Mahgoub, Imad

According to a March 2019 publication by the National Highway Transportation Safety Administration(NHTSA), 62% of all police-reported accidents in the United States between 2011 and 2015 could have been prevented or mitigated with the use of five groups of collision avoidance technologies in passenger vehicles: (1) forward collision prevention, (2) lane keeping, (3) blind zone detection, (4) forward pedestrian impact, and (5) backing collision avoidance. These technologies work mostly by reducing or removing the risks involved in a lane change maneuver; yet, the Broward transportation management system does not directly address these risk. Therefore, we are proposing a Machine Learning based approach to real-time accident prediction for Broward I-95 using the C5.1 Decision Tree and the Multi-Layer Perceptron Neural Network to address them. To do this, we design a new measure of volatility, Lane Change Volatility(LCV), which measures the potential for a lane change in a segment of the highway. Our research found that LCV is an important predictor of accidents in an exit zone and when considered in tandem with current system variable, such as lighting conditions, the machine learning classifiers are able to predict accidents in the exit zone with an accuracy rate of over 98%.

Vehicular Ad hoc Networks (VANET) is a wireless ad-hoc network that includes
two types of communications, Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure
(V2I). In VANET there are two types of messages. The first type is the event-driven
messages that are only triggered in case of emergency. The second type is the periodical
messages named beacons that are exchanged frequently between vehicles. A
beacon message contains basic information about the sending vehicle such as id, location
and velocity. Beacons are frequently exchanged to increase the cooperative
awareness between vehicles. Increasing beacon frequency helps increasing neighborhood
awareness and improving information accuracy. However, this causes more
congestion in the network, specially when the number of vehicles increases. On the
other hand, reducing beacon frequency alleviates network congestion, but results in
out-dated information.
In this dissertation, we address the aforementioned challenges and propose a
number of smart beaconing protocols and evaluate their performance in di↵erent environments
and network densities. The four adaptive beaconing protocols are designed
to increase the cooperative awareness and information freshness, while alleviating the network congestion. All the proposed protocols take into account the most important
aspects, which are critical to beaconing rate adaptation. These aspects include channel
status, traffic conditions and link quality. The proposed protocols employ fuzzy
logic-based techniques to determine the congestion rank, which is used to adjust beacon
frequency.
The first protocol considers signal to interference-noise ratio (SINR), number
of neighboring nodes and mobility to determine the congestion rank and adjust the
beacon rate accordingly. This protocol works well in sparse conditions and highway
environments. The second protocol works well in sparse conditions and urban environments.
It uses channel busy time (CBT), mobility and packet delivery ratio
(PDR) to determine the congestion rank and adjust the beacon rate. The third protocol
utilizes CBT, SINR, PDR, number of neighbors and mobility as inputs for the
fuzzy logic system to determine the congestion rank and adjust the beacon rate. This
protocol works well in dense conditions in both highway and urban environments.
Through extensive simulation experiments, we established that certain input
parameters are more e↵ective in beacon rate adaptation for certain environments
and conditions. Based on this, we propose a high awareness and channel efficient
scheme that adapts to di↵erent environments and conditions. First, the protocol
estimates the network density using adaptive threshold function. Then, it looks at
the spatial distribution of nodes using the quadrat method to determine whether
the environment is highway or urban. Based on the density conditions and nodes
distribution, the protocol utilizes the appropriate fuzzy input parameters to adapt
the beaconing rate. In addition, the protocol optimizes the performance by adapting
the transmission power based on network density and nodes distribution.
Finally, an investigation of the impact of adaptive beaconing on broadcasting
is conducted. The simulation results confirm that our adaptive beaconing scheme
can improve performance of the broadcast protocols in terms of reachability and bandwidth consumption when compared to a fixed rate scheme.

The United States has been going through a road accident crisis for many
years. The National Safety Council estimates 40,000 people were killed and 4.57
million injured on U.S. roads in 2017. Direct and indirect loss from tra c congestion
only is more than $140 billion every year. Vehicular Ad-hoc Networks (VANETs) are
envisioned as the future of Intelligent Transportation Systems (ITSs). They have a
great potential to enable all kinds of applications that will enhance road safety and
transportation efficiency. In this dissertation, we have aggregated seven years of real-life tra c and
incidents data, obtained from the Florida Department of Transportation District 4.
We have studied and investigated the causes of road incidents by applying machine
learning approaches to this aggregated big dataset. A scalable, reliable, and automatic
system for predicting road incidents is an integral part of any e ective ITS. For this
purpose, we propose a cloud-based system for VANET that aims at preventing or at
least decreasing tra c congestions as well as crashes in real-time. We have created,
tested, and validated a VANET traffic dataset by applying the connected vehicle
behavioral changes to our aggregated dataset. To achieve the scalability, speed, and fault-tolerance in our developed system, we built our system in a lambda architecture
fashion using Apache Spark and Spark Streaming with Kafka.
We used our system in creating optimal and safe trajectories for autonomous
vehicles based on the user preferences. We extended the use of our developed system in
predicting the clearance time on the highway in real-time, as an important component
of the traffic incident management system. We implemented the time series analysis
and forecasting in our real-time system as a component for predicting traffic
flow.
Our system can be applied to use dedicated short communication (DSRC), cellular,
or hybrid communication schema to receive streaming data and send back the safety
messages.
The performance of the proposed system has been extensively tested on the
FAUs High Performance Computing Cluster (HPCC), as well as on a single node
virtual machine. Results and findings confirm the applicability of the proposed system
in predicting traffic incidents with low processing latency.

