Heterogeneous Multiprotocol Vehicle Controls Systems in Cloud

Computing Environment

Vladimir S. Zaborovski

, Mikhail Chuvatov

, Oleg Gusikhin

, Abdulaziz Makkiya

and David Hatton

Telematics Department, St. Petersburg State Polytechnic University, St. Petersburg, Russia

Research and Advanced Engineering, Ford Motor Company, Dearborn, Michigan 48121, U.S.A.

Connected Services Systems Organization, Ford Motor Company, Dearborn, Michigan 48121, U.S.A.

Keywords: Cloud Technology, Roads Services, Heterogeneous Multiprotocol Cloud Computing Environment, Mesh

Network, Wi-Fi (802.11), LTE Technology, and DSRC.

Abstract: The paper introduces a heterogeneous multiprotocol cloud computing environment for vehicle monitoring

and applicable services. It also highlights the concept behind virtual vehicle communication with multi-

protocol data access. Cloud based vehicle servicers require constant access to the Internet, which can be

organized using a combination of local (Wi-Fi, DSRC) and the wireless communication technologies. The

paper discusses the advantages of multi-protocol communication with the possibilities of alternative ways of

delivering messages between vehicle and a cloud. The paper presents simulation of different scenarios with

respect to the technology used.

1 INTRODUCTION

With the emergence of cloud computing many

existing vehicular applications may be enhanced

along new functional dimensions, including online

access to high performance computing algorithms

and practically unlimited data storage resources.

Cloud services are now being considered as a source

of various innovations in the information and

management spheres which can further empower

existing Vehicle Controls Systems (VCS). These

new services would need to be in accordance with

performance and security requirements of future

intelligent control systems. Expanding the range of

these services leads to an increase in the volume

and/or value of data that flows between vehicles and

creates new content for automotive protocols.

Objects and subjects of cloud services can be

divided into three main categories: 1)

communication infrastructures which support active

safety features in avoiding accidents and traffic

congestions 2) elastic access systems shared with

high performance data processing, monitoring and

channel integration services 3) ubiquitous storage

systems to store data sessions and archive traffic and

vehicle information. The applications of such cloud

services can be utilized by various industries and

parties including automakers' analytic centres,

drivers, passengers, emergency departments, vehicle

owners, regional traffic centres, dealerships, and

road services. Operative interaction between users

and information resources forms scalable

frameworks for:

 Monitoring the technical condition of vehicles;

 Traffic control and emergency assistance

 Insurance settlements with customers;

 In-vehicle entertainment;

 Solutions for intelligent transportation systems.

In the existing VCS, most operations use

resources of embedded computing. An alternative

approach is to perform some non-safety critical

operations remotely while exchanging information

with the vehicle using on-board wireless

communications appliances. Although the idea of

transferring some of the computational tasks to a

remote server is straightforward (see figure 1); its

technical implementation is challenging and so far

has been focused mainly on infotainment

applications. In this paper we propose a multi-

communication protocol solution that can facilitate

intelligent computation of the data exchanged

between the vehicle and the infrastructure and vice

versa. This solution shall eliminate some of the

current existing complexities accompanied by the

single protocol approach.

555

S. Zaborovski V., Chuvatov M., Gusikhin O., Makkiya A. and Hatton D. (2013).

Heterogeneous Multiprotocol Vehicle Controls Systems in Cloud Computing Environment.

In Proceedings of the 10th International Conference on Informatics in Control, Automation and Robotics, pages 555-561

DOI: 10.5220/0004621505550561

 SciTePress



Figure 1: Basic cloud environment running local and remote service.

Each vehicle is treated as a source of information

to the cloud environment and to other vehicles

simultaneously. The vehicle control system itself

shall have a non-zero probability to crash or fail.

However, the flow of data from vehicle via wireless

mediums is susceptible to interference that can

disrupt connectivity and interaction of network

nodes and service components. In this case, the

reasonable metric of effectiveness of the tasks listed

above is the probability of delivering the message in

a given time interval from the source to a particular

destination within the shared cloud infrastructure.

Delivery probability is a controlled parameter which

is a function of the cloud resources needed to

process requests for services. For example, fault

tolerant implementation of supervisory tasks, such as

route planning or vehicle speed set-point can be

achieved with requested level of probability relying

on the distributed cloud computing power.

Furthermore, service agents which interact with

vehicle applications and run on a remote virtual

machine (VM) can improve or even optimize vehicle

performance by analyzing historical driving patterns

along the same route over many different conditions.

