This version is available at https://doi.org/10.14279/depositonce-8259
© 2015 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained
for all other uses, in any current or future media, including reprinting/republishing this material
for advertising or promotional purposes, creating new collective works, for resale or redistribution
to servers or lists, or reuse of any copyrighted component of this work in other works.
Terms of Use
Bischoff, J.; Maciejewski, M.; Sohr, A. (2015). Analysis of Berlin’s taxi services by exploring GPS traces.
2015 International Conference on Models and Technologies for Intelligent Transportation Systems (MT-
ITS), 209–215. https://doi.org/10.1109/MTITS.2015.7223258
Joschka Bischoff, Michal Maciejewski, Alexander Sohr
Analysis of Berlin's taxi services by
explorin
g
GPS traces
Accepted manuscript (Postprint)Conference paper |
2015 Models and Technologies for Intelligent Transportation Systems (MT-ITS)
3-5. June 2015. Budapest, Hungary
Analysis of Berlin’s taxi services
by exploring GPS traces
Joschka Bischoff
Technische Universit¨
at Berlin
Department of Transport Systems Planning
and Transport Telematics
Berlin, Germany
Email: bischof[email protected]
Michal Maciejewski
Poznan University of Technology
Division of Transport Systems
Poznan, Poland
Technische Universit¨
at Berlin
Department of Transport Systems Planning
and Transport Telematics
Berlin, Germany
Alexander Sohr
German Aerospace Center (DLR)
Institute for Transport technology
Berlin, Germany
Abstract—With current on-board GPS devices a lot of data
is being collected while operating taxis. This paper focuses on
analysing travel behaviour and vehicle supply of the Berlin taxi
market using floating car data (FCD) for one week each in 2013
and 2014. The data suggests that there is generally a demand peak
on workday mornings and a second peak over a longer time in
the afternoon. On weekends, the demand peaks shift towards the
night. On the supply side, drivers seem to adapt to the demand
peaks very efficiently, with fewer taxis being available at times of
low demand, such as during midday. A spatial analysis shows that
most taxi trips take place either within the city centre or from/to
Tegel Airport, the city’s largest single origin and destination.
Drivers spend a large amount of their work time on waiting for
customers and the taxi rank at Tegel Airport is the most popular
one.
Keywords—Taxi Berlin, floating car data, FCD, taxi demand,
taxi supply, Tegel Airport
I. INTRODUCTION
Although taxi services are one of the core elements of
urban transport systems, the scientific focus on them had
been limited until recent developments in information and
communications technology (ICT). Nowadays taxis equipped
with mobile devices may be monitored and managed remotely.
This has opened up a broad range of new services, such as
shared taxis, Uber and similar, or even autonomous taxis in
the future. Due to the onboard devices a lot of data is being
collected while operating taxis. These data make up a very
extensive source of information on urban traffic that can be
used, for example, to estimate time-dependent travel times.
They can also provide an insight into the travel behaviour,
which is the subject of this paper.
The paper deals specifically with taxis in Berlin. Being
Germany’s capital and largest city, the city forms a hotspot
for business, public services and tourists alike. In combination
with a comparingly low percentage of private car ownership
among its inhabitants, the city is an interesting location for
taxi businesses.
II. RELATED WORK
There have been several studies on analysing taxi data all
over the world, most of them focusing on the Asian market.
In Europe, taxi trips in Lisbon were analysed to explore
relationships between origins and destinations of taxi trips and
predict taxi demand [1]. The results of such analyses may be
used to improve the quality of taxi dispatch. In Hangzhou the
trajectories of 5 500 taxis were analysed to determine taxi
drivers’ best strategies to pick up passengers at a given time
and location [2]. A study in Shenzhen explored taxidrivers’
operation patterns with the focus on differences between the
behaviour of top drivers and of ordinary ones [3]. For Bejing,
a recommendation system for taxi drivers and passengers was
implemented to help them in efficient mutual finding [4]. In
Shanghai, spatio-temporal profitability maps were introduced
to help taxi drivers with reducing cruising miles [5]. Similarly,
prediction of taxi demand hotspots has been studied for Taipei
[6] and for Hangzhou [7].
Taxi trajectories have been used for estimating urban link
travel times, e.g. in Manhattan [8], and for time-dependent
path finding. In Berlin, fleet travel times were reduced by
measuring real-time travel times using floating car data from
taxis [9]. Similarly, taxi trajectories were used to devised
shortest path searching algorithms for Wuhan [10] and Bejing
[11]. In Bejing, a parallel algorithm to process massive taxi
traces for detecting traffic hot spots was proposed [12].