Vehicular Ad hoc Networks (VANETs) have the potential to enable various
kinds of applications aiming at improving road safety and transportation efficiency.
These applications require uni-cast routing, which remains a significant challenge due
to VANETs characteristics. Given VANET dynamic topology, geographic routing
protocols are considered the most suitable for such network due to their scalability
and low overhead. However, the optimal selection of next-hop nodes in geographic
routing is a challenging problem where the routing performance is highly affected by
the variable link quality and bandwidth availability.
In this dissertation, a number of enhancements to improve geographic routing
reliability in VANETs are proposed. To minimize packet losses, the direction and
link quality of next-hop nodes using the Expected Transmission Count (ETX) are
considered to select links with low loss ratios.
To consider the available bandwidth, a cross-layer enchantment of geographic
routing, which can select more reliable links and quickly react to varying nodes load
and channel conditions, is proposed. We present a novel model of the dynamic behavior of a wireless link. It considers the loss ratio on a link, in addition to transmission
and queuing delays, and it takes into account the physical interference e ect on the
link.
Then, a novel geographic routing protocol based on fuzzy logic systems, which
help in coordinating di erent contradicting metrics, is proposed. Multiple metrics
related to vehicles' position, direction, link quality and achievable throughput are
combined using fuzzy rules in order to select the more reliable next-hop nodes for
packet forwarding.
Finally, we propose a novel link utility aware geographic routing protocol,
which extends the local view of the network topology using two-hop neighbor information.
We present our model of link utility, which measures the usefulness of a
two-hop neighbor link by considering its minimum residual bandwidth and packet
loss rate. The proposed protocol can react appropriately to increased network tra c
and to frequent topology dis-connectivity in VANETs.
To evaluate the performance of the proposed protocols, extensive simulation
experiments are performed using network and urban mobility simulation tools. Results
confirm the advantages of the proposed schemes in increased traffic loads and
network density.

A cross-layer design architecture featuring a new network
stack component called a controller is presented. The
controller takes system status information from the protocol
components and uses it to tune the behavior of the network
stack to a given performance objective. A controller design
strategy using a machine learning algorithm and a simulator
is proposed, implemented, and tested. Results show the
architecture and design strategy are capable of producing a
network stack that outperforms the existing protocol stack for
arbitrary performance objectives. The techniques presented
give network designers the flexibility to easily tune the
performance of their networks to suit their application. This
cognitive networking architecture has great potential for high
performance in future wireless networks.

This research proposes a cluster-based target tracking strategy for one
moving object using wireless sensor networks. The sensor field is organized in 3
hierarchal levels. 1-bit message is sent when a node detects the target.
Otherwise the node stays silent. Since in wireless sensor network nodes have
limited computational resources, limited storage resources, and limited battery,
the code for predicting the target position should be simple, and fast to execute.
The algorithm proposed in this research is simple, fast, and utilizes all available
detection data for estimating the location of the target while conserving energy.
lbis has the potential of increasing the network life time.
A simulation program is developed to study the impact of the field size
and density on the overall performance of the strategy. Simulation results show
that the strategy saves energy while estimating the location of the target with an
acceptable error margin.

With the issuance of the Notice of Proposed Rule Making (NPRM) for Vehicle
to Vehicle (V2V) communications by the United States National Highway Tra c
Safety Administration (NHTSA), the goal of the widespread deployment of vehicular
networking has taken a signi cant step towards becoming a reality. In order for
consumers to accept the technology, it is expected that reasonable mechanisms will
be in place to protect their privacy. Cooperative Caching has been proposed as an
approach that can be used to improve privacy by distributing data items throughout
the mobile network as they are requested. With this approach, vehicles rst attempt
to retrieve data items from the mobile network, alleviating the need to send all requests
to a centralized location that may be vulnerable to an attack. However, with
this approach, a requesting vehicle may expose itself to many unknown vehicles as
part of the cache discovery process.
In this work we present a Public Key Infrastructure (PKI) based Cooperative
Caching system that utilizes a genetic algorithm to selectively choose members of the
mobile network to query for data items with a focus on improving overall privacy. The
privacy improvement is achieved by avoiding those members that present a greater risk of exposing information related to the request and choosing members that have a
greater potential of having the needed data item. An Agent Based Model is utilized
to baseline the privacy concerns when using a broadcast based approach to cache
discovery. In addition, an epidemiology inspired mathematical model is presented to
illustrate the impact of reducing the number of vehicles queried during cache discovery.
Periodic reports from neighboring vehicles are used by the genetic algorithm to
identify which neighbors should be queried during cache discovery. In order for the
system to be realistic, vehicles must trust the information in these reports. A PKI
based approach used to evaluate the trustworthiness of each vehicle in the system is
also detailed. We have conducted an in-depth performance study of our system that
demonstrates a signi cant reduction in the overall risk of exposure when compared
to broadcasting the request to all neighbors.

