Machine Translated by Google

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/333117435

Google Maps

Article in International Journal of Computer Applications · May 2019

DOI: 10.510/ijca2019918791

CITATIONS READS WILL

18 14,361

3 authors, including:

Pratik Kanani
Dwarkadas J. Sanghvi College of Engineering
79 PUBLICATIONS 216 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Security in next generation networks View project

Political implications of Xenophobia View project

All content following this page was uploaded by Pratik Kanani on May 9, 2020.

The user has requested enhancement of the downloaded file.

Machine Translated by Google

International Journal of Computer Applications (0975 – 8887)

Volume 178 – No. 8, May 2019

Google Maps
Heeket Mehta Pratik Kanani Priya Lande
Dept. of Computer Engineering Dept. of Computer Engineering Dept. of Computer Engineering
Dwarkadas J Sanghvi college of Dwarkadas J Sanghvi college of Dwarkadas J Sanghvi college of
Engineering, University of Engineering, University of Engineering, University of
Mumbai, India Mumbai, India Mumbai, India

ABSTRACT This devised an algorithm to find the shortest route from the source
to the destination.
paper gives a perspective about how Google Maps, one of the
world's most influential application works. Google Maps was ALGORITHM:-
initially coded in C++ programming language by its founders - 1.1.1Choose an initial node in the weighted graph which serves
Lars and Jens Eilstrup Rasmussen. Formerly it was named as a source node.
'Where 2 Technologies', which was later acquired by Google
Inc. in 2004, which renamed this web-application to Google 1.1.2The distance of the other nodes is intialised as infinity
Maps. Earlier it had limited features restricted to navigation, but with respect to the source node.
today it provides overwhelming features like street-view, ETA
1.1.3An empty set N containsns the nodes to which the shortest
and other interesting features.
path has been found.
It gives an overview about the algorithms and procedures
1.1.4The distance of the source node with respect to itself is
employed by Google Maps to carry out analysis and enable zero. It is added to N and thus termed as 'active node'.
users to carry out desired operations. Various features provided
by Google Maps are portrayed in this paper. It describes the 1.1.5Consider the neighboring nodes of the active node which
algorithms and procedures used by Google Maps to find the are connected to it by a weighted edge. Sum up the
shortest path, locate one's position, geocoding and other such distance with the weight of the edge.
elegant features it provides its users.
1.1.6If the neighboring nodes already have some distnace
Keywords assigned, then update the distance if the newly obtained
Heuristics, A* algorithm, Dijkstra's algorithm, Street View, distance in step 5 is less than the current value.
Rosette, Optical flow, Ceres Solver, Geocoding, Triangulation, 1.1.7Now, choose the node which has the minimum distance
Trilateration, WAAS, GNSS, Geo-Positioning System [GPS]. assigned to it and insert it into N. The newly inserted node
is now the 'active node'.
1. INTRODUCTION Google 1.1.8Repeat steps 5 to 7 until the destination node is included
Maps is one of the most sought after innovation in the history of in the set N or if there are no more labeled nodes in N.
technology. The advent of this feature by the tech giant Google
PSEUDO CODE:-
Inc. Google Map enables people to navigate and find the
shortest and most convenient route to their desired destination. function dijkstra_algorithm (graph, S) // S represent source
According to a recent survey, Google Maps has acquired almost for each node N in graph:
64 million users. Moreover, it has included new features like
street-view, location of hospitals, cafes, police-stations and distance[N] := infinity
many more helpful features. The algorithms, techniques and
prev[N] := NULL
technology used by Google Maps is cutting-edge and highly
advanced. The team of engineers at Google, preserve and if (node N != S), add node N to the set Q
analyze myriad datasets including historic and real-time data,
distance[S] := 0
which is what makes Google Maps so progressive and accurate.
The prediction models continuous valued functions ie, predicts Q:= set of all nodes in graph
new values. Also ML models can be used and applied on
traditional computing techniques to improve the existing while Q is not empty:
accuracies [1-5]. Z := node in Q with minimum distance
2. GOOGLE MAPS: ALGORITHMS AND TECHNIQUES for each unvisited neighbor N of Z:
1.1 Dijikstra's Algorithm dist := distance[Z]+ edge_weight(Z,N)
Google maps employs Graph data
if dist<distance[N] // new value less than previous distance
structures for calculation of the shortest path from the source
(point A) to the destination (point B). Graph data structure distance[N] := distprev := Z
connecting of various nodes and multiple edges these nodes.
return distance[ ] , prev[ ] ……[6]
Dijkstra's algorithm is an effective and proficient algorithm The time complexity of Dijkstra's original algorithm is given
proposed by Edsger.W. Dijkstra to navigate the shortest distance
and path to reach a given destination. The nodes of the graph by O( | | ) , but when it is implemented based on minimum-
are connected by weighted edges, which represents the distance priority queue implemented by Fibonacci Heaps, then the time
to be traversed to reach there. Thus Dijkstra complexity reduces to O (|E| + |V|log|V|).