Taxi trajectories have been used also in many other areas.
One approach is to use them to detect flawed urban plan-
ning [13], another one to determine land use by analysing taxi
pick-ups and drop-offs [14]. Also spatial interaction models
were calibrated using taxi trajectories [15].
III. DATA
In December 2012, around 8 000 taxis were licensed to
operate in the city. Those vehicles were operated by around
18 000 taxi drivers which were organised in 3 000 taxi
ISBN 978-963-313-142-8 @ 2015 BME P – 1
2015 Models and Technologies for Intelligent Transportation Systems (MT-ITS)
3-5. June 2015. Budapest, Hungary
companies. Of these companies, 2 368 operate only one vehicle
each, 206 companies have two vehicles and 427 three or more
taxicabs [16]. Metered taxi fares are strictly regulated in Berlin.
Market access for new operators comes only with technical
entrance barriers, but, in contrast to many other countries, there
is no limitation in the amount of available taxi licenses.
The data used in the study is based on historical trajectories
of taxicabs managed by Taxi Berlin, Berlin’s largest taxi
association dispatching roughly 70% of the city’s taxi fleet.
Additionally a small fleet of about 30 taxis in Potsdam is
connected to the system. The GPS position of each taxi is
logged every 30–60 seconds depending of its current status. If
the cab has a customer on board, the position is sent once in
a minute, because it is not available for dispatching purposes
and thus its position is less important for the taxi association.
Consequently empty vehicles are tracked more frequently to
warrant an allocation of the closest vehicle to new customers.
Overall, around 2 million GPS positions are recorded per day.
This way, a database of historical GPS traces arises. Each entry
contains the following fields:
•temporary taxi identifier (ID) – changing several times
a day to assure that taxi drivers’ privacy is not com-
promised (Taxi Berlin is an association and not the
company hiring taxi drivers)
•coordinates – geographical location
•time – time the record is logged
•status – one of ten different statuses such as driving
to a customer, driving with a customer, waiting for a
customer, waiting at a taxi rank, etc.
To build trajectories from raw GPS positions, first some
basic plausibility checks are performed. These include:
•the taxi’s location
•the entry’s uniqueness
•improbable location variations
•a validation of the currently driven speed
In the next step, based on the status information, the start and
the end of the trajectories is determined. These trajectories
are then map matched and travel times within the city can be
derived, which is the main purpose of data collection both for
commercial and research reasons [17].
Extraction of taxi demand and spatiotemporal supply dis-
tribution from the GPS traces results in the following data:
•hourly zone-based1OD-matrices based on the vehi-
cles’ statuses
•the amount of taxicabs per status and zone in five-
minute intervals
1For Berlin, zones are determined by LORs http://www.stadtentwicklung.
berlin.de/planen/basisdaten stadtentwicklung/lor/, for the surrounding areas
according to municipalities
•the duration and time stamp of each temporary taxi
ID with the relevant vehicle status
There are, however, some limitations related to the proper-
ties of the trajectory database. First of all, due to the automatic
anonymisation (IDs are not fixed over a day), it is not possible
to provide full daily trajectories of individual taxi drivers.
Secondly, data sets only include vehicles that are associated
within Taxi Berlin. Consequently, it does not include taxi
drivers which exclusively pick up passengers at taxi ranks or by
roaming through streets. These drivers supposingly constitute
for a large share of taxi supply at places with a high likelihood
for on-the-spot trips, e.g. airports or night clubs. Last but not
least, recent estimation suggest that up to 40% of all taxi
trips in Berlin are black-market rides without proper fiscal
records [18]. This may result in errors while analysing status
data, as drivers will set their status to ’available’ when, in fact,
they have a customer on board. Hence, not all trips could be
tracked correctly for this paper. Thus, the demand and amount
of taxi trips may in fact be higher.
Two weekly data sets of GPS traces are used in this study.
The first set dates from 15 to 21 April 2013, whereas the
second includes data from 7 to 13 April 2014. The demand
for taxi trips is depending on several outer impacts such
as holidays, availability of transport alternatives and weather
conditions. Overall, these outer impacts were rather similar in
both weeks monitored in this study: Both weeks lie outside
major German holiday periods, however, during the week in
2014, the federal states Niedersachsen and Bremen had school
holidays. In neither week any large-scale event took place.
