Read IJIPM-4.pdf text version

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory

Emad I Abdul Kareem, School of Computer science, Universiti Sains Malaysia (USM), [email protected] Aman Jantan School of Computer science, Universiti Sains Malaysia (USM), [email protected]

doi: 10.4156/ijipm.vol2. issue2.4

Abstract

The traffic in urban areas is mainly regularized by traffic lights, which may contribute to the unnecessary long waiting times for vehicles if not efficiently configured. This inefficient configuration is unfortunately still the case in a lot of urban areas, where most of the traffic lights are based on a 'fixed cycle' protocol. To improve the traffic light configuration, this paper proposed monitoring system to be as an additional component (or additional subsystem) to the intelligent traffic light system, this component will be able to determine three street cases (empty street case, normal street case and crowded street case) by using small associative memory. The proposed monitoring system is working in two phases: training phase and recognition phase. The experiments presented promising results when the proposed approach was applied by using a program to monitor one intersection in Penang Island in Malaysia. The program could determine all street cases with different weather conditions depending on the stream of images, which are extracted from the streets video cameras. In addition, the observations which are pointed out to the proposed approach show a high flexibility to learn all the street cases using a few training images, thus the adaptation to any intersection can be done quickly.

Keywords: Computer Vision Systems, Vehicle Detection, Intelligent Traffic Light Monitoring System,

Neural Network and Associative Memory.

1. Introduction

Transportation research's goal is to optimize transportation flow of people and goods. As the number of road users constantly increases while resources provided by current Infrastructures are limited, intelligent control of traffic will become a very important issue. Traffic in the urban areas system regularized by traffic lights, which is in many cases contribute to the unnecessary long waiting times for cars if not efficiently configured [13]. The conventional traffic light control methods include fix-time control, time of day control, vehicle actuated control, semi-actuated control, green wave control, area static control and area dynamic control. However, there is no system meeting the adaptive characteristic. This is because the traffic control system is non-linear, fuzzy and nondeterministic, and thus traditional methods of modeling and control cannot work very well. In order to solve the above mentioned problem, there are many researchers gropes have performed a lot of researches. In recent years the application of image processing techniques in automatic traffic monitoring and control has been investigated to optimize methodologies for traffic. Traditionally, the traffic problems has been managed by using the Trial-And-Error method. For example an expertise or team decided on traffic parameters and depending on the resulting traffic behavior some feedback corrections will be done. This philosophy hasn't changed so much in past decades, except for the use of simulators instead of real traffic tests as feedback source. Recently some micro-simulators ­ based on the vision of traffic as a collection of independent vehicles ­ have been proved to be very accurate. Traffic optimization comprises a set of different problems, from which one of the most relevant ones is the traffic light cycles optimization. This paper aim to propose an intelligent traffic light additional component that is a monitoring system to improve traffic light configuration. It will be in the intelligent traffic light system. Thus, an

- 23 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

intelligent traffic light system will be composed of two components(or subsystems): a monitoring system and a control system will able to integrated together to aid the traffic light system to take the intelligent decisions efficiently. This research will be investigating on the additional component (monitoring system) by using small associative memory. This system will be able to determine three street cases (empty street case, normal street case and crowded street case). The input of this system will be a sequence of images of vehicles of the intersection to be monitored. Technically, the proposed approach has two phases: training phase and recognition phase by use one video camera to every street in the intersection. These cameras fixed in a suitable position regardless camera distance from the street zebra crossing to obtain the best possible view. The proposed approach will use multi-connect architecture associative memory [12][11] which is developed from Hopfield associative memory to learn all the street cases in the intersection to recognize the case of the street during recognition phase.

2. Related works

This subsection provide a survey of the literature related to traffic light systems, highlighting most of the traffic light models (i.e., queue traffic light, fuzzy traffic light, Petri-net traffic light and LED traffic light) that were developed to improve traffic light efficiency.

