle with getting to and from

different places. Acknowledging this helps the

community define the problem as a need for

alternative transportation services and to see

carpooling as a viable solution.

5.6 Socialization Via Carpooling

Dynamic carpooling might have significant impacts

on the local community. In many communities, for

example, there is still sufficient free space so that

much of what we do does not have to be public. In

big cities like New York, on the other hand, citizens

are surrounded by each other and most have learned

to behave more overtly. New Yorkers can be

described as not respecting the distinction between

people they know and strangers (Acocella, 2008).

Initiating interactions with strangers implies a good

deal of trust and comfort. Dynamic carpooling may

likely lead to an expansion of a person’s network of

acquaintances, if not friends.

5.7 Survey of Feasibility

An online survey was administered to all faculty,

staff, and students at our university, to investigate

whether transportation was considered a problem

and to measure the potential for community

involvement and adhesion around the topic of

transportation. Some of the survey results are shown

in Tables 1-4.

Table 1: The percentage of each group making up the

survey results based on 1945 respondents.

Faculty Staff Undergrads Grads

5% 31% 55% 9%

Table 2: Survey results (% of 1945 respondents).

Females 66%

Males 33%

Access to Car 71%

Live on Campus 20%

Drive Alone 30%

Carpool 6%

Bus 26%

No Convenient Bus Stop 40%

Table 3: Survey results (% of 1945 respondents).

Why Would You Carpool?

Better for the environment 82%

Save Money 81%

Meet New People 47%

More Convenient than Bus 67%

Table 4: Survey results (% of 1945 respondents).

Why Wouldn’t You Carpool?

Inconvenient 58%

Difficult to Organize 64%

Like to Travel Alone 12%

Wouldn’t Feel Safe 16%

Other 16%

The survey provided some informative

outcomes. One might have guessed there would not

be enough cars and free seats to support carpooling

in a university community but, in fact, almost 3 out

4 people have access to a car, and 3 out of 10 people

drive to campus alone. Carpooling would also seem

to be attractive to 4 out of 10 people because they

don’t have a convenient parking place. At least 78%

of the respondents said they would carpool with

faculty, staff, undergrads, grads, and members of the

opposite sex.

As can be seen in Tables 3 and 4, the majority of

community would carpool for the environment, to

save money, convenience, and half of them would

be interested in meeting new people. Although more

than half the respondents view carpooling as

inconvenient and difficult to organize, we expect

just-in-time carpooling would alleviate these

concerns. Finally, only a small fraction do not

appear to be among the potential membership

because of valuing being alone in commuting or

safety concerns. Even these individuals might

eventually be recruited to a successful carpooling

service.

6 BUSINESS OPERATIONS

The aforementioned research indicates that we must

take a multi-dimensional approach to the operation

of a dynamic carpooling service. This approach

must at a minimum include a consideration of

driving safety, trust, attitude change, and community

adhesion and do so in complex ways. Moreover, the

service must provide a benefit to both the drivers

and the passengers. The benefits to the passenger are

straightforward; the benefits to the driver come in

the form of a quicker commute (because it allows

CARPOOLNOW - Just-in-Time Carpooling without Elaborate Preplanning

223

use of the HOV lanes) with almost zero cost (since

the passenger is taking the same route as the driver)

and, of course, a financial incentive.

There are many issues to be resolved regarding

the payment by passengers for rides. The amount

each rider pays must be balanced: it needs to be

enough to be beneficial to the driver but not so much

as to be seen as detrimental by the passenger(s). The

carpool service should handle the payment to the

driver, making the experience simpler for both

driver and passenger, and provide its primary source

of income. Once the service reaches a critical

number of rides, it should be economically self-

sufficient. What is truly amazing about dynamic

carpooling is that it requires so little. It uses the

vehicles, infrastructures, routes, drivers, and

technologies that are already on the road today. With

our sustained effort, we believe these can be made to

work together to bring a greener future to private

transportation today.

