The following organizations and institutions are duly acknowledged for funding and supporting our research:

European Union (EU) Organizations:

Swedish Organizations:

 


 
Project: Modeling and Analysis of Random Spatial Systems for 5G Networks (EU-MARSS-5G)

Keywords: 5G Networks, Vehicular Communications, Stochastic Modeling, Road Safety
Fellowship: Marie Sklodowska-Curie Individual Fellowship (Dr. Mouhamed Abdulla)
Program/Call: H2020-MSCA-IF-2014
Grant No.: SEP-210186649
Reference No.: 659933
Host: Chalmers University of Technology (Prof. Henk Wymeersch + Prof. Erik G. Ström)
Start Date: Sep. 07, 2015 
E.U. Website: http://cordis.europa.eu/project/rcn/195863_en.html
Chalmers URL:

https://research.chalmers.se/en/project/?id=6579

http://www.chalmers.se/en/projects/Pages/Modeling-and-Analysis-of-Random-Spatial-Systems-for-5G-Networks.aspx 

 


 

It is important to highlight that ~ 40% of all accidents occur at the intersection, and so carefully studying the so-called "intersection problem" is crucial. The following 4 movies (2min each!) show the reliability per random traffic realization as a transmitting vehicle (TX) communicates with a receiving vehicle (RX) while it maneuvers around the intersection. We consider the granular reliability per traffic in a build-up metropolitan and in the suburbs. The results are quite remarkable. More on the "significance" of what the results really means for real-world deployment on vehicles with autonomous capability soon.....

 

Practical Values for Real-World Analysis

Extreme Values for Fundamental Limits and Stress-Test Analysis

Case I: Urban Intersection, Small Road Segment (R = 200 m)

Case II: Urban Intersection, Sizeable Road Segment (R = 10 km)

 

 

 

 

Case III: Suburban Intersection, Small Road Segment (R = 200 m)

Case IV: Suburban Intersection, Sizeable Road Segment (R = 10 km)

 

 

 

 

 

 

 

More deatils about the above 4 movies are available here. Click the "TAB" for the case of interest to learn more about the outcome of V2V reliability.

Case I: Urban Intersection, Small Road Segment (R = 200 m)

The following short movie (~ 2 min only) shows the vehicle-to-vehicle (V2V) communications reliability for each standalone traffic realization around a corner in an urban intersection using Monte Carlo techniques based on plausible real-world traffic scenarios, channel propagation and system parameters. Press play and see the result!

 

 

We design the vehicular ad hoc network (VANET) in such a way that it inevitably meets a predefined target reliability. In other words, we want to determine the optimum transmit probability; i.e. the percentage of vehicles that can simultaneously transmit at the same time-frame and frequency-band of the wanted TX/RX pair. The design requirements and relationships are shown by the plots below for different a priori target reliability values evaluated for the worst-case TX/RX positions around an urban intersection. For 5G communications, the maximum TX/RX separation is 100 meters; and so we consider a TX and RX where both are located 50 meters away from a junction point on orthogonal roads.

 

Below, we show 1,000 random instances of the considered vehicular traffic. The intensity of each road is set to 0.01 vehicles per meter. Each speck that is shown below represent a vehicle (say a typical car of length ~ 5 meters) driving on one of the two roads forming the urban intersection. Each road is finite with road segment: R = 200 meters (in other words, the street length is 2R = 400 meters).

The previous "tab" showed the physical vehicular traffic considered. What is depicted below is the vehicular traffic composed of a TX and RX with deterministic positions around the urban intersection. Moreover, we show (with red specks) interfering vehicles that are actively transmitting at the same time-frame and frequency-band of the TX/RX pair. This graph is in fact the tolerated vehicular traffic while target reliability of 90% at the RX is still met. As you can see, in some instances, no interferers are tolerated by the RX. As such, we find it insightful to display the "void probability" of this tolerated traffic (i.e. the likelihood of not seeing a generated red speck or specks).

We analytically derived the average V2V reliability among communicating vehicles around an urban intersection with finite road segments. The plot below shows the outage probability for a network that is designed to meet a target reliability of 90% at the maximum TX/RX separation of 100 meters; which again is prescribed by the 5G requirements for V2V communications.