Cloud based approaches allow the extension of the

existing vehicle on-board resources by taking

advantage of collaborative services. However, the

probabilistic nature as well as latency in delivering

urgent information can delay the delivery of safety

critical applications running over shared cloud

resources. Today, not all vehicles are equipped with

wireless access appliances and the wireless network

coverage is not optimum for various wireless

technologies. To overcome this obstacle we

introduce the concept of a heterogeneous

multiprotocol cloud computing environment which

integrates different architectures of MESH networks,

DSRC, and LTE. Adaptive capabilities of such

infrastructure are formed by the use of resources and

protocols that under specific circumstances can

increase the probability of timely delivery of

information necessary for control purposes using

different communication media, for example,

embedded vehicle’s controllers equipped with Wi-

Fi, and LTE interfaces.

ICINCO2013-10thInternationalConferenceonInformaticsinControl,AutomationandRobotics

556

2 HETEROGENEOUS CLOUD

COMPUTING CONCEPT

The use of different communication mediums and

protocols which deliver data from the vehicle to the

cloud services and vice versa can improve traffic

safety and route planning only if we can expand the

functionality of embedded computing systems by

using the concept of multi-protocol access gateway

between vehicle and distributed network resources.

In this case each vehicle in the cloud environment

can be viewed as virtual agent (Gusikhin, 2011).

This virtual agent, that we will call virtual vehicle

(VV) encapsulates an information trail on the data

being processed by a vehicle. VV is an active agent

of information relations which holds information

about the current resources and local properties of

the vehicle environment. In heterogeneous cloud

environments VV may play a role of active data

source which contains essential static (do not

changes during control operations) and dynamic

parameters (environment conditions, GPS

coordinates, supervisory commands, fuel

consumption, or control signals, etc), information

about vehicle motion and/or driver behavior

(Zaborovsky, 2011) (Mulukha, 2011). VV is a

logical node in MESH segments and a virtual

resource on IaaS level which combines actual

vehicle’ data with wealth of information available

from computational and storage resources. The

programmable and multiservice nature of VV

provides a wide variety of opportunities to realize

intelligent control algorithms and merge together

different models of vehicle motion with global data

and goal oriented forecasts in which VV is

represented as an intelligent agent of heterogeneous

the cloud/MESH environment. VV may be viewed

as a subject of appropriate control decisions which

are based on ubiquitous high performance cloud

computing platform (“hpcc-platform”) and storage

resources that are available as a service. Once a

vehicle has been driven on the road, its VV image

becomes a part of cloud-based services which

monitors the vehicle during all control sessions.

There are several advantages behind such a

heterogeneous architecture. High-computational and

communication power associated with the cloud

environment is a major source that can significantly

extend the operational and optimization capabilities

of a new generation of VCS. Cloud-centric VCS

form composite decisions based on the fusion of

different factors including vehicle dynamics,

specific constraints, on-line information from road

sensors about surrounding moving and non-moving

objects. All VCS operations such as routing or

vehicle parameters optimization may be done at

different performance rates. This rate can be

accelerated by reconfigurable (Msadaa, 2010)

available virtual and computing resources in

accordance with vehicle speed.

The concept of virtual machine in the

heterogeneous environment of the cloud

architecture, that we are proposing, provides the

implementation of fault-tolerant, powerful and

flexible tool for managing the traffic services, roads

infrastructure, and vehicle data. The implementation

of such concept which is based on the organization

of access to cloud services is executed through the

currently existing wireless and hard-wire

connections of the various technologies. Those

connections can be placed on vehicles, on objects of

road’s infrastructure and services providers. That

leads to simultaneous use of communication

channels of various technologies for improved

accessibility to cloud services which requires

integration to the shared wireless multiprotocol

network of data communication.

3 MULTIPROTOCOL

COMMUNICATIONS

To fully leverage cloud-based infrastructure in

providing different vehicle services, especially

within mission critical tasks we need to ensure a

fault tolerant communication infrastructure with a

given probability of message delivery within a given

time interval. The lack of a stable 2G/3G/4G

coverage in the highways and transportation

infrastructure along with the impracticality to equip

each vehicle with an access device to the local and

global networks leads to the need of developing

alternative methods of data communication with the

cloud.

The model proposed in this paper is centred

around using existing industry standards represented

by mesh networks, Wi-Fi (802.11), DSRC and LTE

(

Yoon, 2010) (Navarro, 2007) technology. However, in

practice, the implementation of this approach faces a

number of challenges:

 Incompatibility of network equipment;

 High costs of implementation of new

technologies;

 Difficulty to ensuring reliability in case of high

message rate.