Multi-hop broadcast is one of the main approaches to disseminate data in
VANET. Therefore, it is important to design a reliable multi-hop broadcast protocol,
which satis es both reachability and bandwidth consumption requirements.
In a dense network, where vehicles are very close to each other, the number of
vehicles needed to rebroadcast the message should be small enough to avoid a broad-
cast storm, but large enough to meet the reachability requirement. If the network
is sparse, a higher number of vehicles is needed to retransmit to provide a higher
reachability level. So, it is obvious that there is a tradeo between reachability and
bandwidth consumption.
In this work, considering the above mentioned challenges, we design a number
of smart broadcast protocols and evaluate their performance in various network den-
sity scenarios. We use fuzzy logic technique to determine the quali cation of vehicles
to be forwarders, resulting in reachability enhancement. Then we design a band-
width e cient fuzzy logic-assisted broadcast protocol which aggressively suppresses
the number of retransmissions. We also propose an intelligent hybrid protocol adapts
to local network density. In order to avoid packet collisions and enhance reachability, we design a cross layer statistical broadcast protocol, in which the contention window
size is adjusted based on the local density information.
We look into the multi-hop broadcast problem with an environment based
on game theory. In this scenario, vehicles are players and their strategy is either
to volunteer and rebroadcast the received message or defect and wait for others to
rebroadcast. We introduce a volunteer dilemma game inspired broadcast scheme to
estimate the probability of forwarding for the set of potential forwarding vehicles. In
this scheme we also introduce a fuzzy logic-based contention window size adjustment
system.
Finally, based on the estimated spatial distribution of vehicles, we design a
transmission range adaptive scheme with a fuzzy logic-assisted contention window
size system, in which a bloom lter method is used to mitigate overhead.
Extensive experimental work is obtained using simulation tools to evaluate the
performance of the proposed schemes. The results con rm the relative advantages of
the proposed protocols for di erent density scenarios.

A Vehicular Ad-hoc Network (VANET) is a wireless ad-hoc network that
provides communications among vehicles with on-board units and between vehicles
and nearby roadside units. The success of a VANET relies on the ability of a
routing protocol to ful ll the throughput and delivery requirements of any applications
operating on the network. Currently, most of the proposed VANET routing protocols
focus on urban or highway environments. This dissertation addresses the need for an
adaptive routing protocol in VANETs which is able to tolerate low and high-density
network tra c with little throughput and delay variation.
This dissertation proposes three Geographic Ad-hoc On-Demand Distance
Vector (GEOADV) protocols. These three GEOADV routing protocols are designed
to address the lack of
exibility and adaptability in current VANET routing protocols.
The rst protocol, GEOADV, is a hybrid geographic routing protocol. The second
protocol, GEOADV-P, enhances GEOADV by introducing predictive features. The
third protocol, GEOADV-PF improves optimal route selection by utilizing fuzzy logic
in addition to GEOADV-P's predictive capabilities.
To prove that GEOADV and GEOADV-P are adaptive their performance is demonstrated by both urban and highway simulations. When compared to existing
routing protocols, GEOADV and GEOADV-P lead to less average delay and a
higher average delivery ratio in various scenarios. These advantages allow GEOADV-
P to outperform other routing protocols in low-density networks and prove itself
to be an adaptive routing protocol in a VANET environment. GEOADV-PF is
introduced to improve GEOADV and GEOADV-P performance in sparser networks.
The introduction of fuzzy systems can help with the intrinsic demands for
exibility
and adaptability necessary for VANETs.
An investigation into the impact adaptive beaconing has on the GEOADV
protocol is conducted. GEOADV enhanced with an adaptive beacon method is
compared against GEOADV with three xed beacon rates. Our simulation results
show that the adaptive beaconing scheme is able to reduce routing overhead, increase
the average delivery ratio, and decrease the average delay.

In this thesis we extend the VANET-based approach to counting vehicles at a traffic
light by implementing a Geo-fence Based Vehicle Counting Algorithm which supports the
use of RFID technology. This implementation utilizes the concept of geo-fencing to create
a Zone of Interest (ZOI) that sections off a roadway that is relevant to a traffic intersection.
All vehicles in this ZOI are used to determine the required length of the green-cycle time.
By utilizing vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) technologies, we
broadcast beacons that are propagated to all vehicles in the ZOI from the infrastructure
which in this case is the traffic light controller.
These beacons are used to determine the last vehicle location in the ZOI. A timing
algorithm ensures that the last vehicle broadcasts first. The beacons are sent using the
IEEE 1609.4 Wireless Access in Vehicular Environments Standard Vendor Specific Action
(VSA) frames on the Smart Drive Initiative Vehiclular Ad Hoc Networks testbed. This
work is implemented in conjunction with the Vehicular Multi-technology Communication
Device (VMCD) supported by the National Science Foundation.