A METHOD TO IMPROVE URBAN VEHICULAR NETWORKS BY TRAJECTORY DATA DELIVERY
Maan Nawaf Abbood1*
*alanimm2@gmail.com Zahraa Adnan Abdalkareem1 zahraa20102015@gmail.com
1Alimam aladham university college
*corresponding author
ABSTRACT
This report proposes a Trajectory-Based Data Forwarding (TBD) plan, custom-made for the information sending in light-activity vehicular impromptu systems. We look at the spots in which Internet access focuses are scantily sent to bring the street side reports on time-basic data, for instance, driving mishap or danger. Since the Internet access focuses have constrained correspondence scope, a vehicular specially appointed system is required to forward information bundles to the entrance focuses. Best in class programs have shown the adequacy of their information exploiting so as to send methodologies known vehicular movement measurements (e.g., Densities and rates) in such a system. These results are empowering; on the other hand, further alterations can be caused by exploiting the growing notoriety of GPS-based route frameworks. This paper presents the first enterprise to explore how to viably use vehicles' direction data in a protection safeguarding way. In our schema, the way data is consolidated with the movement insights to enhance the implementation of information sending in street systems. Through hypothetical examination and broad recreation, it is shown that our configuration outflanks the current program as far as both the information conveyance deferral and parcel conveyance proportion, uncommonly under light-movement circumstances.
Keywords: Driving accident, Trajectory-Based Data Forwarding, GPS systems, vehicular ad-hoc network.
1. INTRODUCTION:
A vehicular system is a system of vehicles which com-municate with one another by means of remote interchanges. Vehicular systems have numerous engaging applications, for instance, driving wellbeing, Internet access, clever transport, and framework observing. Proficient information conveyed is of focal significance to vehicular systems, which concentrates on execution measurements, including conveyance proportion, deferral and throughput. Contrasted and customary versatile, specially appointed systems, information conveyance in vehicular systems confronts an arrangement of new troubles. In the inaugural blank space, vehicular systems are liable to constant interruptions. It is tough to discover a joined way between more than one source and destination in vehicular structures. That is conveyed approximately through high portability and uneven circulate of cars over the employer. The set off gadget topology framed by 2500 taxies in the shanghai metropolitan metropolis, china. 2nd, automobiles normally move at a speedy. Automobiles can deliver simply while they're inside of the corresponding range. Recent examine has provided that the touch time period if there ought to be an occurrence of a car and a static get right of entry to point is as low as 10 seconds on average (Zhu, Y., Wu, Y., & Li, B ,2014).
All the more imperatively, at that place is a great deal of instability associated with vehicle portability.
Vehicles move at their own volitions. It is troublesome, if not impossible, to pluck up the complete information about the vehicle hint of future growth. For maneuvering in a vehicular system, a transfer hub must choose to what extent a bundle ought to be saved and which hub a given parcel ought to be committed to. The existing study demonstrates that it is conceivable to locate an ideal steering way, when
the learning of future hub follows is accessible; this is NP-hard however. Be that as it may, it is unrealistic to have previous learning about future hints of hops. Various computations have been meant for the information store exceptionally in vehicular systems, which can freely be isolated into two categories.
The main class basically expects the availability of future development, that is, the hints of hubs are altered and can be recognized in advance. GeOpps expect that the lead of a hub can be gained through the route framework prepared locally available in the vehicle. Such calculations are constrained by the accessibility of route frameworks and the affinity of drivers (Zhao, J., & Cao, G.,2008).
The second class tests to get the information of future portability by expectations or ideas. It has been exposed that the between meeting time is exponentially circulated construct just in light of the chronicled gatherings. In this manner, the conveyance deferral of a bundle in view of the between meeting time circulation can be seen. MobySpace keeps up appearance frequencies of hubs in a given spatial place, and gauges meeting opportunities between any two hubs. In Max-Contribution tries to together upgrade connections and parcels, planning, furthermore accept the exponential distribution of between meeting times. In this room, the learning of future vehicle follows assumes a key function of ideal information conveyance. Existing steering calculations intensely depend on forecasts of vehicle versatility. Be that as it may, they have embraced just straightforward portable examples, for instance, the spatial dis-tribution and between meeting time dissemination, which bolster coarse - grained expectations of vehicle developments. A few calculations accept arbitrary portability in which vehicles move haphazardly in an open space or a street system. This example is straightforward, however is a long path from the genuine versatility. Some different calculations accept straightforward portable examples, for model, exponential interesting times and general spatial dispersions.