The first problem can be solved by using

network access devices with reconfigurable multi-

HeterogeneousMultiprotocolVehicleControlsSystemsinCloudComputingEnvironment

557

frequency radio interfaces that are at the software

level compatible with the protocols of global and

local wireless networks.

To support processing of information

interactions mentioned above requires designing a

new generation of multi-protocol routers for DSRC,

LTE, MESH, WI-FI networks. Such routers should

synthesize optimal paths taking into account the

nominal and available bandwidth as well as ongoing

delays of delivering data packets or control

information. These routers should have a well-

defined relationship between: 1) data rate and the

available virtual channel bandwidth, 2) routing

policy and the current network topology, and 3)

routing algorithms and the data characteristics type

the characteristics of the data link layer protocols.

The basis of these routers are the multi-core multi-

threaded network processors whose functionality are

extended through the use of FPGA co-processors

which provide high-performance internal operations

for multi-band processing of the incoming traffic.

The solution for the second problem can be done

by using the method of "seamless" roaming between

Wi-Fi, 2G/3G/4G, which are based on technology

Hotpsot 2.0 (Next Generation Hotspot - NGH).

NGH technology provides an authentication

procedure and handover, as well as support for

automatic logon of mobile devices to the network

without the need for re-authorization (Heller, 2006)

(Andreev, 2010).

The effect of the third problem can be mitigated

by using recommendations from 802.11 x protocol

suite that allows creating Wi-Fi network that

supports alternative data routes. Development of

mesh Wi-Fi infrastructure based on the 802.1x

specification provides self-organization of networks

and increases the ability of networking recovery in

case of failure of the switching nodes.

IEEE 802.11v specification provides support

mechanisms for controlling the radio parameters in

order to reduce energy consumption which is

important to support self-contained modules that are

unable to regularly recharge (for example, in the

vehicle). The promising solution that can improve

communication sessions for short-range areas is to

use 60 GHz frequency band with peak data rates up

to 7 Gbit/s in accordance with the recommendations

of Wireless Gigabit Alliance (Arbabi, 2010).

A common solution to these problems is to create

a multi-protocol access point (MPAP) which helps

to raise the reliability of message transmission in

areas with low density of network stations or

difficult terrain. From the technical point of view,

MPAP is a modular hardware and software system

that supports concurrent messaging technology and

robust routing algorithms. Functional requirements

of the MPAP are the follows:

1. Connecting to the vehicle.

2. Connecting vehicle to the wireless data

networks.

3. Relaying messages between networks of

different technologies.

4. Sustained interaction with the objects of road

infrastructure.

5. Effective interaction with providers of data

networks.

Figure 2 shows the block diagram of MPAP.

Currently, stationary transmitters of road

infrastructure and vehicle messages regarding road

conditions, which inform drivers and embedded

controllers about the state of traffic lights, signs,

specific facts of reducing distances and intervals

between vehicles, or the presence of congestion and

accidents on the road.

DSRC is an optimum solution to send short

messages between vehicles and emergency

infrastructures in a very short period of time. The

advantage of technology is the short time connection

between the stationary and mobile transmitters, high

speed data transfer and the maintenance of a stable

connection with vehicles moving at high speeds.

The use of DSRC technology in the network of

stationary and mobile devices allows creating

effective message delivery system for cloud

services. In this case, the DSRC technology is

implemented at the transport layer and controls

delivery of high-priority messages that are generated

as a result of the integrated emergency call system.

These messages transmitted not only via

dedicated channel associated with the specific

emergency network but also using LTE or 3G

wireless networks (Gramaglia, 2011). Simultaneous

and parallel transmission of emergency messages

using different protocols and two independent

network infrastructures can significantly reduce the

time and increase the likelihood of message delivery

to the specified address.

4 SIMULATION EXPERIMENT

A study of the possibility of delivery of emergency

messages from the vehicle to vehicle with an output

interface to the cloud computing environment by

means of LTE channels was also conducted in this

paper.



ICINCO2013-10thInternationalConferenceonInformaticsinControl,AutomationandRobotics

558

Figure 2: The block diagram of MPAP.

The first experiment consists of fragmenting an

800 meters road that contains 4 turns at 90 degrees.

The road has 4 lanes broken down by 2 on each side,

on the fragment of the road there are always 8 or 16

vehicles (1 or 2 vehicle for every 100 meters)

equipped with 802.11s (number of vehicles leaving

the fragment is equal to the number of road vehicles

that fall in this fragment of the road) - one half of the

total number of vehicles are moving in one direction

and the other half of the vehicles are moving in the

opposite direction. Traffic is moving according to

the IDM model with the following parameters:

maximum speed of 10 to 100 miles/hr - varied in

increments of 10 miles/h (parameter of the

experiment), the distance between any two vehicles

is at least 4 m, the acceleration of 5 m/s

. One

vehicle is equipped with a dedicated LTE modem.