2.1. Queue traffic light model

The queue traffic light model was developed in traffic engineering studies. Vehicles arrive at an intersection controlled by a traffic light and form a queue. Many researchers evaluated the queue lengths in each lane using different techniques depending on street width and the number of vehicles that are expected at a given time of day. In this model, traffic light efficiency is effected when unexpected events happen (traffic accidents) causing disruption to the flow of vehicles. For example, Fathy and Siyal (1995) proposed a queue detection algorithm that consists of motion detection and vehicle detection operation. Both are based on extracting the edges of the scene to reduce the effects of variations in lighting conditions [15]. Jin and Ozguner described (1999) a combination of multidestination routing and real time traffic light control based on a concept of cost-to-go to different destinations. This traffic light model is also a decentralized control approach [29]. A general formulation for delays on a road section was developed by De Schutter (1999). He discussed how optimal and suboptimal traffic light switching schemes can be determined. First, they constructed a model that described the evolution of queue length (as continuous variables) in each lane [7]. Next, he showed how optimal and suboptimal traffic light switching schemes (with possibly variable cycle lengths) can be determined. Xiaohua and Yangzhou (2003) introduced a hybrid optimization system. They used the average queue length over all queues as an objective function to find an optimal traffic light switching scheme [57]. Leeuwaarden (2006) derived a probability generation function of both the queue length and delay from which the whole queue length and delay distribution could be obtained. This allowed for the evaluation of performance characteristics other than the mean, such as the variance and percentiles of distribution [39]. A non-cooperative approach, which gives rise to a non-cooperative game, was studied by Alvarez and coworkers (2008). In this model, signalized intersections were considered as finite controlled Markov chains and a solution was sought to optimize the congestion into an avenue. Using a game theory method, they considered each intersection as a non-cooperative game where each player tries to minimize its queue [3]. Helbing and Mazloumian (2009) discussed elements of signal control based on the minimization of overall travel times or vehicle queues. They found different operation regimes, some of which involve a "slower-is-faster effect", where delayed switching reduced the average travel time. These operation regimes characterized different ways of organizing traffic flows in urban road networks. Besides the optimize-one-phase approach, they discussed the procedure and advantages of optimizing multiple phases as well. To improve the service of vehicle platoons and support the self-organization of "green waves", considering the price of stopping newly arriving vehicles was proposed [19].