Accordingly, forecast results in light of these straightforward examples are of constrained quality to effective information conveyance in vehicular systems. What's more, a number of existing calculations disregard the way that connections in a vehicular system have special characters. From one viewpoint, a association is regularly brief.. This recommends the limit of the connection is constrained. In this way, the request for sending parcels gets to be decisive. And so again, connects in a densely populated zone may meddle with one another. This shows connection booking gets to be primal. To capture the constraints in the current deliberations, this paper proposes directing calculations that exploit direction forecasts. By mining the broad follow Data units of greater than 4000 taxies over a length of time of over two years amassed in shanghai metropolitan town, china, we happen that there may be stable spatiotemporal normality with car mobility. All of the more specifically, our effects in mild of restrictive entropy exam display that the future management of a vehicle signal corresponds with its beyond control. Hence, we building up diverse request markov chains for waiting for future directions of the motors. With the available destiny directions of cars, we endorse an explanatory model and hypothetically infer the conveyance chance of a packet(Zhao, J., & Cao, G. ,2008).
Because the foremost routing problem with given automobile trajectories continues to be np- difficult, we acquire an green international set of rules for computing routing paths when predicted trajectories are to be had. For maximum realistic situations, we develop a fully allotted set of rules which wishes best localized statistics. The two algorithms at the same time recall packet scheduling and hyperlink scheduling. We compare the algorithms with sizable trace pushed simulations, based along the trace datasets accrued in shanghai. The answers display that our algorithm significantly outperforms other algorithms in terms of transport probability and transport efficiency.
By way of mining, vast hint datasets of taxies, we demonstrate that there is robust spatiotemporal regularly with automobile mobility via conditional entropy evaluation. We make powerful trajectory predictions by growing more than one order markov chains primarily based on the good sized ancient lines. We provide an analytical version and theoretically derive shipping chance with anticipated vehicle trajectories. We provide both a international algorithm and a fully dispensed set of rules. We demonstrate with hint-driven simulations that the algorithms attain higher transport ratios at lower value compared with different algorithms.
2. RESEARCH PROBLEM:
Effective between vehicle information conveyance is of focal significance to vehicular systems and such significance has been perceived by numerous current studies. In this paper we concentrate on such vehicular systems that are scanty and do no expect that all vehicles out and about are part hubs of the vehicular system. Such inadequate vehicular systems highlight rare correspondence opportunities.
Between vehicle information conveyance may present nonneligible conveyance dormancy as a result of regular topology separation of a vehicular system. In this manner, we ought to stretch that the between vehicle correspondence in vehicular system is suited for those applications which can endure certain conveyance inactivity. For illustration, in the connection of urban detecting, vehicles constantly gather helpful information, for lesson, street movement conditions and street terminations. A vehicle may send a query for a peculiar sort of data and the one that receives the information ought to react the questioning hub with the data.
Such correspondence requires multi-jump information conveyance in vehicular systems. Unlike examples of such applications incorporate distributed document sharing, excitement, commercial, and record downloading. It has received just straightforward versatility examples, for instance, the spatial circulation and between meeting time dispersion, which bolster coarse-grained expectations of vehicle developments.
It ignored the way that connections in a vehicular system have interesting characteristics (Lee, U., Zhou, B., Gerla, M., Magistretti, E., Bellavista, P., & Corradi, A. ,2006).
3.RESEARCH METHOD:
To conquer the constraints of existing calculations, this report proposes a way to deal with misusing the concealed versatility consistency of vehicles to predict future directions. By mining the full dataset of vehicular follows from more than 4,000 taxis in Shanghai, China, we manifest that there is solid spatiotemporal normality with vehicle portability. All the more particularly, our results in light of contingent entropy examination show that the future management of a vehicle incredibly corresponds with its past management.
In this way, we add to different request Markov chains for anticipating future directions of the vehicles.
With the accessible future directions of vehicles, we suggest a consistent model and hypothetically determine the conveyance likelihood of a packet. It makes up a productive worldwide calculation for registering directing ways when anticipated directions are accessible. It impressively beats different calculations as far as conveyance likelihood and conveyance efficiency (Mo, Z., Zhu, H., Makki, K., &
Pissinou, N. ,2006).