We show the meta distribution (MD) of reliability for each value of outage probability conditioned on a vehicular traffic. The accuracy of the displayed MD results improve as we consider a larger number of vehicular traffic realizations (i.e. as nppp increases). The plots utilizes 5,000 fading iterations to estimate each outage probability value associated with a particular vehicular traffic realization. These results are based on a designed vehicular network to meet a target reliability of 90% at the RX.

This meta distribution plot is similar to the previous "tab". The only exception is that the considered vehicular network is not designed to meet a certain predefined target reliability. Here, irrespective of the resulted reliability at the worst-case TX/RX separation at 100 meters, we consider a fixed Aloha transmit probability of 2%. 

 


 

Case II: Urban Intersection, Sizeable Road Segment (R = 10 km)

The following short movie (~ 2 min only) shows the vehicle-to-vehicle (V2V) communications reliability for each standalone traffic realization around a corner in an urban intersection using Monte Carlo techniques based on plausible real-world traffic scenarios, channel propagation and system parameters. Press play and see the result!

 

 

We design the vehicular ad hoc network (VANET) in such a way that it inevitably meets a predefined target reliability. In other words, we want to determine the optimum transmit probability; i.e. the percentage of vehicles that can simultaneously transmit at the same time-frame and frequency-band of the wanted TX/RX pair. The design requirements and relationships are shown by the plots below for different a priori target reliability values evaluated for the worst-case TX/RX positions around an urban intersection. For 5G communications, the maximum TX/RX separation is 100 meters; and so we consider a TX and RX where both are located 50 meters away from a junction point on orthogonal roads.

 

Below, we show 1,000 random instances of the considered vehicular traffic. The intensity of each road is set to 0.01 vehicles per meter. Each speck that is shown below represent a vehicle (say a typical car of length ~ 5 meters) driving on one of the two roads forming the urban intersection. Each road is finite with large road segment: R = 10 km (in other words, the street length is 2R = 20 km).

The previous "tab" showed the physical vehicular traffic considered. What is depicted below is the vehicular traffic composed of a TX and RX with deterministic positions around the urban intersection. Moreover, we show (with red specks) interfering vehicles that are actively transmitting at the same time-frame and frequency-band of the TX/RX pair. This graph is in fact the tolerated vehicular traffic while target reliability of 90% at the RX is still met. As you can see, in some instances, no interferers are tolerated by the RX. As such, we find it insightful to display the "void probability" of this tolerated traffic (i.e. the likelihood of not seeing a generated red speck or specks).

We analytically derived the average V2V reliability among communicating vehicles around an urban intersection with finite road segments. The plot below shows the outage probability for a network that is designed to meet a target reliability of 90% at the maximum TX/RX separation of 100 meters; which again is prescribed by the 5G requirements for V2V communications.

We show the meta distribution (MD) of reliability for each value of outage probability conditioned on a vehicular traffic. The accuracy of the displayed MD results improve as we consider a larger number of vehicular traffic realizations (i.e. as nppp increases). The plots utilizes 5,000 fading iterations to estimate each outage probability value associated with a particular vehicular traffic realization. These results are based on a designed vehicular network to meet a target reliability of 90% at the RX.

This meta distribution plot is similar to the previous "tab". The only exception is that the considered vehicular network is not designed to meet a certain predefined target reliability. Here, irrespective of the resulted reliability at the worst-case TX/RX separation at 100 meters, we consider a fixed Aloha transmit probability of 2%. 

 


 

Case III: Suburban Intersection, Small Road Segment (R = 200 m)

The following short movie (~ 2 min only) shows the vehicle-to-vehicle (V2V) communications reliability for each standalone traffic realization around a corner in an suburban intersection using Monte Carlo techniques based on plausible real-world traffic scenarios, channel propagation and system parameters. Press play and see the result!

 

 

We design the vehicular ad hoc network (VANET) in such a way that it inevitably meets a predefined target reliability. In other words, we want to determine the optimum transmit probability; i.e. the percentage of vehicles that can simultaneously transmit at the same time-frame and frequency-band of the wanted TX/RX pair. The design requirements and relationships are shown by the plots below for different a priori target reliability values evaluated for the worst-case TX/RX positions around an suburban intersection. For 5G communications, the maximum TX/RX separation is 100 meters; and so we consider a TX and RX where both are located 50 meters away from a junction point on orthogonal roads.

 

Below, we show 1,000 random instances of the considered vehicular traffic. The intensity of each road is set to 0.01 vehicles per meter. Each speck that is shown below represent a vehicle (say a typical car of length ~ 5 meters) driving on one of the two roads forming the suburban intersection. Each road is finite with road segment: R = 200 meters (in other words, the street length is 2R = 400 meters).