- 24 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

2.2. Knowledge based traffic light models

Knowledge based systems are artificial intelligent tools that work in a narrow domain to provide intelligent decisions with justification. Knowledge is acquired and represented using various knowledge representation rules, frames and scripts. Many researchers have used knowledge based systems to developed traffic light systems. For example, Findler and coworkers (1997) described a distributed, knowledge-based system for real-time and traffic-adaptive control of traffic signals. The first of a two-stage learning process optimizes the control of steady-state traffic at a single intersection and over a network of streets. The second stage of learning deals with predictive/reactive control in responding to sudden changes in traffic patterns [16]. Wen (2008) proposed a framework for a dynamic and automatic traffic light control expert system. The model adopts inter-arrival time and interdeparture time to simulate the arrival and leaving number of cars on roads. This model used a knowledge base system and rules. Depending on the traffic light data, which are collected by a RFID reader, this model makes decisions that are needed to control the intersections [52]. Other researchers have developed fuzzy control systems, which are mathematical systems that analyze analog input values in terms of logical variables that take on continuous values between 0 and 1 based on fuzzy logic. This type of control system consists of three components. First, fuzzy elements have degrees of membership. Second a membership function is created, which is a curve that defines how each point in the input space is mapped to a membership value (or degree of membership) between 0 and 1. Third, if-then rules are applied, which are used to formulate the conditional statements that comprise fuzzy logic. In general, many theoretical papers on control of traffic systems using fuzzy statements have been published. For example, Kaur and Konga (1994), described the design of a fuzzy traffic light controller at the intersection of two streets that changes cycle time depending upon the densities of cars behind green and red lights and the current cycle time. A fuzzy model of the system has been built and tested to predict the behavior of the model under different traffic conditions [31]. A realistic approach to fuzzy control of urban road traffic lights was described by Hoyer and Jumar (1994a) [23]. Fuzzy application is encouraging since the design procedure of vehicle actuated traffic light systems is very transparent, and an adaptation to the changing situations of traffic is easy to accomplish [24]. Khalid and coworkers (2004) proposed a fuzzy traffic light controller to be used at a complex traffic junction. The proposed fuzzy traffic light controller is capable of communicating with neighboring junctions and manages phase sequences and phase lengths adaptively. Average flow density, average delay time and link overflow of the intersections are used as performance indices for comparison purposes [32]. GiYoung and coworkers (2001) created an optimal traffic signal using fuzzy control. Electro sensitive traffic lights have better efficiency than fixed preset traffic signal cycles because they are able to extend or shorten the signal cycle when the number of vehicles increases or decreases suddenly. They used fuzzy membership function values between 0 and 1 that estimate the uncertain length of a vehicle, vehicle speed and width of a road. They stored different kinds of conditions such as car type, speed, delay in starting time and the volume of cars in traffic [17]. A traffic light controller based on fuzzy logic was proposed by Kulkarni and Waingankar (2006) to be used for optimum control of fluctuating traffic volumes, such as over saturated or unusual road conditions. The rules of a fuzzy logic controller are formulated by following the same protocols that a human operator would use to control the time intervals of the traffic light. The length of the current green phase is extended or terminated depending upon the 'arrival'; i.e. the number of vehicles approaching the green phase, and the 'queue' that corresponds to the number of queuing vehicles in red phases [35]. Pedraza and coworkers (2008) detailed the design of a traffic system model for vehicles that examined the traffic traveling through a series of traffic lights on a main road. The adaptive network-based fuzzy inference system was used to synchronize the time of duration and phase angle of the traffic lights, and also maintain the maximum possible velocity of the vehicles traveling on the road [44]. A complex adaptive system (CAS) is a network of communicating, intelligent agents where each agent adapts its behavior in order to collaborate with other agents to achieve overall system goals. The overall system often exhibits emergent behavior that cannot be achieved by any proper subset of agents

- 25 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

alone. The classifier event action block can implement both crisp and fuzzy rules. This system uses one network of traffic light controller agents at each intersection. Each traffic controller agent uses a fuzzy classifier block to make decisions about traffic light timing in order to minimize local vehicle wait time. Hong and coworkers (2001) presented concepts that described a main urgent phase and minor urgent phase. The traffic data are acquired from the detectors in the intersections and lanes. Based on the concepts of main and minor urgent phases, a set of novel fuzzy control rules is developed to control the phases and delay of traffic lights according to the dynamic characters of some correlative traffic intersections [21]. An electro sensitive traffic light, using a smart agent algorithm to reduce traffic congestion and traffic accidents, was proposed by Hong and coworkers (2007). Specifically, they designed and implemented a system to create optimum traffic signals in congested conditions using smart agent algorithms. This approach antecedently created an optimal traffic cycle of passenger car units at the bottom traffic intersection. Mistakes were possible due to different car lengths, car speeds, and the length of the intersection. Therefore, this approach consequently reduced car waiting times and start-up delay times using fuzzy control of feedback data [22]. An application of diffuse systems in traffic lights for the road control of urban transit was proposed by Lopez and coworkers and Alejandro and coworkers (2007). Given vehicular problems of the city, it was intended to look for options to make vehicular traffic more agile. With this in mind, three proposals for diffuse control design were formulated. The first proposal was the control of two traffic lights for cars placed in a crossing of a few streets. The functioning of the traffic light was typical (green-amber-red). The system included sensors entrusted to indicate the pace of arrival of the cars and the length of the trail of cars at a certain moment. The principal street had one sensor and the lateral street had another sensor. Second, the proposal had, as a basic principle, the modification of the traffic light timing around a predetermined nominal value. Such a nominal value was calculated based on normal traffic conditions in a determinate cross road, using standard traffic theory and criteria. The objective of the fuzzy controller was to dynamically adjust the timing of each light stage to support variations in vehicular load, such as during rush hours. Finally, the third proposal was to optimize the flow of vehicles in the street. This was carried out by defining the times that each light of the traffic light remained lit. This proposal has a fuzzy inference system (FIS) control where the input variables for the control are car density and waiting times[2]. The weakness of all these approaches is the fact that the systems use expert system technology but do not provide any guarantee about the quality of the rules; i.e., an expert system/rule based approach is not ideal for problems that require considerable knowledge. In addition, easily creating and modifying rules can destroy any system. A knowledgeable user can add no value rules or rules that conflict with existing ones.