The simulation also included one vehicle that

transmits emergency messages to the infrastructure

and the other vehicles via the mesh network. The

experiment evaluated the percentage of delivered

emergency messages.

Figure 3. shows a graphical representation for the

results obtained from simulating the mesh network

under different conditions. The results of the

experiments with MESH network using NS3

simulator (Yoon, 2010) (Heller, 2006) which is

based on 802.11s shows clear relationship between

vehicle speed and the distance between vehicles and

the total loss rate (Fig. 3).

Analysis of the data presented in Figure 4 shows that

at speed of 100 miles per hour and traffic density of

1 or 2 vehicles on a road segment of length of 100

meters a loss percentage of 78-92%. It can be

concluded that in such condition of movement it is

necessary to increase the density of vehicles that are

equipped with on-board 802.11s transmitter or to

create alternative ways of delivering emergency

messages.

In the second experiment we consider the same

simulation parameters: four-lane road and traffic

model with 16 vehicles available on the road.

HeterogeneousMultiprotocolVehicleControlsSystemsinCloudComputingEnvironment

559

Figure 3: The dependence of the loss rate to deliver a

message from the emergency vehicle.

Among these vehicles there are vehicles equipped

with output interface to the cloud by means of LTE.

The percentage of vehicles equipped with LTE

modules is 3-25% (the parameter of the experiment).

In case of accident the vehicle initiates the

transmission of emergency messages addressed to:

any vehicle equipped with output interface to the

cloud, using transport mesh (802.11s). The

experiment evaluated the percentage of delivered

emergency messages.

Figure 4: The dependence of the messages delivered from

vehicle speed and their concentration.

Figure 4 shows the percentage of messages loss

at a rate of 2 vehicles a 100 meter long road when

the vehicles equipped with 802.11s transmitter and

the share of vehicles equipped with LTE modules is

the following: 25%, 12%, 6%, 3%.

Analysis of the obtained dependency leads to the

conclusion that the use of protocols for mesh

networking (802.11s) increases the reliability of

message delivery to nodes equipped with output

LTE interface to the multiservice cloud computing

environment. The number of the vehicles with such

interfaces should be at least 25%. Otherwise, more

than 50% of sent messages may be lost. It is

advisable to use a reconfigurable data transmission

routing platform which is capable of finding an

alternative path for messages and reach the cloud via

DSRC, Wi-Fi, or even others network infrastructure.

The third experiment illustrates the alternation

between the different protocols controlled by the

bandwidth variances. The operation of combined

mobile vehicles networks equipped with

multiprotocol device with two interfaces: LTE and

802.11p. In case of low LTE bandwidth usage, the

majority of messages are transmitting over LTE.

While when the LTE network is overloaded, the

high-priority emergency short messages are

transmitted over 802.11p channel as an alternative

way. This situation shows how a multiprotocol node

increases the ability of transferring messages using

802.11p channels at moments of peak loads on LTE

network.

This experiment consists of fragmenting an 800

meters road that contains 4 turns at 90 degrees. The

road has 4 lanes broken down by 2 on each side, on

the fragment of the road there are always 16 vehicles

(8 vehicles on each direction). Each vehicle is

equipped with LTE and 802.11p interfaces. On the

roadside there is 802.11p-receiver working as a

repeater. Main goal of experiment is to check the

ability to transmit short (1 KByte) emergency

messages from moving vehicles through 802.11p

transceiver to 802.11p access point. Transmission

speed of every node is changing from 8 Kbit/s to

256 Kbit/s. 802.11p network load is measured using

TCP and UDP transport layer protocols (figure 5).

The dependency shows that network utilization

reaches critical value at 256 Kbit/s transfer speed,

wherein UDP protocol shows 25 percent less

network utilization compared to TCP. It is worth

Figure 5: Dependency between 802.11p network load and

messages transmission speed.

ICINCO2013-10thInternationalConferenceonInformaticsinControl,AutomationandRobotics

560

noting that the amount of undelivered messages

reaches 40 percent at 256 kbps transfer speed (based

on transport layer protocols statistics data of this

experiment). At transfer speeds of 128 Kbit/s or less

the amount of undelivered messages is insignificant;

messages are delivered up to 40 percent faster (in

average) over UDP compared to TCP.