3.1 DISTRIBUTED ALGORITHM:
The accord issue can be expressed as takes after: given a gathering of n procedures, they must concede to a character. Any answer for accord must fill three prerequisites: understanding, end and legitimacy. The main prerequisite says that all the procedure that choose need to pick the same character. The second alleges that all non-defective procedures in the long run choose, and the third states that the normal yield wort is an information estimation of some operation.
A surely understand result in the disseminated calculation field, called the FLP outlandish possibility result, states that there is no deterministic convention that fulfills the Ascension, end, and non-technicality conditions for a nonconcurrent message passing framework in which any single process can fall flat imperceptibly. Another comparative result, because of Loui and Abu-Amara, states the same outlandish possibility when the appropriated model is an offbeat shared-memory framework.
There are various different options for evading the FLP result. For exemplar, the utilization of clocks approximates the offbeat framework to a synchronous situation. Also, disappointment finders can evacuate slammed processors. It is additionally conceivable to ensure understanding by utilizing solid primitives, for example, test-and-set, or move-and-barter. At long last, in his newspaper, James Aspnes presents three unique calculations that utilization randomization keeping in mind the end goal to bypass
the FLP result. One of these calculations is for the common memory model. The other two are for the message-passing worldview.
In a randomized calculation, assention and legitimacy are as in the deterministic model, however end is marginally diverse: the calculation dependably ends with likelihood 1 the length of it has enough time to be executed. With a specific end goal to expand the computational force of the conveyed framework, randomized calculations utilize an operation that allows processes to create irregular numbers. In this model, there is the thought of the foe: a sequence of executions that endeavors to break up the calculation.
There are two sorts of enemies. The primary website is known as the solid foe, and it can pick up all the inner shape of the memory and the result of coin flips before picking the following operation to do. Its partner, the purported feeble enemy, can exactly pick the next procedure to execute (Mo, Z., Zhu, H., Makki, K., & Pissinou, N. ,2006).
The main calculation exhibited by Aspnes is the Ben-Or's Consensus convention. This was the First convention to carry through an agreement with probabilistic end in a mannequin with an in number foe (1983). It endures t < n/2 crash disappointments, and may require exponential anticipated that time would match in the most pessimistic scenario. The calculation runs in circles, each round having two phases. The main point is experienced as the recommendation step. In this level, every procedure transmits its worth, and retains up to get notification from different central processing units. The second phase is experienced as the choice step.
In this stage, if the larger part esteem is learned, the procedures take this world; else, they throw a coin to select another neighborhood esteem. The essential idea of this calculation is not to sit tight for messages from all the subroutine with a specific end goal to involve choices, in light of the fact that more or less of them may come up short. In the event that enough procedures distinguish the larger role, then choose for this caliber.
On the off chance that it is understood that a procedure has identified the greater region, every different procedure that receives such data must change to the larger part's worth. The calculation in the long run ends on the grounds that all the procedure will throw the same estimation of the coin, given countless (Jerbi, M., Senouci, S. M., Rasheed, T., & Ghamri-Doudane, Y. ,2009). A theatrical performance of the calculation is presented as follows:
The calculation dependably ends, even in the typeface of an in number foe, in light of the fact that the likelihood of a contradiction for a limitless time is 0. (It is tantamount to the likelihood that each bend there will be one 1 and one 0 until the end of time). In whatever event, the facts may confirm that the calculation presents exponential many-sided quality in the more regrettable case, considering that, if a large dowry of the procedure falls flat, the possibility of all the irregular coins to bear the same worth is 2- n, where n is the aggregate number of processes.
