design of an infrastructure using wi-fi for the pot-hole detection system m. tech. project –...
Post on 11-Jan-2016
212 Views
Preview:
TRANSCRIPT
Design of an Infrastructure using Wi-Fi for the pot-hole detection system
M. Tech. Project – Dissertation
Shonil VijayCSE (KReSIT)
July 13, 2007
© Shonil Vijay MTP Dissertation 2
Outline Introduction
Motivation Problem Statement
Obstacle Detection Methods Off-Road Obstacle Detection Vision Based Obstacle Detection RADAR (Radio Detection And Ranging) LADAR (Laser Detection And Ranging)
Vehicles & Wi-Fi Integration Localization using Wi-Fi Vehicular Internet using Wi-Fi
Pot-Hole Detection System Our Approach
© Shonil Vijay MTP Dissertation 3
Cont… Experiments
Network Simulator – Qualnet Experimental Setup Variation of Speed & Packet Size Variation of Speed & Packet Rate Variation of Number of Mobile Nodes & Packet rate at MNs Use Case Analysis – Pothole Detection System
Conclusion & Future Work Conclusion Future Work
References
© Shonil Vijay MTP Dissertation 4
Introduction Motivation source: WHO report on road traffic injury prevention, 2004
Worldwide, around 1.2 million people are killed in road crashes every year
Around 50 million people get injured By 2020, road accidents will be the world’s third leading cause of death
Problem Statement Design of an Infrastructure, with Wi-Fi (802.11b) access points on the
road sides and vehicles equipped with on-board Wi-Fi equipments, to support applications involving vehicle to road-side communication
The system should be easily adaptable to metropolitan scales and should utilize off-the shelf or pre-deployed hardware so as to lower the cost metrics
The Pot-hole detection System should fit this architecture
© Shonil Vijay MTP Dissertation 5
Defects on roads There can be various degrees of defects on the roads:
Cracks, Pop-outs, wear-out & pot-holes
The one we are interested in are the pot-holes We can also focus some attention on cracks or pop-outs, as
they can help reduce the number of pot-holes
© Shonil Vijay MTP Dissertation 6
Steps to Obstacle Detection Requirements for such a system:
Look sufficiently far ahead of the host car to be able to discern potential obstacles
Scan the perimeter of the traveling direction, and allow for the fact that this traveling direction changes very often, both vertically and horizontally
Scan the perimeter of the traveling direction, also because objects might approach the path from the sides or from above/below, as well as being stationary and in the path of travel
Evaluate what is being seen Do all this in sufficiently good time for the control system to brake
the car before impact
© Shonil Vijay MTP Dissertation 7
Methods used for Obstacle Detection Infra-red Sensors
Near Infra-Red Sensors (NIR) 0.8~1.2μm NIR device can detect the objects regardless of the light situation even
under dark condition
Far Infra-Red region (FIR) 7~14μm FIR device can detect the presence of objects that emit heat Such devices is that they work well under various illumination
conditions devices are usually expensive
© Shonil Vijay MTP Dissertation 8
Cont… Optical Flow
Optical flow uses two or more images taken at different times to perform obstacle detection
CCD/CMOS sensors are used, and they are active in the visibility range
Optical flow methods are best at detecting moving objects which have motion fields significantly different from the background
Stereo Vision Stereo Vision uses the method of simple region matching between
images, observed from a stereo camera Conceptual difference from Optical Flow is that in one set of images
is taken with a spatial base line, and one with a temporal baseline As has been observed, spatial baseline provides better detect ability
© Shonil Vijay MTP Dissertation 9
Cont… RADAR - Radio Ranging And Detection
RADAR is an excellent means of detecting other vehicles because radar works at long ranges and is relatively unaffected by rain or snow
RADARs are not able to reliably detect small objects at ample distances
Metal surfaces are good RADAR reflectors, and hence make vehicle detection fairly easy
Other Obstacles which are not good reflectors, including defects on the roads, are not easily detected
© Shonil Vijay MTP Dissertation 10
Cont… LADAR – Laser Detection And Ranging
Laser beams are eminently suitable for the task of identifying obstacles
They keep a narrow beam over long distances Methods for measurements involve time lag between the
transmitted and reflected laser or the intensity variation Problem to detect an object that does not reflect the laser beam, eg.