One Internet-based ride-share matching service,

ZoomPool (2009), has independently begun work to

add several mobile applications to their service to

incorporate dynamic carpooling options. This

membership service performs multiple security

filtering to facilitate trust and encourages vouching

and other forms of recommendations among

members. The service shares the transportation costs

automatically between drivers and passengers to

decrease the awkwardness associated with

negotiating finances. ZoomPool aims to decrease the

barriers to carpooling resulting in a verifiable

improvement in personal carbon footprint. It is a

recently launched company that has yet to prove its

model.

REFERENCES

Acocella, J. (2008). You got a problem with that?

Smithsonian, May, 26-29.

Aktakita (2009). Retrieved on January 18, 2009, from

http://aktalita.com/.

Behind the Numbers (2009). Retrieved on January 18,

2009, from www.aaaexchange.com/.

Bhattacherjee, A. (2002) Individual trust in online firms:

Scale development and initial test. Journal of

Management Information Systems, 19, 211-241.

Ferguson, E. (1997). The rise and fall of the American

carpool: 1970-1990. Transportation, 24, 349-376.

Flannerlly, K.J. & McLeod, M.S. (1989). A multivariate

analysis of socioeconomic and attitudinal factors

predicting commuters' mode of travel. Bulletin of the

Psychonomic Society, 27 (1), 64-66.

García-Larrea, L., Perchet, C., Perrrin, F., & Amenedo, E.

(2001). Interference of cellular phone conversations

with visuomotor tasks: An ERP study. Journal of

Psychophysiology, 15, 14-21.

Hartwig, S, & Buchmann, M. (2007). Empty Seats

Traveling," Nokia Research Center Bochum, February

14, 2007.

iCarpool (2009). Retrieved on January 18, 2009, from

http://www.icarpool.com.

Massaro, D. W. (1998). Perceiving talking faces: From

speech perception to a behavioral principle.

Cambridge, Massachusetts: MIT Press.

Oppenheim, N. (1979). Carpooling: Problems and

potentials. Traffic Quarterly, 33, 253-292.

Paine Schofield, C. B. & Joinson, A. N. (2008), Privacy,

trust, and disclosure online. In A. Barak (Ed.)

Psychological aspects of cyberspace: Theory, research,

and applications (pp. 13-31). Cambridge: Cambridge

University Press.

PickupPal (2009). Retrieved on January 18, 2009, from

http://www.pickuppal.com

Pisarski, A.E., (1987). Commuting in America. Westport,

Connecticut: Eno Foundation for Transportation, Inc.

Sharpe, R. (2007). We're all alone on our drive to work.

USA Weekend Magazine. Sept. 9, usaweekend.com

Törnros, J.E.B, & Bolling, A.K. (2005). Mobile phone use

– effects of handheld and handsfree phones on driving

performance. Accident Analysis and Prevention, 37,

902-909.

Transportation and Climate (2009). Retrieved on January

18, 2009, from

http://www.epa.gov/OMS/climate/index.htm

Treffner, P.J., & Barrett, R. (2004). Hands-free mobile

phone speech while driving degrades coordination and

motor control. Transportation Research Part F, 7, 229-

246.

Ungemah, D., Goodin, G., Dusza, C., & Burris, M. (2007).

Examining Incentives and Preferential Treatment of

Carpools on Managed Lane Facilities. Journal of

Public Transportation, Vol. 10, No. 4, 151-170.

Virki, Tarmo. (2007). Global Cellphone Penetration

Reaches 50 pct. http://investing.reuters.co.uk/news.

Thu Nov. 9th

Yomiuri, S. (1999). Ban on cell phone use by drivers

having impact. The Daily

Yomiuri (2009). Retrieved on January 18, 2009, from

http://www.yomiuri.co.jp/newse/1216so04.htm,

ZoomPool (2009). Retrieved on January 18, 2009, from

http://www.zoompool.com/.

WEBIST 2009 - 5th International Conference on Web Information Systems and Technologies

224