The following calculation displayed by Aspnes is the Chor-Israeil-Li Protocol. Its center thought depends on a race between procedures. Every routine is composed in the mutual memory its tune and its disposition. On the off chance that there is a procedure sufficiently far out front, alternate procedures take its quality. Every procedure flips a coin to pick if to propel a turn or not, with possibility of 1/2n to
Input: Boolean initial consensus value Output: Boolean final consensus value Data: Boolean preference, integer round begin
preference := input round := 1
while true do
send (1, round, preference) to all processes wait to receive n – t (1, round, *) messages if received more than n / 2 (1, round, v) messages then send (2, round, v, ratify) to all processes else send (2, round, ?) to all processes end
wait to receive n – t (2, round, *) messages If received a (2, round, v, ratify) message then preference = v
if received more than t (2, round, v, ratify) messages then output = v
end
else preference = CoinFlip() end
round = round + 1 end
end
progress. Probabilistic results guarantee that, after the normal time of n rounds, there will be a pioneer enough ahead. The calculation is as per the following:
The Chor-Israeil-Li Protocol gives frail accord, in light of the fact that a powerless foe can defer specific procedures, however can't recognize in which stage every procedure is. On the other hand, this convention neglects to crush on in number foe. The triumphant system for an in number foe is to terminate a procedure that increased its quality, until every different procedure have additionally done it.
Then as to illuminate the solid enemy accord in a common memory framework with preferred many-sided quality over exponential more awful case, Aspnes presents the idea of shared coin. This convention gives back a bit, 0 or 1, to every subroutine that takes part in it. The likeliness that the same character is coming back to all conjuring procedures is in any event some settled parameter 'e'. Furthermore, the likelihood that the convention returns 1 to all procedures, levels with the likelihood that the convention returns 0 to every one of them. At last, this convention is hold up free (Wu, Y., Zhu, Y., & Li, B. ,2011).
3.2 GLOBAL ALGORITHM:
Calculations of succession arrangement are the central instruments for PC helped investigations of polymers. Understandably, it is vital to look at the "quality" of the got arrangements, i.e. how nearly the
Input: Initial preference Output: consensus value
Local data: preference, round, maxround
Shared data: one single-writer multi-reader register for each process.
begin
preference := Initial preference;
round := 1;
while true do
write (preference, round) read all registers R
maxround := maxR R.round
if for all R where R.round >= maxround - 1, R.preference = v then return v
else if exists v such that for all R where R.round = maxround, R.preference = v then preference := v
end end
with probability 1/2n do
round := round(round + 1, maxround - 2) end
end
calculations figure out how to restore the "highest quality level" arrangement (GS-arrangement), which superimposes positions starting from the same position in the basic progenitor of the looked at groupings.
As a guess of the GS-arrangement, a 3D-arrangement is regularly utilized not exactly sensibly. Among the right now utilized calculations of a couple insightful arrangement, the best character is carried out by utilizing the calculation of ideal arrangement taking into account relative punishments for erasures (the Smith-Waterman calculation). In whatever event, the practicality of utilizing nearby or worldwide variants of the computation has not been studied. Utilizing model arrangement of amino corrosive grouping sets, we focused on the relative "quality" of solutions delivered by nearby and worldwide arrangements versus (1) the relative length of comparative parts of the successions (their "centers") and their nonhomologous parts, and (2) relative positions of the core areas in the looked at groupings. We got numerical estimations of the normal quality (measured as precision and certainty) of the worldwide arrangement technique and the nearby arrangement system for developmental separations between homologous succession parts from 30 to 240 PAM and for the center length, making from 10% to 70% of the aggregate length of the groupings for every single conceivable position of homologous arrangement parts with respect to the centers of the sequences (Skordylis, A., & Trigoni, N. ,2008).
We uncovered criteria permitting to determine the status of favored pertinence for the nearby and the worldwide arrangement calculations relying upon the locations and relative distances of the centers and nonhomologous parts of the groupings to be corrected. It was proven that when the center a portion of one grouping was situated over the middle of the other succession, the worldwide calculation was more steady at longer transformative separations and bigger nonhomologous parts than the nearby calculation. Despite what might be expected, when the centers were situated lopsidedly, the nearby calculation was more steady at longer transformative separations and bigger nonhomologous parts than the worldwide algorithm(Skordylis, A., & Trigoni, N. ,2008). This opens a probability for production of a consolidated s y s t e m p e r m i t t i n g e r a o f m o r e p r e c i s e p l a c e m e n t s .
!
Altered successions with consoles. Here L1, L2, K1, K2, R1, R2 is left consoles, centers and right consoles of the first and second altered successions; their lengths |L1| = |R2| = 9, |L2| = |R1| = 11, |K1| ≠ | K2| (the disparity is clarified by varying length of irregular insertions and erasures). At that point r = (|L1|
+ |R1|)/|P| = (9+11)/200 = 0.1, c = (|R1| - |L1|)/|(|L1| + |R1|) = 2/20 = 0.1, where |P| = 200 is the length of the beginning succession.