Asphalt – the material used in road construction Laser Detectors as well as source can be comparatively costlier
© Shonil Vijay MTP Dissertation 11
Cont… GPS based method
The Method involves equipping the host cars with GPS navigation systems that can help them by providing information regarding their position relative to other obstacles
GPS based systems have very high operational costs as they need to obtain the precise positions of the obstacles in every few milliseconds
© Shonil Vijay MTP Dissertation 12
Vehicles & Wi-Fi Integration The growth of Wireless LANs based on Wi-Fi (802.11), clearly
predicts the pathway to a practical networking solution offering mobility, flexibility and low-cost deployment
Increased Safety, traffic management and Mobility are the main drivers towards the integration
Commercial benefits may be used as a start-up push and then consumed in maintaining the system
The stake holders include government agencies, automobile manufactures & many Associations and Universities
Research in this area include localization, providing vehicular internet using opportunistic communication
© Shonil Vijay MTP Dissertation 13
Localization using Wi-Fi Location systems have been identified as an important
component of emerging mobile applications Experiments show that we can estimate a user's position
with a median positioning error of 13-40 meters using wi-fi PlaceLab, an open source program, that enables localization
using the signal strengths received from the nearby APs Researchers have used methods including centroid, finger-
printing, Particle filters & bayesian framework for spatial location estimation using Wi-Fi signal strengths
War-Driving helps these systems by updating the data about the access points over the internet
© Shonil Vijay MTP Dissertation 14
Vehicular Internet using Wi-Fi Researches show that using off-the-shelf Wi-Fi hardware,
a vehicle could maintain a connection to a roadside access point for 600 m, and transfer 9 MB of data at 80 km/h
They found that connections pass through three phases: the entry phase, the production phase, and the exit phase, each lasting 200 m in their experiments
They postulate that existing protocols are not optimized for operation, as they suffer tremendously in the presence of high packet loss
Also the multiple round trips required for HTTP traffic reduced the total throughput by a great amount
© Shonil Vijay MTP Dissertation 15
Pot-Hole Detection System We need an infrastructure enabling the mobile nodes to get
the information of the road conditions in their vicinity The mobile nodes get this information from the road-side
access points in the form of broadcasts over UDP. This setup, helps by removing the overhead of connection
setup and hence improves utilization The mobile node’s equipment can warn the driver by using
a beep or some visual signal, if a pothole is in the vicinity The mobile nodes can also send back packets containing the
validity of the potholes’ information to the access points
© Shonil Vijay MTP Dissertation 16
Cont… The packet from the mobile nodes can help the system
update the data with the current conditions of the road The access points can also broadcast their own location co-
ordinates in the packet to enable the mobile nodes to estimate their own location
We avoid use of RTS/CTS to reduce the overhead of the protocol and in-turn increase the probability of success
© Shonil Vijay MTP Dissertation 17
Experiments Used the Qualnet network simulator provided by Scalable Network
Technologies under the Qualnet University Program Qualnet was preferred as even the simulation of large number of
nodes with heavy traffic does not reduce the efficiency Qualnet also has the support of complete Java based GUI The three main components in the Qualnet include simulator,
analyser and packet tracer In the simulator we are able to see the events as they occur at each
layer, and can control the speed of the simulation Packet Tracer can be used after the simulation to track all the
network, transport and application layer packets transferred during the simulation
© Shonil Vijay MTP Dissertation 18
Qualent Simulator
© Shonil Vijay MTP Dissertation 19
Experimental Setup The experimental setup included mobile nodes moving on a four-
lane road driving past an access point at the centre of the road The speed were varied from 30 km/h to 180 km/h to test the
feasibility of the system at various scenario configurations The number of mobile nodes indicate the density of the traffic, it
was varied from 4 (sparse) to 132 (dense) The packet sizes were also varied from 500 to 2000 bytes The packet rates were also varied from 1 packet per second to
1 packet in 10 seconds for both the AP and the MNs All other parameters, were kept fixed at the default values, so as to
mimic the behavior of off-the shelf hardware components
© Shonil Vijay MTP Dissertation 20
Variation of speed and packet size This experiment was conducted to study the effects of the
speed of mobile nodes, and the application layer packet size on the number of successful transmissions of the packets
The access point is broadcasting packets at the rate of 1 packet per second and there are 36 mobile nodes broadcasting data at the rate of 1 packet in 10 seconds
The speed is varied from 30 to 180 km/h and the packet size from 500 to 2000 bytes
© Shonil Vijay MTP Dissertation 21
Packets received by the MNs & AP
© Shonil Vijay MTP Dissertation 22
Average number of packets received by each of the mobile nodes from the access point at different speeds and the same packet sizes
0
10
20
30
40
50
60
70
80
90
30 Km/h 60 Km/h 90 Km/h 120 Km/h 150 Km/h 180 Km/hSpeed of Mobile Nodes
No.