41
Machine Translated by Google

International Journal of Computer Applications (0975 – 8887)

Volume 178 – No. 8, May 2019

Google maps deals with a mammoth of data, thus the number of Owing to the high accuracy, flexibilty, ability to deal with mammoth
nodes in the graph are myriad. Thus, graph algorithms like Dijkstra's graphs and its stark advantages over Dijkstra's algorithm, Google
algorithm may fail due to the increased time and space complexity. Maps has shifted to deploying A* algorithm for obtaining the shortest
The drastic increase in the size of the graph limits the efficiency of path between the source and destination.
the algorithm. Heuristic algorithms like A* algortihm prove to be
implementable.
ALGORITHM:-
A* is similar to Dijkstra's algorithm, which uses a heurisitc function to 1) Initialise two lists -
navigate a better and more efficient path. A* algorithm assigns higher
priority to the nodes which are supposed to be better (checks for a) Open_list
parameters like time requirement, distance and other such
b)Closed_list
parameters) than the others, while Dijkstra explores all the nodes.
Therefore it is meant to be faster than Dijkstra's algorithm, even if the 2) Insert the source node into the open_list
memory requirement and operations per node are more, since it
explores a lot less nodes and the gain is good in any case. [7] 3) Keep the total cost of the source node as zero

4) While open_list is not NULL:

a) Check the node with minimum total cost (least f(x)) present
1.2 A* (A-star) Algorithm A* algortihm is a in the open_list
flexible, more efficient and cutting-edge algorithm which Google Maps
b) Rename the above node as X
currently utilises to find the shortest path between a given source and
destination. It has a wider range of contexts therefore, it can deal with c) Pop X out of the open_list
larger graphs which is preferable over Dijkstra's algorithm.A* is like
d) Generate the successors to X
Greedy Best-First-Search. A* algorithm is widely used since its
mechanism involves combining the pieces of information that Dijkstra e) Set X as the parent node of all these successors
algorithm uses (It favors the vertices close to the source) and
information that Greedy Best-First-Search uses (choosing nodes 5) for each successor:
close to the goal).[8 ]
a) if successor is the same as the destination:stop

b) else:
One of the difference between A* and Dijkstra's algorithm is that A*
algorithm, being a heuristic algorithm, has two cost functions : successor.g(x) = Xg(x) + distance between successor
and X

g(x) :- The actual cost to reach a vertex (same applies to successor.h(x) = distance from destination [goal] to
Dijkstra) successor (using heuristics)

h(x) :- Approximate cost to reach goal node from node n. (This is the successor.f(X) = successor.g(x) + suucessor.h(x)
heuristic part of the cost function)
c) if a node with same position as successor is present in
f(x) :- Total estimated cost of each node open_list, but having a lower cost [f(x)] than successor,
then skip the successor
The cost calculated by A* algorithm shouldn't be overestimated, since
it is a heuristic algorithm. Hence, the actual cost to reach the d) if a node with the same position as the successor is present
destination node, from any given node n should be either greater in the closed_list , but having a lower cost [f(x)], then skip
than or equal to h(x). This is called Admissible Heuristic. this successor, else insert the node into the open_list