4. RESULTS
We first present execution aftereffects of our calculations against different calculations. We then study the effect of lattice size. We fluctuate the measure of parcels and look at the five algorithms as far as conveyance proportion, normal defer and aggregate expense under diverse heaps of system. For a given context of measure of parcels, the same arrangement of packages and the arrangement of hubs are utilized for all the five calculations. Since the conveyance proportion is to a large extent influenced when length recreated, we employ the metric of relative conveyance proportion rather than exposed conveyance proportion, which is the conveyance proportion of every calculation standardized by that of Flooding. In Figure 4, the execution of the five calculations as far as relative conveyance proportion is appeared. We can assure that our calculations per-structures superior to anything P-Random and Max-humility. Among the five calculations, Flooding performs the best, of course. All in all, the calculations that utilization expectations create preferable conveyance proportions over the computations that make no forecasts. Our calculations are superior to anything Max-Contribution in light of the fact that our calculations use the recorded follow data as well as the present state and the past states. Anticipated trajectories of vehicles discover better steering ways. The worldwide calculation is superior to anything the conveyed one since it receives the worldwide, complete system data. We can similarly see that when the quantity of bundles builds, the general conveyance proportions of all calculations diminish. The understanding is that the universal limit of the scheme is limited. By infusing more bundles, the shares may go after system assets and subsequently, less parcels can be channeled at least.
In normal deferral against the quantity of bundles is plotted. Our calculations have a lower postponement than the P - Random and Max-Contribution. Not surprisingly, Flooding has the least delay. The normal postponements of our calculations are marginally larger than that of flooding. This execution increase of our calculations is largely because of the way that steering ways with high conveyance probabilities as a rule lead to shorter time lags. Since our calculations can choose to address ways of high conveyance probabilities, the resultant deferral is low. Disbursement of the five calculations. We can see that our calculations have lower expenses, superior to the various calculations. The principle cause is that by adequately foreseeing the directions of vehicles, we barely count the ways that prompt an inevitable conveyance with high likelihood. Accordingly, superfluous parcel exchanges are enormously decreased.
In our face at the efficiencies of the five calculations. We can assure that our calculations have higher productivity, superior to the various calculations. Flooding and P-Random have a comparative productivity and are a lot more awesome than the rest three. This is because of the visual impairment of Flooding and P-Random when they are squaring off on exchange choices. Max-Contribution has a bigger productivity than Flooding and P-arbitrary, nevertheless delivers a lower effective than our calculations, since Max-Contribution only uses a straightforward versatile example of between meeting time (Lee, U., Magistretti, E., Gerla, M., Bellavista, P., & Corradi, A. ,2009).
The framework size impacts expectation and hence the calculation execution. On the off chance that we employ a more prominent framework size, two vehicles in the same network would be reckoned to experience with one another, however as a general rule they don't. Be that as it may, in the meantime, a more noteworthy matrix size prompts a lower calculation unpredictability. We focus on the issue of network size on conveyance execution. The technique of the dispersed calculation is plotted against diverse matrix sizes (Chen, W., Guha, R. K., Kwon, T. J., Lee, J., & Hsu, Y. Y. ,2011). We see that when all is said in done a littler matrix result in a superior product. At the period when the framework size expands, the productivity de-creases. The understanding is that when the matrix size is bigger, the direction expectation turns out to be more regrettable. Entropy CDFs when the matrix size is 3 kilometer.
By contrasting them and the CDFs, we can see that a more noteworthy matrix size results in a small crack between the trifling and the restrictive CDFs (Lee, U., & Gerla, M. ,2010). This recommends a more prominent network size lessens the viability of direction forecast.
5. DISCUSSIONS
The appropriated calculation requires the vehicles to trade metadata about versatile examples of vehicles.
From one standpoint, the arrangement of vehicles may change after some time. Be that as it may, we watch that in an urban situation, the arrangement of vehicles is fairly firm. In our theoretical account, for instance, taxies in Shanghai shape the vehicular system, which does not change regularly. And so again, the versatile example of a vehicle might also change after some time. All things weighed, the consistency of a vehicle is additionally steady. All the more critically, we circulated calculation backings overhauling of portable examples. This is carried out by every vehicle proliferating its new model. Living with an overhauled design, different vehicles can upgrade the vehicle's example accordingly (Fiore, M., &
Barcelo-Ordinas, J. M. ,2009).