2.3. Petri net traffic light models

Petri net (PN) models consist of places (graphically represented as circles) and transitions (graphically represented as bars) connected via a set of directed arcs. Places may contain tokens (represented by dots inside the circle) that move through the network (i.e., from place to place) according to certain rules. PN models have been used as a tool for various kinds of discrete event systems, simulation and control logic. However, these models have disadvantages. The main disadvantage is that PN models are quite primitive, not only in that a significant burden is placed on the analyst, but also the graphical representation may become too complex to be useful and there is an inability to model similar processes using one graphical representation. This means, in traffic light cases, each intersection would have its own PN graph representation. Another disadvantage is that the representation of priorities or ordering is hard to manage, although priority queues are important in performability modeling. A PN model can be applied to traffic control. For example, Di Febbraro and coworkers provided a modular representation of urban traffic systems regulated by signalized intersections. Considering such systems to be composed of elementary structural components; namely, intersections and road stretches, the movement of vehicles in the traffic network is described with a microscopic representation and is realized via timed PNs. An interesting feature of the model is the possibility of representing the offsets

- 26 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

among different traffic light cycles as embedded in the structure of the model itself [9]. Li et al. (2003) [40] proposed an approach via a programmable logic controller (PLC) and PN synthesis depending on the fact that a network of traffic lights regulating the urban vehicle can be viewed as a complex discrete-event system (DES). It was shown that PNs are one of the methodologies for modeling, analyzing and controlling a DES. First, they described typical urban traffic light control using PNs. Having compared three methods for PN realization, the best implementation method for a traffic lights control system was one that was made with a PLC. The sequential function chart (SFC) of the PLC system was also presented in the thesis. An urban network of signalized intersections can be suitably modeled as a hybrid system in which the vehicle flow behavior is described by means of a time-driven model and traffic light dynamics are represented by a discrete event model. Di Febbraro et al. (2004) used a model of such a network via hybrid PNs to state and solve the problem of coordinating several traffic lights with the aim of improving the performance of some classes of special vehicles; i.e., public and emergency vehicles [10]. Yi-Sheng et al. (2006b) modeled such a network via timed colored PNs (TCPNs) to state and solve the problem of coordinating several traffic lights with six and two phases. Moreover, the analysis of control TCPN models is performed using an occurrence graphs (OG) method. The relation of the liveliness and reversibility of the control TCPN is thus obtained [61]. Yi-Sheng et al. (2006) developed a model for a sophisticated urban traffic light control system for nine crossroads. Variations in vehicle directions in crossroads were allowed with more cases of transitions for each crossroad [60].