Rapid public transit was running without any planned major
interruptions during both time periods. For the week in 2013,
the outside temperatures were overall higher than 2014. In
neither week high rainfall had been measured.
IV. FINDINGS
A. Time-based taxi demand and supply
A detailed distribution of taxi supply (the number of taxis
logged into the system) and demand (the number of requests
submitted per hour) over a week are presented in Fig. 1 for
2013 and Fig. 2 for 2014. For the demand side, the following
applies:
•Taxi demand during weekdays (Monday morning –
Friday morning) follows a clear pattern with a major
peak around 9 am and a smaller one during the
afternoon.
•Friday is demand-wise the busiest day, with a some-
what smaller morning peak followed by a high demand
during evening and night hours.
•The weekly peak-hour for taxi demand is Saturday
night.
•Besides Friday and Saturday night, the demand for
taxi services during weekends is low.
ISBN 978-963-313-142-8 @ 2015 BME P – 2
2015 Models and Technologies for Intelligent Transportation Systems (MT-ITS)
3-5. June 2015. Budapest, Hungary
0
500
1000
1500
2000
2500
3000
3500
4000
Mon Tue Wed Thu Fri Sat Sun
amount
weekday
requests 2013
taxis 2013
Fig. 1. Request submissions per hour and active taxis (2013)
0
500
1000
1500
2000
2500
3000
3500
4000
Mon Tue Wed Thu Fri Sat Sun
amount
weekday
requests 2014
taxis 2014
Fig. 2. Request submissions per hour and active taxis (2014)
As both figures show, the supply adapts surprisingly well to
the demand. Peaks in demand are generally matched by a larger
supply. For most of the time, the number of available taxis is
higher than of hourly request submissions, which means that
every taxi serves less than one request per hour. Saturday night
in 2013 is the only exception to this with less taxis being in
service than requests per hour.
Overall, the week in 2013 has a higher demand for taxi
trips, with 198 445 trips in total versus 169 651 trips for the
week in 2014. This is mainly due to a higher demand during
Saturday night. However, the demand in 2014 is lower at most
other times as well (see Fig. 3) On the other hand, the amount
of active vehicles is higher during most times in 2014 (see
Fig. 4). This goes in line with the trend of the steadily growing
taxi fleet in Berlin over the last years[16] .
B. Trip distances
The average trip distances, calculated as the beeline dis-
tance between the zone centres, are at most weekday hours
between four and five kilometres (Fig. 5). However, the
weekday mornings (5–7 am), with average distances of six
to eight kilometres, form a notable exception. The day with
the highest average trip distance is, on the other hand, Sunday.
Overall, average trip distance is similar for 2013 and 2014.
As to the distribution of trip distances (cf. Fig. 6), there are
0
500
1000
1500
2000
2500
3000
3500
4000
Mon Tue Wed Thu Fri Sat Sun
amount
weekday
requests 2014
requests 2013
Fig. 3. Request submissions per hour (2013 and 2014)
0
500
1000
1500
2000
2500
3000
3500
4000
Mon Tue Wed Thu Fri Sat Sun
amount
weekday
taxis 2014
taxis 2013
Fig. 4. Active taxis (2013 and 2014)
few trips of less than two kilometre beeline distance, whereas
a trip distance between two and three kilometres is the most
common. Longer distances are less and less likely – not even
ten percent of all trips are ten kilometres or longer.
C. Location-based taxi demand
Origins and destinations of taxi trips are generally spread
all over the city. The majority commences and ends within the
city centre (defined as the inner railway circle), as Figure 7
shows. The data suggest that there are more trips from the
city centre to the outside than vice versa, though this could be
due to the fact that many trips from Tegel Airport are made
by taxis that are not affiliated with a radio-cab operator2and
most trips from Sch¨
onefeld airport are made by taxis which
are registered in a surrounding parish rather than Berlin. Both
types of taxis are not a part of the data set.
Fig. 8 shows the overall top 100 OD-relations and the
attractiveness, measured as the number of incoming trips, of
zones during the week in 2013. Notable are the following
points:
2For unaffiliated drivers it is easier to wait for passengers at the airport or
other high-demand locations than to search for them cruising through the city.
ISBN 978-963-313-142-8 @ 2015 BME P – 3
2015 Models and Technologies for Intelligent Transportation Systems (MT-ITS)
3-5. June 2015. Budapest, Hungary
0
2
4
6
8
10
Mon Tue Wed Thu Fri Sat Sun
distance [km]
weekday
2014
2013
Fig. 5. Average beeline trip distance at different times (2013 and 2014)
•Trips between the city centre and Tegel Airport make
up for most of the top OD relations and is hence also
the most sought-after destination.