The previous "tab" showed the physical vehicular traffic considered. What is depicted below is the vehicular traffic composed of a TX and RX with deterministic positions around the suburban intersection. Moreover, we show (with red specks) interfering vehicles that are actively transmitting at the same time-frame and frequency-band of the TX/RX pair. This graph is in fact the tolerated vehicular traffic while target reliability of 90% at the RX is still met. As you can see, in some instances, no interferers are tolerated by the RX. As such, we find it insightful to display the "void probability" of this tolerated traffic (i.e. the likelihood of not seeing a generated red speck or specks).

We analytically derived the average V2V reliability among communicating vehicles around an suburban intersection with finite road segments. The plot below shows the outage probability for a network that is designed to meet a target reliability of 90% at the maximum TX/RX separation of 100 meters; which again is prescribed by the 5G requirements for V2V communications.

We show the meta distribution (MD) of reliability for each value of outage probability conditioned on a vehicular traffic. The accuracy of the displayed MD results improve as we consider a larger number of vehicular traffic realizations (i.e. as nppp increases). The plots utilizes 5,000 fading iterations to estimate each outage probability value associated with a particular vehicular traffic realization. These results are based on a designed vehicular network to meet a target reliability of 90% at the RX.

This meta distribution plot is similar to the previous "tab". The only exception is that the considered vehicular network is not designed to meet a certain predefined target reliability. Here, irrespective of the resulted reliability at the worst-case TX/RX separation at 100 meters, we consider a fixed Aloha transmit probability of 2%. 

 


 

Case IV: Suburban Intersection, Sizeable Road Segment (R = 10 km)

The following short movie (~ 2 min only) shows the vehicle-to-vehicle (V2V) communications reliability for each standalone traffic realization around a corner in an suburban intersection using Monte Carlo techniques based on plausible real-world traffic scenarios, channel propagation and system parameters. Press play and see the result!

 

 

We design the vehicular ad hoc network (VANET) in such a way that it inevitably meets a predefined target reliability. In other words, we want to determine the optimum transmit probability; i.e. the percentage of vehicles that can simultaneously transmit at the same time-frame and frequency-band of the wanted TX/RX pair. The design requirements and relationships are shown by the plots below for different a priori target reliability values evaluated for the worst-case TX/RX positions around an suburban intersection. For 5G communications, the maximum TX/RX separation is 100 meters; and so we consider a TX and RX where both are located 50 meters away from a junction point on orthogonal roads.

 

Below, we show 1,000 random instances of the considered vehicular traffic. The intensity of each road is set to 0.01 vehicles per meter. Each speck that is shown below represent a vehicle (say a typical car of length ~ 5 meters) driving on one of the two roads forming the suburban intersection. Each road is finite with large road segment: R = 10 km (in other words, the street length is 2R = 20 km).

The previous "tab" showed the physical vehicular traffic considered. What is depicted below is the vehicular traffic composed of a TX and RX with deterministic positions around the suburban intersection. Moreover, we show (with red specks) interfering vehicles that are actively transmitting at the same time-frame and frequency-band of the TX/RX pair. This graph is in fact the tolerated vehicular traffic while target reliability of 90% at the RX is still met. As you can see, in some instances, no interferers are tolerated by the RX. As such, we find it insightful to display the "void probability" of this tolerated traffic (i.e. the likelihood of not seeing a generated red speck or specks).

We analytically derived the average V2V reliability among communicating vehicles around an suburban intersection with finite road segments. The plot below shows the outage probability for a network that is designed to meet a target reliability of 90% at the maximum TX/RX separation of 100 meters; which again is prescribed by the 5G requirements for V2V communications.

We show the meta distribution (MD) of reliability for each value of outage probability conditioned on a vehicular traffic. The accuracy of the displayed MD results improve as we consider a larger number of vehicular traffic realizations (i.e. as nppp increases). The plots utilizes 5,000 fading iterations to estimate each outage probability value associated with a particular vehicular traffic realization. These results are based on a designed vehicular network to meet a target reliability of 90% at the RX.

This meta distribution plot is similar to the previous "tab". The only exception is that the considered vehicular network is not designed to meet a certain predefined target reliability. Here, irrespective of the resulted reliability at the worst-case TX/RX separation at 100 meters, we consider a fixed Aloha transmit probability of 2%. 