Hence, one can represent the estimated total cost of each node by : end for loop

6) Add X into the closed list

f(x) = g(x) + h(x).
7) End the while loop
A node is expanded by A* search, only if it is considered to be
8) The route is obtained by the closed_list
promising. This algorithm focuses solely on reaching the goal node
from the node which is currently under consideration. It does not try 9) Stop.
to reach other nodes.
PSEUDO CODE:-
Thus, if the heuristic function taken into consideration approximates
the future cost accurately, then one needs to explore considerably function A_star_algorithm (start, goal)
less nodes as compared to Dijkstra's algorithm. [9]
open_list = Initialised set contains start //Initialisation of the open list
closed_list = NULL set //Initialisation of the closed list
A* algorithm terminates when the destination node has been reached
to offer that the total cost of reaching the destination node is the
start.g = 0
minimum compared to all the other paths. It can also terminate if the
destination node has not been reached despite all the nodes being start.f = start.g + h(start, goal) // f(x)= g(x) + h(x) where h is the
explored. Unlike, Dijkstra's algorithm, it does not explore all nodes heuristic function
possible, thus A* proves to be more proficient.
while open_list is not empty:

node = element in open list with lowest cost //f(x) should be minimum

42
Machine Translated by Google

International Journal of Computer Applications (0975 – 8887)

Volume 178 – No. 8, May 2019

if node = goal// when destination is reached, current node = goal

return route(goal) // route found

remove node from open_list

insert node into closed_list

for neighbors in neighbors(node):

if neighbor not in closed_list:

Fig.1. Cars and Camera Rig used.
neighbor.f = neighbor.g + h(neighbour, goal) //f(x)= g(x) + h(x)
where h is the heuristic function The development and creation of these panoramic views is a
strenuous task, involving image capturing from multi-camera setup
if neighbor is not in open_list:
called Rosette. The manually captured images are blended and
insert neighbor into open_list stitched using various computer editing and blending algorithms
and softwares. However, impediments like parallax, miscaliberation
else:
of the Rosette camera geometry, time difference can destroy or
x = neighbor node already present in open_list affect affect the image-processing and blending.

if neighbor.g < xg:

Despite such cautious procedure/steps, visible seams in overlap-
xg = neighbor.g region can still occur. In order to provide more seamless images
and to overcome the above drawbacks, the team at Google has
x.parent = neighbor.parent
developed a new algorithm based on Optical-Flow. The concept
return False // no path exists behind this is to warp/adjust each image in a way such that the
content aligns with the region of overlap. To eliminate/avoid any
function neighbors(node)
chance of distortion or drawback the process is to be carried out
neighbors = set of valid neighbors to node // check for obstacles meticulously. A robust and potent approach is required to
here counteract adverse lighting conditions, varying scene geometry,
calibration and parallax distortion, and other inevitably conditions.
for neighbor in neighbors: if neighbor is diagonal: [10]
neighbor.g = node.g + diagonal_cost // Pythagoras Theorem eg.
5^2 = 3^2 + 4^2 -> here 5 is the diagnol
else:

neighbor.g = node.g + normal_cost // eg. 3

neighbor.parent = node

return neighbors

function route(node) / Displays the route

path = the set which contains the current node

while node.parent exists: Fig.2. Parallax distortion due to miscalibration.

node = node.parent Furthermore, the images captured by rosettes are aligned and
adjusted corresponding to all the points from the overlap region.
add node to path These images are then stitched using softwares to obtain desired
panoramic, 3-D views. To generate the final results, the warping
return path
is formatted as a spline-based flow field with spatial regularization.
3. STREET VIEW – GOOGLE MAPS In 2007, the Google Google's open source Ceres Solver is used to compute the spline
parameters. [ten]
Maps team introduced an outstanding feature called Street View.
This brilliant innovation proved to be a milestone in the history of Therefore, after carrying out such complicated procedures, the
navigation and GPS. Street View, the newly introduced feature, images can be corrected, aligned and warped as pleased.
provided a 3- dimensional, HD, Panaromic view of many Google Maps' street-view feature is being refined and remodeled
neighborhoods, streets, avenues, and other such areas merely since their team is exploring more and new every places day. The
from one's mobile phone or computer. Surprisingly, these images expanse of street-view coverage is almost fully completed in
are manually taken by Google Maps team using a car specially numerous developed countries and is dramatically increasing in
fitted with multi-camera rig and other necessary equipments for the developing countries.
actually taking pictures of the neighborhoods they drive by. The Statistically, the Government of India reported that 14 miles are
following images depict the cars and cameras used to generate being covered each day by the Google Maps' Street View team.
the street view. The implementation of such a 3-D, Panoramic feature which
enables the users to view and virtually travel along the extreme
corners of the world is truly a technological breakthrough. [11]