•Other frequently used relations are almost exclusively
within the city centre, with one notable exception (see
below).
•In general, the central zones are more likely to be
destinations of taxi trips. Yet the picture is rather
diverse with some outskirts also being a popular
destination.
•The city’s second airport, Sch¨
onefeld, is neither a
destination nor an origin in the top 100 relations.
The notable exception is one relation are trips from southern
Pankow to the northern part of the district, with a total of
106 trips. In the opposite direction only 44 trips have been
measured. The relation seems unusual, but can be explained
by construction works on the tram line M1 which connects the
area during the week monitored. Especially in the northbound
direction the offered bus replacement may not have provided
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
(0, 1]
(1, 2]
(2, 3]
(3, 4]
(4, 5]
(5, 6]
(6, 7]
(7, 8]
(8, 9]
(9, 10]
(10, 11]
(11, 12)
(12, 13]
(13, 14]
(14, 15]
(15, 16]
(16, 17]
(17, 18]
(18, 19]
(19, 20]
(20, +∞)
share
beeline distance [km]
Fig. 6. Taxi trip distance distribution (2013 and 2014)
Fig. 7. Taxi trip allocation within the city (2013 and 2014)
the expected level of service, for example, due to lost connec-
tions. This may have lead to an increase in taxi demand.
Points of special interest are often a destination for taxi
trips. These include train stations, the fair grounds and major
event locations (the airport-related taxi traffic is analysed in
Section IV-D). Trips to a selection of such locations are
displayed in Figure 9. Of all train stations with regular long-
distance connections, Hauptbahnhof (Central Station) has the
biggest attraction for taxi traffic, with more than ten weekly
trips from almost any zone in the city centre. Ostbahnhof (in
the east) and Spandau (in the far west) stations attract taxi
business to a much lesser degree and mainly on a district-wide
level. Some zones notably have an equally high attraction to
both Hauptbahnhof and Ostbahnhof. Finally, S¨
udkreuz (in the
south) and Gesundbrunnen (in the north) are almost exclusively
destinations for short-range taxi trips. The fair ground has a
Fig. 8. Number of incoming taxi trips per zone (blue colour scale) and the
top 100 OD-relations (variable-width purple lines) for the week in 2013.
Map source: OpenStreetMap contributors.
ISBN 978-963-313-142-8 @ 2015 BME P – 4
2015 Models and Technologies for Intelligent Transportation Systems (MT-ITS)
3-5. June 2015. Budapest, Hungary
Fig. 9. Selection of Berlin’s points of interest and the incoming taxi flows
of at least 10 trips per week (2013)
Map source: OpenStreetMap contributors.
certain amount of trips from Tegel Airport and the western
city centre around Kurf¨
ustendamm. Since there was no bigger
trade fair, this seems reasonable. For event locations, no clear
pattern emerges. Taxi trips to zones near them are generally
short-distance trips from within the city centre, with a notable
exception around the Tempodrom. However the likelihood of
travelling by taxi to a mass event is also depending on the
actual event itself.
D. Airports
As already seen in the previous analyses, taxi trips to
and from Tegel and Sch¨
onefeld airports constitute for a large
amount of all taxi trips in Berlin. The following numbers are
once more based on the week in April 2013. The amount of
trips to Tegel is much higher (14 144 trips) due to its proximity
to the city, the lack of a direct rail link and higher air passenger
numbers. To Sch¨
onefeld, only 1 938 trips took place. On the
other hand, with 7.48 km, the average beeline trip distance to
Tegel is only half as long compared to Sch¨
onefeld (14.27 km).
As Fig. 10 shows, origins for trips to the airports are similar in
the city centre, but tend to differ in the outskirts. One possible
interpretation would be that passengers departing in southern
parts of the city would rather use a taxi to Sch¨
onefeld than
to Tegel and vice versa. Thus the likelihood of taxi usages
decreases with a growing distance from the airport.
E. Symmetries in taxi demand
For taxi dispatching, one of the more important questions
is, whether taxi demand between different city areas is sym-
metric and whether trips in one direction occur at different
times than in the other direction. In general, demand for taxi
trips is rather symmetric between zones. The most notable
exceptions may be explained by the aforementioned reduced
share of Taxi Berlin in serving trips originating at both airports.