 

Overview:

 
The United Nations International Telecommunications Union (ITU) has recently reported that the number of mobile users is expected to surpass the world population. In fact, it is expected that by 2020 mobile users will exceed an astonishing 9 Billion subscribers. With this unprecedented market penetration and growth, many communication engineering challenges are anticipated. In particular, next generation 5G wireless systems must be designed using sophisticated and innovative strategies and techniques. While 5G applications are expected to be diverse, the network architectures and devices need to ensure and deliver reliable, pervasive, and high-speed interconnection for various data-intensive applications (e.g., interactive multimedia streaming). These requirements must be accomplished while necessitating limited resources for a continuously expanding consumer demographics. Thus, deploying such complex networks is a formidable engineering feat that requires novel ways of modeling, evaluating, and designing extremely dense radio systems. This project aims, through the study of spatial geometry of randomly deployed mobile units, to develop several analytical tools to model, design, and analyze complex 5G networks, and validate them through experimental datasets. Ultimately, our broad goal is to conceptualize an engineering research idea, and then transitioning it into innovative applications that can be replicated for real-world cellular networks operated by established service providers and mobile manufacturers. 

 

 

 
 

Research:

 
Track no.1 — Packet Detection Performance of V2V Communications about an Urban Intersection:

Vehicle-to-vehicle (V2V) communication about an intersection (either based on roundabouts or traffic signals) is a challenging problem that researchers are carefully investigating; among them, giant automotive manufacturers in Sweden, such as Volvo and Scania. Such research has various implications, particularly for accident avoidance systems, seamless connectivity, and future driverless vehicles. In addition to upper-layer analysis, extensive communications-based research is required at the link-layer level with strong emphasis on the network, the protocol, and the physical stack. In our research, the focus will be on the PHY and MAC layers, where we are especially interested to conceptualize, characterize and eventually derive predictive analytical models that offer some insight into this highly complex problem. The modeling and analysis that we intend to tackle will evaluate the quality of service (QoS) pertaining to the successful detection of packets between V2V communicating cars on orthogonal streets near an intersection.

The researchers interested into the V2V interconnection viability at the intersection have identified various intricacies that require careful investigation. As a consequence, we ought to derive performance evaluation models that can effectively be tuned for intersections in a built-up urban setting, or suburban and rural environments. Each of these situations will exhibit channel characteristics that are unique. The modeling that we will consider for the propagation channel of these environments will in fact be based on recent empirical measurements that have been conducted by the automotive industry. Meanwhile, in addition to the randomness introduced by the channel, the communications fidelity will also be impacted by random interfering vehicles. Overall the analysis for deriving the generic performance tools near an intersection will require a solid foundation in wireless communications, vehicular networks, stochastic geometry, RF propagation, probability theory, and Monte Carlo simulations. The benefit of the anticipated theoretical analysis that we aim to solve, will serve as a first-order tool that evaluates the transmission quality and transmission capacity as a function of vehicular positions, traffic intensities, and channel environments.

 

 
Track no.2 —VANET Communication Link over Highways and Interchanges:

A vehicular ad hoc network (VANET) is formed through the transmission of data packets among vehicular nodes by automatically hoping from source to destination. At core, and in its most basic form, this so-called spontaneous network requires that the data transfer among two communicating vehicular nodes be successfully accomplished through the effective minimization of the outage probability at the link-layer; and this must be fulfilled for each hop. In itself, the VANET principle is quite complex to establish; and if such network is desired to be formed over highways, then the related engineering endeavor will substantially be more intricate. Despite the overwhelming challenge, preliminary research in this topic could provide some basic understanding of the problem. In our research, we are interested to explore the quality of the inter-vehicular communication (IVC) while taking into account the irregularity of highways. Granted, expressways and freeways are generally straight roads; however, these roads may in fact be irregular, particularly at interchanges that connect different highways together. Overall, the aim for modeling and analyzing predictive connectivity tools for VANET over highway interchanges is expected to enlighten us on the interrelation among: the IVC communication range required to sustain a link; the status and profile of the road; the vehicular traffic; and the characteristics of the channel.

 

  
Track no.3 :

...multilane roads ....  ... they each have a density ... 

 

 
Track no.4 :

...cell overlap (aka the cell-edge problem) is one of the most crucial topic in modern communications .... particularly as the size of the cell shrinks.... interference, spatial correlation of shadowing of AP during downlink ..... network densification ... 