43
Machine Translated by Google

International Journal of Computer Applications (0975 – 8887)

Volume 178 – No. 8, May 2019

technique, which can be extended to the mathematical principle

of Trilateration.To achieve 3-D trilateration, one need to take
consider spheres instead of two dimensional circles. Unlike 2-
D trilateration, the radii of the sphere is said to spread in all
directions, thus forming imaginary 3- dimensional spheres
around a given point. Thus the intersection of these three
sphere, gives the location of the object. In case of 3-D
trilateration, two intersection points are obtained because of
the 3 spheres. The Earth helps by eliminating one of the points
by acting as the fourth sphere.
This interested the search and makes it possible to locate the
exact position of the given object, place or person. [14]The
concept of Trilateration is echoed better by the following
diagrams:-

This is a system of satellites as well as ground stations which

works with GPS to refine and improve the quality of the system
and also helps in error detection and correction. A GPS receiver
Fig.3. Status of Street View compatibility in the world.
is capable of gauging any given position accurately within three
meters, with the assistance of WAAS. It has been developed
4. LOCATING THE POSITION (OF AN
by the Federal Aviation Administration (FAA) and the
OBJECT, PERSON OR PLACE) Department of Transport (DOT) to counteract the signal errors
USING GPS AND GEOCODING:- caused as a result of disturbances in the ionosphere, time
Global Position System (GPS) is an outstanding innovation errors and satellite orbit errors. WAAS is essential for obtaining
used for tracking an object using satellite systems currently the knowledge about the conditions of each satellite of the
present in the space. This type of tracking can be used for GPS network. [15]
military as well as civilian purposes. Google Maps uses the
civil GPS tracking and locates people and places. The tracking
system used by the Global Positioning System is the GNSS
network (Global Navigation Satellite System Network). This
system not only stores the location, but also has the memory
to store the speed, direction, time and other parameters.

The Global Positioning System consists of 27 GPS satellites

orbiting around the Earth, out of which 24 are operational and
3 are spare (To be used in case of failure). These revolve
around the Earth each 12 hours and emit civilian and military
navigation data signals, on two L-band frequencies. These
signals are emitted in space and are gathered by the GPS
receiver. Five monitoring stations along with four ground
antennas include the control of the Global Positioning System.
These are located across the globe which gathers the data Fig.4. Three Dimensional Trilateration
about the satellite's position which is further relayed to the
master control station at Schriever Air Force Base in Colorado,
USA [12]. Passive tracking by GPS involves storage of the
data obtained during tracking, while in active tracking some
information is relayed regularly through a modem within the
GPS system unit (2-way GPS) to a centralized database. [13]

The location tracking is based on a mathematical principle

called Trilateration. There are 2 types of trilateration - 2-D
trilateration and 3-D trilateration. This requires two parameters
–

ÿ First, the location of the place is required to be known,

which will subsequently be traced by atleast 3
satellites. The range of these satellites should be
Fig.5. Two Dimensional Trilateration
able to cover the specific position.

ÿ Next, the distance between the object (eg. a vehicle)

and one of the satellites is required to be known.
WAAS (Wide Area Augmentation System):-
With the help
Hence, for this, a miniature receiver/device is installed on the of the satellites, the receiver can now compute and detect the
vehicles/phones to capture the radio (electro-magnetic signals) latitude, longitude and altitude of the vehicle which are
emitted by the satellite. This helps to compute the distance supposed to include the hypothetical geometrical co-ordinates
between the vehicle and satellite as well as the satellite's of the vehicle. Thus, assigning and accessing the following
location. Thus, by locking the signals onto 3 satellites, the values is termed 'Geocoding', where the location of
location of the vehicle can be gauged by using Triangulation

44
Machine Translated by Google

International Journal of Computer Applications (0975 – 8887)

Volume 178 – No. 8, May 2019

a particular object is accessed using a hypothetical grid system, Thus, in the modern time, myriad factors apart from average
and then locating/plotting the specific co-ordinates (latitude, speed of the user on a specific route, the shortest path and
longitude, altitude) to obtain the required location of the object distance between two points have to be taken into consideration
on a given map. in order to determine and compute the estimated time of arrival.