2.4. Traffic light models based on wireless communication

Wireless communication is the transfer of information over a distance without the use of enhanced electrical conductors or wires. The distances involved may be short (a few meters, as in television remote control) or long (thousands or millions of kilometers for radio communications). There are many problems with systems that use wireless communication. One of the problems is the already limited spectrum available for communications. Another problem is splitting up the environment into a number of small cells, which increases the overall accessible bandwidth of the communication system, but also increases the cost as more cell sites are required. Some form of encryption is required for communications to avoid interception of data transmitted over the network by devices not taking part in the communications. In addition to security considerations from external devices accessing the network, interfering signals can be generated by other devices in the environment. These devices can temporarily disrupt a communication link through the noise that they generate. Other researchers used a wireless sensor network to develop such a system. For example, Ibrahim et al. (2005) developed software planning tools for wireless LAN link optimization as an intelligent traffic light system control. This software was based on the combination of Mat Lab and MapInfo software, which gives grouping parameters to build up the software development. The traffic light site selections must include line-of-sight (LOS) field strength predictions for either point to point or point to multipoint situations [25]. Miguel et al. (2006) proposed the use of a wireless sensor network to enable car drivers to have more energy efficient city driving via finding an interaction communication between drivers and traffic lights [42]. Gradinescu et al.(2007) presented an adaptive traffic light system based on wireless communication between vehicles and fixed controller nodes deployed in intersections. A smart transport and road communications model was presented by Kun-chan et al. (2007) based on a wireless mesh network architecture, connecting a hierarchy of several thousand devices, from individual traffic light controllers to regional computers and the central traffic management centre (TMC), that placed stringent requirements on the reliability and latency of the data exchanges [36]. Tubaishat et al. (2008) studied the performance with one sensor and two sensors and designed corresponding controllers. In the case of one sensor, two models were developed: the first one detected passing vehicles only; whereas, the second one detected vehicles that passed the sensor or stopped at it. In both methods, the change of the sensor location was relative to traffic light location. They then used two sensors to calculate the number of vehicles waiting or approaching a traffic light. They tested different distances between the two sensors. Researches who use wireless sensor networks face serious

- 27 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

challenges, such as the problem of providing power to such sensors, in addition to the difficulty of determining the distance between the sensors in the case of needing more than one sensor [49].

2.5. LED traffic light models

Light-emitting diodes (LEDs) are semiconductor devices that are capable of fast switching ON and OFF. This means LEDs can be used for communication purposes. In addition to the normal function of being an indication and illumination device, LED traffic lights can also be used as transmitters. Although, research on wireless optical communication using LED traffic lights has been performed, this type of system would require replacement of all traffic lights. In addition, receivers (e.g., cameras) would need to be located in the front of vehicles. Ibrahim and Beasley (1998) discussed the technical aspects of LED traffic lights and provided estimates on expected savings if all the traffic lights were to be replaced by LEDs [26]. Akanegawa et al. (2001) proposed a traffic information system using LED traffic lights, focusing on light visibility and power used for traffic control, the number and location of traffic lights, and movement toward use of LED traffic lights [1]. Finally, Wada et al. (2005) proposed a parallel wireless optical communication system for road-tovehicle communication that uses a LED traffic light as a transmitter and a high-speed camera as a receiver. The proposed system enables multi-channeling in a two dimension arrangement and spatial dividing ability. LED transmitters arranged in the shape of a plane are modulated individually and a camera is used as a receiver for demodulating the signals by using image processing techniques [51].

2.6. Traffic light models using an extension neural network (ENN)

The extension neural network (ENN) consists of extension theory and a neural network that uses a modified extension distance (ED) to measure the similarity between data and a cluster center. ENN is another traffic light control system developed to deal with object recognition in outdoor environments [6]. In outdoor environments, lighting conditions cannot be controlled or predicted, objects can be partially occluded, and their position and orientation is not known a priori. The chosen objects are traffic or road signs, due to ease of sign maintenance and inventory in highways and cities, driver support systems and intelligent autonomous vehicles. A genetic algorithm is used for the detection step, allowing invariance localization to changes in position, scale, rotation, weather conditions, partial occlusion, and the presence of other objects of the same color. A neural network can achieve classification [6]. Kuie et al. (2008) and chao et al. (2009) have presented an intelligent traffic light control method based on extension theory for crossroads. First, the number of passing vehicles and maximum passing time of one vehicle within the green light time period are measured in the main-line and sub-line of a selected crossroad. Then, the measured data are adopted to construct the extended matter-element model and accordingly the correlation degrees are calculated for recognizing the traffic flow of a standard crossroad. Some experimental results were obtained to verify the effectiveness of the proposed intelligent traffic flow control method. The diagnostic results indicated that the proposed estimated method can discriminate the traffic flow of a standard crossroad rapidly and accurately [34] and [5]. These researchers, however, did not take into account unexpected situations that may cause disruption in the flow of vehicles, where the ENN is used for estimation. Also, the ENN tends to be slower to train than other types of networks (e.g., a single layer neural network) for two reasons. First, a large number of iterations are needed to finish learning all the prototypes. Second, the large amount of data needs large networks. Therefore, the network size should be as small as possible to allow for efficient computations. Sometimes, reducing the size of these data sets leads to ignoring some factors that could improve the estimation process for the flow of vehicles.