of
Pkt
s. 500 Bytes
1000 Bytes
1500 Bytes
2000 Bytes
© Shonil Vijay MTP Dissertation 23
Average number of packets received by the access point from each of the mobile node at different speeds and the same packet sizes
0
1
2
3
4
5
30 Km/h 60 Km/h 90 Km/h 120 Km/h 150 Km/h 180 Km/hSpeed of Mobile Nodes
No.
of
Pkt
s.
500 Bytes
1000 Bytes
1500 Bytes
2000 Bytes
© Shonil Vijay MTP Dissertation 24
Average number of packets received by each of the mobile nodes from the access point at different packet sizes for the same speed of the mobile nodes
0
10
20
30
40
50
60
70
80
90
500 Bytes 1000 Bytes 1500 Bytes 2000 BytesPacket Size
No.
of P
kts.
30 Km/h
60 Km/h
90 Km/h
120 Km/h
150 Km/h
180 Km/h
© Shonil Vijay MTP Dissertation 25
Average number of packets received by the access point from each of the mobile node at different packet sizes for the same speed of the mobile node
0
1
2
3
4
5
500 Bytes 1000 Bytes 1500 Bytes 2000 BytesPacKet Size
No.
of
Pkts
.
30 Km/h
60 Km/h
90 Km/h
120 Km/h
150 Km/h
180 Km/h
© Shonil Vijay MTP Dissertation 26
Results from this experiment: The speeds of mobile nodes have a direct relationship with
the amount of time they spend in the influence of the access point and hence directly influences the number of packets that it receives
Change in the packet size has a negligible or a very small effect on the number of packets received by the mobile nodes or the access point
The uplink data suffers more loss, probably due to the crowding of the mobile nodes close to each other
percentage drop in the number of packets is more visible with either higher speeds or with slower rates
© Shonil Vijay MTP Dissertation 27
Variation of speed and packet rates This experiment was conducted to study the effects of the
speed of mobile nodes, and the packet rates on the number of successful transmissions of the packets
There are 36 mobile nodes and an access point broadcasting data packets of 500 Bytes
The speed is varied from 30 to 180 km/h The packet rates are varied as 1 packet in 1, 5 or 10 seconds
interval, individually for both the access point and the mobile nodes
© Shonil Vijay MTP Dissertation 28
Packets received by the MNs & AP
© Shonil Vijay MTP Dissertation 29
Average number of packets received by each of the mobile nodes from the access point at different speeds and the same packet rates
0
10
20
30
40
50
60
70
80
90
30 Km/h 60 Km/h 90 Km/h 120 Km/h 150 Km/h 180 Km/h
Speed of the Mobile Nodes
No.
of
Pkt
s.
AP_1 & MN_1
AP_1 & MN_5
AP_1 & MN_10
AP_5 & MN_1
AP_5 & MN_5
AP_5 & MN_10
AP_10 & MN_1
AP_10 & MN_5
AP_10 & MN_10
© Shonil Vijay MTP Dissertation 30
Average number of packets received by the access point from each of the mobile nodes at different speeds and the same packet rates
0
5
10
15
20
25
30
35
40
45
30 Km/h 60 Km/h 90 Km/h 120 Km/h 150 Km/h 180 Km/h
Speed of the Mobile Nodes
No.
of
Pkt
s.