Google Maps uses this algorithm of Global Positioning System

to detect and locate places on the map. Thus the signals However, Google Maps cannot take into consideration a user's
obtained from satellites in space and other such ground personal speed [16], but can only determine the approximate
networks, towers and other such facilities help in accurately time of arrival by taking into consideration factors like distance
locating the position of an object, place or person using remaining, average speed due to real-time traffic conditions,
algorithms like Triangulation, Trilateration and other such historic data and officially recommended speed . They can thus
complex algorithms. predict the approximate time needed to reach the destination
using any desired mode of transport (car, walking, train, cycle
5. ESTIMATED TIME OF ARRIVAL (ETA):- The most and other available modes). The average speed of a person in
basic and each of these modes are recorded and used for determining the
important criterion for Google Maps to estimate the time taken ETA.
to reach a particular destination is based on the route taken to The technology giant Google is one of the best company at
reach the given destination from the source. This is calculated collecting, manipulating, analyzing data and drawing very
using the A-star (A*) algorithm, which helps in obtaining the accurate conclusions about ETA, shortest path and other such
shortest route from the starting point to the desired destination. real-time data. Since it has so much access to data through
This would be the shortest path recommended by Google Maps cellular devices and the data signals emitted by them, Google
without taking into consideration the real-time traffic, which was Maps can almost correctly and efficiently predict the time taken
a major drawback. To overcome this drawback, Google collects by the user to reach his destination.
continuous data from all cellular devices on that particular way
and other routes possible. Manipulating this data, the average 6. CONCLUSION Therefore,
speed of any user can be determined and the shortest path can the algorithms, techniques, procedures and technology used by
then be decided. A user can find a route adding a few halts, Google Maps in order to render accurate and real-time
avoid freeways, avoid bridges and choose a location according information and data on their app, is understood.
to their desire, despite the shortest path recommended by Obsolete algorithms like Dijkstra's algorithm may not work since
Google Maps. the time complexity would increase and faster since a more
efficient, accurate and algorithm like A* has helped replaced it.
Before 2007, Google Maps was only able to calculate the Features like street view need a splendid level of image
estimated time of arrival (ETA) by taking into consideration the processing, editing and stitching.
distance between two given points and the average speed of Mathematical concepts like Trilateration, Triangulation and
the object or person between the source and destination. Geocoding are required to locate any position with the help of
The flaw with this was that a very important criterion of traffic imaginary geographical co-ordinates. Aerial imagery also helps
was overlooked. Today, Google Maps considers the current design the frontend of the app. The pictures captured by aerial
traffic condition on the selected route, which often proves to imagery are processed and subsequently used to give the map
increase the commute time. It provides the user with two ETA – layout which is seen when one opens the application.
one is at the average optimal conditions and the other is under Thus, Google Maps constantly keeps updating the old features
and introduces new ones for the benefit and convenience of its
current traffic conditions, which helps the user estimate the time
required to reach any particular destination. [16] users. Hence, Google Maps is one of the most helpful and
reliable application.

A former Google engineer Richard Russell wrote, answering a 7. REFERENCES [1]

question on Quora - “These things range from official speed P.Kanani, V.Kaul, K.Shah,” Hybrid PKDS in 4G using Securred
limits and recommended speeds, likely speed derived from road DCC”, International Conference on Signal Propagation and
types, historical average speed data over certain time periods Computer Technology (ICSPCT) 2014, pp-323-328.
(sometimes just averages, sometimes at particular times) of
day), actual travel times from previous users and real-time traffic
[2] M. Padole, P. Kanani, “Textimage Ciphering”, 2nd
information. They mix data from sources they have and come
International Conference for Convergence in Technology
up with the best predictions they can make.”[17]
(I2CT), 2017, pp - 926-930.