Fig. 11 provides an overview. The asymmetry is best illustrated
by taxi flows incoming to and outgoing from Tegel Airport.
Fig. 10. Taxi flows incoming to Berlin’s airports (2013).
Map source: OpenStreetMap contributors.
F. Idle taxi distribution
Within Berlin, there are roughly 400 taxi ranks, which is
where most taxi drivers gather. Figure 12 provides an overview
of rank locations and the average number of idle taxis in each
zone at each moment during the week in 2014. Tegel Airport
attracts overall most vehicles. Drivers accept waiting times of
up to several hours at this rank. One reason might be that the
following ride has the chance of being especially rewarding.
Furthermore, the general infrastructure for waiting taxi drivers
is rather good here, so some drivers might prefer it over other
ranks.
0
2000
4000
6000
8000
10000
12000
14000
16000
0 2000 4000 6000 8000 10000 12000
trips to zone
trips from zone
2013
TXL
Fig. 11. Outgoing vs. incoming taxi flows by zone for the week in 2013
ISBN 978-963-313-142-8 @ 2015 BME P – 5
2015 Models and Technologies for Intelligent Transportation Systems (MT-ITS)
3-5. June 2015. Budapest, Hungary
Fig. 12. Average amount of idle taxis per zone (2014)
G. Taxi dispatching statistics
An illustration of the average daily occupation of taxi
drivers is presented in Tab. I, where all types of the taxi statuses
(see Sec. III) have been aggregated into the following four
main categories:
•to-customer – driving to and waiting for a customer
•with-customer – driving with a customer
•at-rank – standing at a taxi rank
•outside-rank – idle but not at a rank, e.g. cruising or
returning to a rank
First of all, one can clearly see that Taxi Berlin typically
operates under low load and taxis remain idle for over 70%
of time. Secondly, the similar share of to-customer and idle
categories proves that after serving a customer, drivers tend to
return to one of the nearest taxi ranks and thus do not have
only one favourite rank (the triangle inequality does not hold,
i.e. to-customer +outside-rank <with-customer). However,
due to the aforementioned anonymisation of the GPS traces, it
is impossible to be exact on the daily mileage of taxicabs and
drivers, or to find out whether drivers have a set of their own
favourite taxi ranks distributed over the city or they do not
have any preference at all. Last but not least, the high share
of the at-rank category may result from not reporting all trips
in order to avoid paying fees or taxes (see Section III).
V. CONCLUSION
Berlin’s taxi traffic under usual demand conditions is in
general relaxed. The supply adjusts to the demand very well
TABLE I. TAXI DRIVER ACTIVITIES AND THEIR TIME SHARE (2014)
Activity Time share [%]
to customer 6.7
with customer 21.8
at rank 62.7
outside rank 8.8
with biggest demand peaks being weekday mornings and
weekend nights. These peak patterns are repeating both in 2013
and 2014. The top single origin and destination is clearly Tegel
Airport, most trips however take place within the city centre.
A correlation between taxi demand and interruptions in public
transit is noticeable in at least one case.
ACKNOWLEDGMENT
The authors would like to thank the Einstein foundation for
co-funding this paper and TAXI Berlin TZB GmbH for data
provision.
REFERENCES
[1] M. Veloso, S. Phithakkitnukoon, and C. Bento, “Sensing urban
mobility with taxi flow,” in Proceedings of the 3rd ACM SIGSPATIAL
International Workshop on Location-Based Social Networks, ser.
LBSN ’11. New York, NY, USA: ACM, 2011, pp. 41–44. [Online].
Available: http://doi.acm.org/10.1145/2063212.2063215
[2] B. Li, D. Zhang, L. Sun, and C. Chen, “Hunting or waiting? discovering
passenger-finding strategies from a large-scale real-world taxi dataset,”
in 8th IEE International Workshop on Managing Ubiquitous Commu-
nications and Services, 2011.
[3] L. Liu, C. Andris, and C. Ratti, “Uncovering cabdrivers behavior
patterns from their digital traces,” Computers, Environment and Urban
Systems, vol. 34, no. 6, pp. 541–548, 2010.
[4] J. Yuan, Y. Zheng, L. Zhang, X. Xie, and G. Sun, “Where to find my
next passenger,” in Proceedings of the 13th international conference on
Ubiquitous computing. ACM, 2011, pp. 109–118.