 

 

 

 

 
Track no.5 — Empirical Validation of Synthetic Random Networks:

...Experimental research/POC  and validation is needed to ascertain 

 

 

 

 

 

Movable Access-Point RF Transceivers Channel Evaluation

 

 
 

Seminars:

Seminars Presented by the Principal Investigator (PI) of the "EU-MARSS-5G Project":


Seminars Organized Under the "EU-MARSS-5G Project":




 

M. Abdulla, "On the Facets of Creative and Comprehensive Engineering Design," Communication Systems (ComSys) Seminar, Dept. of Signals and Systems, Chalmers University of Technology, Gothenburg, Sweden, Sep. 25, 2015.

 

 

 
E. Steinmetz, "Using Stochastic Geometry to Model Packet Reception Probabilities in Vehicular Networks," IEEE VTS Society Workshop on Wireless Vehicular Communications, Halmstad University, Halmstad, Sweden, Nov. 11, 2015.

 
 

Connected Vehicles Team:

 

Mouhamed Abdulla

Marie-Curie Fellow (PI)

Chalmers U. Tech.

Henk Wymeersch

Professor (Co-PI)

Chalmers U. Tech.

Erik M. Steinmetz

Ph.D. Candidate

SP Technical Research

Srikar Muppirisetty

Ph.D. Candidate

Chalmers U. Tech.

Performance Evaluation

 

Network Analysis

V2V Communications

Predictive Modeling

 

Bjarki Vilmarsson

M.Sc., MPSYS

Chalmers U. Tech.

Daoyuan Yang

M.Sc., MPCOM

Chalmers U. Tech.

Suhail Ahmad

M.Sc., MPCOM

Chalmers U. Tech.

Rikard Reinhagen

B.Sc. Student

Chalmers U. Tech.

Martin Dahlgren

B.Sc. Student

Chalmers U. Tech.

System Design

Design Application Tools

Measurement Campaigns

Empirical Analysis

Hardware Configuration

 

 

 

News:

 
Interview with Chalmers University of Technology's INSIDAN, Sep. 17, 2015. [S2 newsletter]

 


 

S. Valji, "Prestigious Marie Curie Fellowship for Mouhamed Abdulla: Faculty of Engineering and Computer Science Graduate off to Sweden this Fall," Advancement and Alumni Relations, Concordia University, Jun. 18, 2015.

 

Letter from Concordia University President and Vice-Chancellor,

Prof. Alan Shepard, Jul. 15, 2015.

Letter from Concordia University Dean,

Prof. Paula Wood-Adams, Jul. 16, 2015.

 

 

Outreach:

 

We are very pleased to announce that were invited by the organizing committee of Europe's leading popular science event (with 70,000 visitors), "The International Science Festival in Gothenburg" to co-organize and participate in various science and technology based activities and exercises, including a panel discussion that will bring R&D engineers from academia and industry working on next-generation telecommunication systems (5G+). The details of our event are as follows:

 

Title of Panel   

Discussion:  

"Defining the Future of our Wireless World"

 

Researchers are trying to solve some of the big challenges of faster, more reliable and high capacity mobile networks for the next generation - 5G. Will 5G bring new dimensions to wireless? What other smart wireless applications can we expect in the future? Our panel of speakers will give exciting peeks into the future enabled by new wireless technologies. Hands-on activities will also be demonstrated!

Moderator:   Ms. Malin Ulfvarson
Panelist:   Dr. Katrin Sjöberg (Connected vehicle technology specialist, advanced technology and research, Volvo Group), Dr. Ayca Ozcelikkale (Research Fellow, Chalmers), Dr. Amina Piemontese (Research Fellow, Chalmers), Mr. Sven Jacobsson (Industrial PhD, Ericsson Research), Dr. Mouhamed Abdulla (Research Fellow, Chalmers).

Date/Time  

and Location: 

Panel: April 17, 2016 at 4pm CET -- Nordstan

Activities: April 15, 18, 19, 2016 -- Chalmers Johanneberg Campus

 

The panel that we co-organized will feature researchers from the following European institutions:

 

 

 

We will regularly share updates of the MARSS-5G project on social-media. For more, you may follow us on the following platforms:

 

 

 
 

Last Update: May 25, 2017 10:06 PM CET | Copyright © 2015-2016. All Rights Reserved. This page is maintained by Mouhamed Abdulla