The dispersed calculation likewise requires the most redesign area of vehicles. The perfect example is that every vehicle can sustain the present fields of whole vehicles, e.g., by a third channel. On the other hand, this may introduce a restrictive cost (Jeong, J., Guo, S., Gu, Y., He, T., & Du, D. H. ,2011). By getting the most overhaul areas in vehicular systems, the area learning about different vehicles kept up by a vehicle may be time slacked. The time slacked area may corrupt the expected quality, however, anticipated directions are still of worth to steering decisions (Skordylis, A., & Trigoni, N. ,2011).
6. CONCLUSION
Vehicular systems have received significant consideration re-cently. In malice of the fact that information conveyance of vehicular systems has been mulled over, few leaving calculations successfully abuse directions of the vehicles. In this report, we demonstrate the solid mountain-o temporal normality with vehicle portability by entropy only-sister. By adding to various request Markov chains, we anticipate vehicle directions. Taking into account the investigative model, we infer the conveyance likelihood with the expected charges. The ace postured calculations exploit vehicle directions. Execution results, check that our calculation beats different calculations. This exhibits anticipated directions do help information conveyance in vehicular systems.
REFERENCES
Chen, W., Guha, R. K., Kwon, T. J., Lee, J., & Hsu, Y. Y. (2011). A survey and challenges in routing and data dissemination in vehicular ad hoc networks. Wireless Communications and Mobile Computing, 11(7), 787-795.
Fiore, M., & Barcelo-Ordinas, J. M. (2009, October). Cooperative download in urban vehicular networks. In Mobile Adhoc and Sensor Systems, 2009. MASS'09. IEEE 6th International Conference on (pp. 20-29). IEEE.
Jeong, J., Guo, S., Gu, Y., He, T., & Du, D. H. (2011). Trajectory-based data forwarding for light-traffic vehicular ad hoc networks. Parallel and Distributed Systems, IEEE Transactions on, 22(5), 743-757.
Jerbi, M., Senouci, S. M., Rasheed, T., & Ghamri-Doudane, Y. (2009). Towards efficient geographic routing in urban vehicular networks. Vehicular Technology, IEEE Transactions on, 58(9), 5048-5059.
Lee, U., & Gerla, M. (2010). A survey of urban vehicular sensing platforms. Computer Networks, 54(4), 527-544.
Lee, U., Magistretti, E., Gerla, M., Bellavista, P., & Corradi, A. (2009). Dissemination and harvesting of urban data using vehicular sensing platforms. Vehicular Technology, IEEE Transactions on, 58(2), 882-901.
Lee, U., Zhou, B., Gerla, M., Magistretti, E., Bellavista, P., & Corradi, A. (2006). Mobeyes: smart mobs for urban monitoring with a vehicular sensor network. Wireless Communications, IEEE, 13(5), 52-57.
Mo, Z., Zhu, H., Makki, K., & Pissinou, N. (2006, July). MURU: A multi-hop routing protocol for urban vehicular ad hoc networks. In Mobile and Ubiquitous Systems: Networking & Services, 2006 Third Annual International Conference on (pp. 1-8). IEEE.
Skordylis, A., & Trigoni, N. (2008, May). Delay-bounded routing in vehicular ad-hoc networks. In Proceedings of the 9th ACM international symposium on Mobile ad hoc networking and computing (pp.
341-350). ACM.
Skordylis, A., & Trigoni, N. (2011). Efficient data propagation in traffic-monitoring vehicular networks.
Intelligent Transportation Systems, IEEE Transactions on, 12(3), 680-694.
Wu, Y., Zhu, Y., & Li, B. (2011, April). Trajectory improves data delivery in vehicular networks. In INFOCOM, 2011 Proceedings IEEE (pp. 2183-2191). IEEE.
Zhao, J., & Cao, G. (2008). VADD: Vehicle-assisted data delivery in vehicular ad hoc networks.
Vehicular Technology, IEEE Transactions on, 57(3), 1910-1922.
Zhu, Y., Wu, Y., & Li, B. (2014). Trajectory improves data delivery in urban vehicular networks. Parallel and Distributed Systems, IEEE Transactions on, 25(4), 1089-1100.