Thus the multiprotocol interface improves

network characteristics in case of overloads at main

data transfer channels and in case of critical network

operation modes.

5 CONCLUSIONS

Current trends in information services establish new

challenges for developers of network protocols and

equipment. Promising areas include:

 the development of universal reconfigurable

wireless devices that adapt to the specifications

of multi-protocol network nodes to ensure

compatibility standards for the transmission of

data in local and wide area networks;

 the use of multi-protocol services, grouped

under an hpcc-platform that provides rapid

reconfiguration of virtual computing resources

with a goal to increasing the productivity of

cloud applications and the probability of delivery

of urgent messages;

 development of methods to increase the

probability of emergency message delivery in

areas with unstable coverage zones. Duplication

of priority traffic via public wireless networks;

 new 802.11x standard which can significantly

reduce the costs of mobile network development

while providing high-speed connectivity of

mobile devices over short distances and secure

access to multimedia services.

 the use of alternative channels of

communication such as DSRC, leads to the

increased quality of communication at the

moment of message transmission when other

channels such as LTE and 802.11s are

unavailable. It follows that along the road

transceivers can be located providing a stable

zone of coverage.

ACKNOWLEDGEMENTS

This research was supported by a grant from the

Ford Motor Company.

REFERENCES

Gusikhin, 2011, Oleg Gusikhin, Erica Klampfl, Dimitar

Filev and Yifan Chen. Emotive Driver Advisor

System (EDAS). In J.A. Cetto, J.-L. Ferrier, J. Filipe

(eds.), Informatics in Control Automation and

Robotics, Lecture Notes in Electrical Engineering, v.

89, 2011, 21-36.

Zaborovsky, 2011, Vladimir Zaborovsky, Alexey

Lukashin, Sergey Kupreenko, Vladimir Mulukha.

Dynamic Access Control in Cloud Services.

Proceedings of The 2011 IEEE International

Conference on Systems, Man and Cybernetics (IEEE

SMC 2011), Anchorage, Alaska, USA, October 9-12,

2011, 1400-1406

Mulukha, 2011, Vladimir.Zaborovsky, Vladimir Mulukha.

Access Control in a Form of Active Queuing

Management in Congested Network Environment.

Proceedings of the Tenth International Conference on

Networks (ICN 2011; GlobeNet 2011), January 23-28,

2011 - St. Maarten, the Netherlands Antilles.

Published by IARIA, 2011, 12-17.

Msadaa, 2010, Chammakhi Msadaa, Ikbal; Cataldi,

Pasquale; Filali, Fethi. A Comparative Study between

802.11p and Mobile WiMAX-based V2I

Communication Networks. Next Generation Mobile

Applications, Services and Technologies (NGMAST),

2010 Fourth International Conference. 27-29 July

2010. Amman, 186 - 191.

Navarro, 2007, N. Coelho, N. Cabral, A. Pereira, A. Rocha

and A. Navarro. Mobile WiMAX Assessment in Sub-

urban Area to Support TV Broadcasting. IEEE ISCE

2007, 2007, Dallas, USA, 12-19.

Yoon, 2010, Hayoung Yoon. Collaborative streaming-

based media content sharing in WiFi-enabled home

networks. Consumer Electronics, IEEE Transactions

on (Volume:56, Issue: 4 ), 2193-2200.

Heller, 2006, Brandon Heller, The Quest for Wireless

Gigabit: Recent Advances in the Wireless Physical

Layer. [online] Available at: <http://www.cse.

wustl.edu/~jain/cse574-06/ftp/phy_trends.pdf>

[Accessed 15 March 2013].

Andreev, 2010, Kirill Andreev, Pavel Boyko. IEEE

802.11s Mesh Networking NS-3 Model.[online].

Available at: http://www.nsnam.org/workshops/wns3-

2010/dot11s.pdf [Accessed 10 March 2013].

Arbabi 2010, Highway mobility and vehicular ad-hoc

networks in ns-3. In Proceedings of the Winter

Simulation Conference (WSC '10), Björn Johansson,

Sanjay Jain, and Jairo Montoya-Torres (Eds.). Winter

Simulation Conference 2010, 2991-3003.

Gramaglia, 2011, Marco Gramaglia, Carlos J Bernardos

and Maria Calderon, Seamless Internet 3G and

Opportunistic WLAN Vehicular Connectivity.

[Online]. Available at: Http://Link.Springer.Com/

Content/Pdf/10.1186%2F1687-1499-2011-183.Pdf

[Accessed 10 March 2013].

HeterogeneousMultiprotocolVehicleControlsSystemsinCloudComputingEnvironment

561