AP_1 & MN_1
AP_1 & MN_5
AP_1 & MN_10
AP_5 & MN_1
AP_5 & MN_5
AP_5 & MN_10
AP_10 & MN_1
AP_10 & MN_5
AP_10 & MN_10
© Shonil Vijay MTP Dissertation 31
Average number of packets received by each of the mobile nodes from the access point at varying packet rate for the mobile nodes, and the same rate for the access point and fixed speeds
0
10
20
30
40
50
60
70
80
90
AP_1 & MN_1 AP_1 & MN_5 AP_1 & MN_10
Varying rate of Mobile Nodes' packet
No.
of
Pkt
s.
30 Km/h
60 Km/h
90 Km/h
120 Km/h
150 Km/h
180 Km/h
© Shonil Vijay MTP Dissertation 32
Average number of packets received by the access point from each of the mobile node at varying packet rate for the mobile nodes, and the same rate for the access point and fixed speeds
0
5
10
15
20
25
30
35
40
45
AP_1 & MN_1 AP_1 & MN_5 AP_1 & MN_10
Varying rate of Mobile nodes' packet
No.
of
Pkt
s.
30 Km/h
60 Km/h
90 Km/h
120 Km/h
150 Km/h
180 Km/h
© Shonil Vijay MTP Dissertation 33
Average number of packets received by each of the mobile nodes from the access point at varying packet rate for the access point, and the same rate for the mobile nodes and fixed speeds
0
10
20
30
40
50
60
AP_1 & MN_1 AP_5 & MN_1 AP_10 & MN_1Varying rate of Access Point' update
No.
of P
Kts
.
30 Km/h
60 Km/h
90 Km/h
120 Km/h
150 Km/h
180 Km/h
© Shonil Vijay MTP Dissertation 34
Average number of packets received by the access point from each of the mobile node at varying packet rate for the access point, and the same packet rate for the mobile nodes at fixed speeds
0
5
10
15
20
25
30
35
40
45
AP_1 & MN_1 AP_5 & MN_1 AP_10 & MN_1
Varying rate of Access Point's Update
No.
of
Pkt
s.
30 Km/h
60 Km/h
90 Km/h
120 Km/h
150 Km/h
180 Km/h
© Shonil Vijay MTP Dissertation 35
Results from this experiment: The more the packet rate the more are the number of
successful transfers For maximum downlink transfers increase the rate at the
access point and reduce it at the mobile nodes Even with low rate at the access point the crowding of the
mobile nodes reduces the uplink transfers possible The speeds show the same effect as observed in the last
experiment The change in the rate at the access point has a negligible
effect on the number of packet that it receives, due to the load of traffic from the number of mobile nodes
© Shonil Vijay MTP Dissertation 36
Variation of the number and the packet rate of the mobile nodes This experiment was conducted to study the effects of the
number, and the packet rate of the mobile nodes on the number of successful transmissions of the packets
The access point is broadcasting packets at the rate of 1 packet per second and the packet size is 500 Bytes
The mobile nodes are moving at an average speed 60km/h The setup includes the density of mobile nodes varying
from 4(sparse) to 132(dense) The mobile nodes are broadcasting data at the rate of 1
packet at an interval of 1, 5 or 10 seconds
© Shonil Vijay MTP Dissertation 37
Packets received by the MNs & AP
© Shonil Vijay MTP Dissertation 38
Average number of packets received by each of the mobile nodes from the access point at changing the number of mobile nodes and keeping the packet rates same
0
5
10
15
20
25
30
35
40
45
50
4 nodes 12 nodes 36 nodes 68 nodes 132 nodes
Number of Mobile Nodes
No.
of
Pkt
s.
AP_1 & MN_1
AP_1 & MN_5
AP_1 & MN_10
© Shonil Vijay MTP Dissertation 39
Average number of packets received by the access point from each of the mobile node at changing the number of mobile nodes and keeping the packet rates same
0
5
10
15
20
25
30
35
4 nodes 12 nodes 36 nodes 68 nodes 132 nodes
Number of Nodes
No.
of
Pkt
s.