2.7. Agent based traffic light models

A number of researchers turned to developing agent based traffic light models. For example, an agent based approach for traffic light control was adopted by Hirankitti and Krohkaew (2007).

- 28 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

According to this approach, the system consists of agents and their world. In the traffic context, the world consists of cars, road networks and traffic lights. Each of these agents controls all traffic lights at a road junction by an observe-think- act cycle. That is, the agent continuously observes the current traffic conditions by collecting traffic data, and then the data is used for reasoning with the traffic-lightcontrol rules by the agent's inference engine to determine how a signal will be changed on each traffic light near each junction. Use of inference engine techniques requires the use of a knowledge base. Therefore, using a large knowledge base will have an effect on the efficiency of the traffic light system because of storage space and the time it would take for decision-making, in addition to the quality of rules for decision-making [20]. Although, a penalty for converting electronic signals to optical signals (to realize optical interconnects) and vice-versa must be paid, because of the clear differences between the speed of the data communication and processor speed, many researchers have continued interest in the marriage between photonics and electronics. For example, an agent based traffic lights logic algorithm, developed by Krajzewicz et al. (2005) used the length of a jam in front of a traffic light as input along with information from the optical information system (OIS) sensors. This research observes the incoming lanes and measures the jam lengths on these lanes. If at one of these lanes the jam gets longer, this lane gets a green for a longer time [33].

2.8. Reinforcement learning traffic light models

Researchers have used reinforcement learning to improve traffic light configurations, which is a sub-area of machine learning concerned with how an agent ought to take actions in an environment to maximize some notion of long-term reward. For example, Wiering et al. (2004) studied simulation and optimization of traffic light controllers in a city and presented an adaptive optimization algorithm based on reinforcement learning. They implemented a traffic light simulator to experiment with different infrastructures and to compare different traffic light controllers [54]. Nijhuis et al. (2005) described an existing approach of reinforcement learning applied to the optimization of traffic light configurations. This approach used implicit cooperation between traffic lights while letting cars take into account the traffic situation of the road ahead [43]. These two researches used different knowledge representation, dependent on the reinforcement learning machine. Also, the action value function, which is difficult to analyze especially in a case of a large domain, was different. The rules are usually much easier to interpret unless there are too many of them. Another disadvantage of these two approaches was the necessity of tuning parameters, which meant the choice of these parameters had a high impact on the results. Finally, due to necessary exploration (exploring the environment), the algorithm's performance is less stable.

2.9. Genetic algorithm traffic light models

A genetic algorithm (GA) is an evolutionary algorithm. GAs have been presented since the early 1960s and they apply the rules of nature, such as evolution occurs through selection of the fittest individuals and individuals can represent solutions to a mathematical problem. Some researchers have used GAs to improve traffic light configurations. For example, Sanchez et al. (2008) presented a new architecture for the optimization of traffic light cycles in a traffic network. The model is based on three basic design items: the use of genetic algorithms as an optimization technique, the use of cellular automata simulators within the evaluation function, and the use of a cluster as a parallel execution environment for this architecture [45]. Although, a great advantage of GAs is the fact that they find a solution through evolution, this is also the greatest disadvantage. Evolution is inductive. In nature, life does not necessarily evolve towards a good solution; it can evolve away from bad circumstances. This can potentially cause a species to evolve into an evolutionary dead end.