[3] Y. Doshi, A. Sangani, P. Kanani, M. Padole, “An Insight into

Furthermore, the collection of the data like average speed of
CAPTCHA”, International Journal Of Advanced Studies In
users traveling on the same paths, officially recommended
Computer Science And Engineering IJASCSE Volume 6,
speeds, comparison between optimum average speed and the
Issue 9, pp. 19-27. 2017.
current real-time speed due to traffic conditions help the
engineers and analysts at Google to calculate the estimated [4] T. Reddy, J. Sanghvi, D. Vora, P. Kanani, “Wanderlust : A
time of arrival (ETA) from point A to point B. Historic data also Personalized Travel Itinerary Recommender”, International
helps the engineers analyze the current traffic scenarios at a Journal Of Engineering Development And Research IJEDR,
certain point between two places, which contributes towards the Volume 6, issue 3, pp.78-83. 2018
fluctuations of the ETA. As rightly stated by former Google
engineer Richard Russell, the companies who have the best, [5] P. Kanani, K. Srivastava, J. Gandhi, D. Parekh, M. Gala,
most accurate and highly advanced real-time data usage tend “Obfuscation: maze of code”, Proceedings of the 2nd
to come up with best forecasts in the medium and long term. [18] International Conference on Communication Systems,
Computing and IT Applications (CSCITA), IEEE (2017),
pp.11-16

45
Machine Translated by Google

International Journal of Computer Applications (0975 – 8887)

Volume 178 – No. 8, May 2019

[6] Dijkstra's Algorithm pseudocode (2019, January 22). Cincinnati“GPS: Location-Tracking Technology” (2019, March
Programiz[Online].Available: 2).
https://www.programiz.com/dsa/dijkstra-algorithm [Online].Available:https://pdfs.semanticscholar.org/e52d /
b090bf49541473de29ca164647578aa70c9f.pdf?_ga=2.1
[7] Difference and advantages between Djiksta and A star
[Duplicate] (2019, January 27).Stack Overflow. 91753129.284091150.1552310180-1625978761.1550899332
[Online].Available:https://stackoverflow.com/questions/1
3031462/difference-and-advantages-between-dijkstra-a [13] Patrick Bertagna, GTX Corp., 10.26.10, “How does a GPS
star tracking system work?” – EE|Times (2019, February 22).

[8] Introduction to A* - From Amit's thoughts on path finding (2019,

[Online].Available:https://www.eetimes.com/document.a sp?
January 29). Theory.stanford.edu. [Online Available:http://
doc_id=1278363#
theory.stanford.edu/~amitp/GameProgra mming/
AStarComparison.html [14] What is Trilateration? (2019, February 27).
[Online].Available:https://www.roseindia.net/technology/gps/
[9] How does Djikstra and A-star compare (2019, February)
what-is-trilateration.shtml
StackOverflow[Online]Available:https://stackoverflow.c om/
questions/1332466/how-does-dijkstras-algorithm and-a- [15] What is WAAS? (2019, February 27).
star-compare [Online].Available:https: //www.roseindia.net/technology/gps/
what-is-waas.shtml
[10] Seamless Google Street View Panoramas - Posted by Mike
Krainin, Software Engineer and Ce Liu, Research Scientist, [16] Kay Ireland, “How Does Google Map Calculate Travel
Machine Perception. (2019, February 13)GOOGLE BLOG. Time” (2019, March 4). [Online].
WHO Available:https://www.techwalla.com/articles/how-does
[Online].Available:https://ai.googleblog.com/2017/11/se google-maps-calculate-travel-time
amless-google-street-view-panoramas.html
[17] Zach Epstein, “Former Google Engineer explains how Google
[11] Google's New Street View Cameras Will HelpAlgorithms Maps determines your ETA” – BGR.com. (2019, March 9).
Index The Real World – Tom Simonite(Business) 2019,
February 18). Wired.com. [Online].Available:https://bgr.com/2013/12/27/google
[Online].Available:https://www.wired.com/story/googles -new- maps-eta-calculation-explanation/
street view-cameras-will-help-algorithms-index the-real-world/
[18] Techcoffee – Google Maps Calculate ETA. (2019, March11)
[Online].Available:https://www.techcoffees.co m/google-
[12] Rashmi Bajaj, France Telecom R&DSamanthaLalinda maps-calculate-estimated-time/
Ranaweera, LSI LogicDharma P. Agrawal, University of

IJCATM : www.ijcaonline.org forty six

View publication stats