AP_1 & MN_1
AP_1 & MN_5
AP_1 & MN_10
© Shonil Vijay MTP Dissertation 40
Average number of packets received by each of the mobile nodes from the access point at different packet rates for the mobile node keeping the number of mobile nodes fixed
0
5
10
15
20
25
30
35
40
45
50
AP_1 & MN_1 AP_1 & MN_5 AP_1 & MN_10
varying rate of Mobile Nodes' packet
No.
of
Pkt
s.
4 nodes
12 nodes
36 nodes
68 nodes
132 nodes
© Shonil Vijay MTP Dissertation 41
Average number of packets received by the access point from each of the mobile node at different packet rates for the mobile node keeping the number of mobile nodes fixed
0
5
10
15
20
25
30
35
AP_1 & MN_1 AP_1 & MN_5 AP_1 & MN_10
Varying rate of Mobile Nodes' Update
No.
of
Pkt
s.
4 nodes
12 nodes
36 nodes
68 nodes
132 nodes
© Shonil Vijay MTP Dissertation 42
Results from this experiment: The interference offered to the packet is directly related to
the number of mobile nodes and hence also affects the number of successful transmissions
With high traffic rates at the mobile nodes, the drop in the packets being received successfully is obvious
Another interesting observation can be that with lower number of mobile nodes the rate at the mobile nodes have a very negligible effect on the downlink
© Shonil Vijay MTP Dissertation 43
Use case analysis- Pothole Detection System Requirements for the system:
Each mobile node should receive at least 3 packets of information from the access point, while in its influence, for supporting redundancy
The system should also support at least one packet of data transfer, from the mobile node to the access point, with no success guarantee
It should also be able to support the same under the heaviest traffic conditions and highway speeds, with the limits to relax the downlink limit to one packet
© Shonil Vijay MTP Dissertation 44
Cont… From the second experiment it can be seen that even at high
speeds only the ones with high packet rate at the access point can satisfy the first requirement
Out of these the one with the high packet rates at the mobile nodes can not satisfy the requirement when the scenario has maximum number of nodes
Therefore we conclude that the choice should be the second one that is the downlink with the rate of 1 packet per second and the uplink with the rate of 1 packet in 5 seconds
© Shonil Vijay MTP Dissertation 45
Conclusion The result of this research is the design of the infrastructure
to support the transfer of information from the access points to the moving mobile nodes.
The access point as well as the mobile nodes can use the broadcast technique to ensure that the overheads in the transfer of information be minimized
In the use-case, the pot-hole detection system, we require the information about the road conditions to be transferred to the vehicles, for which the packet rates of 1 packet per second at the access point and 1 packet in 5 seconds at the mobile nodes seems appropriate
© Shonil Vijay MTP Dissertation 46
Conclusion & Future Work This infrastructure can be supported by the commercial
benefits, that it can provide and hence can be deployed on a metropolitan scale with ease
Future work would include, developing an application that can be used over the framework provided by the research.
This research mainly focuses on the vehicle to road-side communication, so an area that can be further explored is also to look into the inter-vehicle communication in the same manner.
© Shonil Vijay MTP Dissertation 47
References Y. Cheng, Y. Chawathe, A. LaMartha & J. Krumm “Accuracy Charac-
terization for Metropolitan-scale Wi-Fi Localization” MobiSys, 2005 D. Hadaller, S. Keshav, T. Brecht & S. Agarwal “Vehicular Oppor-
tunistic Communication under the microscope”, ACM MobiSys, 2007 J. Ott & D. Kutscher. “Drive-thru internet: IEEE802.11b for
automobile users.” IEEE Infocom, 2004 R. Gass, J. Scott, C. Diot. “Measurements of In-Motion 802.11
Networking.”, Proc. WMCSA, Apr.2006 V. Bychkovsky, B. Hull, A. K. Miu, H. Balakrishnan, & S. Madden, ”A
measurement study of vehicular Internet Access using In-Situ Wi-Fi networks”. ACM MobiCom, 2006
Methods for Obstacle Detection, http://www.swedetrack.com/obstact.htm, 2004
© Shonil Vijay MTP Dissertation 48
Thank You
top related