[5] J. Powell, Y. Huang, F. Bastani, and M. Ji, “Towards reducing taxicab
cruising time using spatio-temporal profitability maps,” Advances in
Spatial and Temporal Databases, pp. 242–260, 2011.
[6] H.-w. Chang, Y.-c. Tai, H. Chen, J. Y.-j. Hsu, and C. Kuo, “itaxi:
Context-aware taxi demand hotspots prediction using ontology and data
mining approaches,” in Proceedings of the 13th Conference on Artificial
Intelligence and Applications (TAAI 2008), 2008.
[7] X. Li, G. Pan, Z. Wu, G. Qi, S. Li, D. Zhang, W. Zhang, and Z. Wang,
“Prediction of urban human mobility using large-scale taxi traces and its
applications,” Frontiers of Computer Science, vol. 6, no. 1, pp. 111–121,
2012.
[8] X. Zhan, S. Hasan, S. V. Ukkusuri, and C. Kamga, “Urban
link travel time estimation using large-scale taxi data with partial
information,” Transportation Research Part C: Emerging Technologies,
vol. 33, no. 0, pp. 37 – 49, 2013. [Online]. Available: http:
//www.sciencedirect.com/science/article/pii/S0968090X13000740
[9] R. Ebendt, A. Sohr, L. Touko-Tcheumadjeu, and P. Wagner, “Dynamis-
che Neuplanung der Touren von Express Trucks unter Einbeziehung
einer FCD-basierten Verkehrslage,” Multikonferenz Wirtschaftsinfor-
matik, 2012.
[10] Q. Li, Z. Zeng, T. Zhang, J. Li, and Z. Wu, “Path-finding through
flexible hierarchical road networks: An experiential approach using taxi
trajectory data,” International Journal of Applied Earth Observation and
Geoinformation, vol. 13, no. 1, pp. 110 – 119, 2011. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0303243410000930
[11] J. Yuan, Y. Zheng, C. Zhang, W. Xie, X. Xie, G. Sun, and Y. Huang, “T-
drive: driving directions based on taxi trajectories,” in Proceedings of the
18th SIGSPATIAL International Conference on Advances in Geographic
Information Systems. ACM, 2010, pp. 99–108.
[12] Z. GUI and H. YU, “Mining traffic hot spots from massive taxi trace,”
Journal of Computational Information Systems, vol. 10, no. 7, pp. 2751–
2760, 2014.
[13] Y. Zheng, Y. Liu, J. Yuan, and X. Xie, “Urban computing with taxicabs,”
in Proceedings of the 13th international conference on Ubiquitous
computing. ACM, 2011, pp. 89–98.
ISBN 978-963-313-142-8 @ 2015 BME P – 6
2015 Models and Technologies for Intelligent Transportation Systems (MT-ITS)
3-5. June 2015. Budapest, Hungary
[14] Y. Liu, F. Wang, Y. Xiao, and S. Gao, “Urban land uses and
traffic source-sink areas: Evidence from GPS-enabled taxi data in
Shanghai,” Landscape and Urban Planning, vol. 106, no. 1, pp. 73
– 87, 2012. [Online]. Available: http://www.sciencedirect.com/science/
article/pii/S0169204612000631
[15] Y. Yue, H. dong Wang, B. Hu, Q. quan Li, Y. guang Li, and A. G.
Yeh, “Exploratory calibration of a spatial interaction model using
taxi GPS trajectories,” Computers, Environment and Urban Systems,
vol. 36, no. 2, pp. 140 – 153, 2012, special Issue: Geoinformatics
2010. [Online]. Available: http://www.sciencedirect.com/science/article/
pii/S0198971511000901
[16] BZP - Deutscher Taxi- und Mietwagenverband e.V., “Zahlen ¨
uber den
Taxi- und Mietwagenverkehr 2013/2014,” 2014. [Online]. Avail-
able: http://www.bzp.org/Content/SERVICE/Geschaeftsbericht/index.
php?Z highmain=6&Z highsub=2&Z highsubsub=0
[17] E. Brockfeld, A. Sohr, and R. Ebendt, “Validation of a Taxi-FCD
system by GPS -testdrives,” 2014, german Aerospace Center, Institute
of Transport Systems. [Online]. Available: http://elib.dlr.de/66060/1/
TP014-4.pdf
[18] S. Sauer, “Wie der Mindestlohn das Taxi-Gewerbe veraendert,”
Berliner Zeitung, 2014. [Online]. Available: http://bit.ly/16MeUBH
ISBN 978-963-313-142-8 @ 2015 BME P – 7