- 29 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

2.10. Vision traffic light models

Video sensors have become particularly important in traffic applications, mainly due to their fast response and easy installation, operation and maintenance. They also have the ability to monitor wide areas. Serrano et al. (2005) designed, constructed and tested an artificial-vision controlled traffic-light prototype to rule and regulate intersections. Methods, algorithms and automatons were created to provide analysis of images and decision making in real time. They developed an intelligent trafficlight capable of capturing the presence or absence of vehicles, pedestrians and their particular situations defined by their trajectories without taking into account the street conditions; for example, street crowding or not [47].

2.11. Other traffic light models

Researchers have developed other traffic light models using various techniques. For example, Viera et al. (2000) represented the traffic light as an example of an object that exhibits a rich behavior set and serves as a case study for a number of interesting design issues. It was implemented using a traffic light internal state and corresponding control information that was constructed using sensors [50]. Wu and coworkers (2006) proposed control of traffic lights in a simple intersection by taking into account individual vehicle arrival behavior to distinguish different kinds of vehicles, such as public transport vehicles and emergency vehicles [56]. As mentioned previously, the use of sensors has a number of limitations. For example, there is a power problem and distance determination problem with two or more sensors. A branch and bound approach to control traffic lights in a simple intersection was proposed by Yan and coworkers (2008). It was based on new technologies for recognizing vehicles and traffic lights that allow distinguishing different kinds of individual vehicles such as public and emergency vehicles [58]. These two proposed systems seem to be relatively slow and complicated because more techniques may be necessary to identify various objects that appear in the scene of a traffic light. Sanchez et al. (2004) designed a traffic light cycle evolutionary optimization architecture that attempted to validate results with a real-world test case [46]. An optimal switching time sequence for a traffic signal controlled intersection was designed by De Schutter (1999) using an extended linear complementary problem, which is a mathematical programming problem. He considered the determination of the optimal switching time (switching sequences are acyclic, but the phase sequence is pre-fixed). Also, he discussed some approximations that lead to suboptimal switching time sequences that can be computed very efficiently, and for which the value of the objective function is close to the global optimum [7]. The system was not feasible in practice due to its computational complexity. In addition, it did not take into account the intersection road conditions. A methodology to model a traffic light system was proposed by Yi-Sheng et al. (2007). The modeling tool was called state chart. State chart has been utilized as a visual formalism for the modeling of complex systems [59]. Wong and Woon (2008) presented a method for optimizing traffic timing plans via the use of clustering algorithms to automatically generate time-of-day (TOD) intervals. By detecting time intervals that share common traffic conditions, it was possible to obtain TOD intervals that better reflected the underlying generators of traffic patterns [55]. However, the abovementioned procedure suffered from a problem; i.e., the data that was used to find the clusters was generated using the original traffic patterns, which meant that the estimates of the true underlying TOD intervals could be inaccurate. A refinement of the method was presented that was based on iteratively re-estimating the TOD intervals. An adjacent intersections traffic light coordinated control model, based on mixed integer programming formulation, was proposed by Yuan et al. (2008). At a given moment, it was possible to compute the departure rate in the upstream intersection. Then, by switching the green light time, they could control the arrival rate of the downstream intersection [62]. Mayukh and Theresa (2009) applied rough set theory to show the rough relationships between historical and current traffic pattern data and they proposed a traffic signal system based on the patterns of historical data. Using the application of probability, the system forecasts the traffic loads on lanes and assigns higher credits to a particular case based on the probabilistic values of historical data [41].

- 30 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

An intelligent traffic light controller based on very-high-speed integrated circuit hardware description language (VHDL) was given and simulated by Shi et al. (2009) using hierarchical design thought. VHDL is a hardware description language used in electronic design automation to describe digital and mixed-signal systems, such as field-programmable gate arrays and integrated circuits [48]. The above methods failed to take into account the exceptional cases that may occur unexpectedly and that affect the flow of vehicles in the road, except for the last method, which did not fully explain the way in which the proposed intelligent traffic light system extracted the data.

3. An intelligent traffic light monitor system using an adaptive associative memory

The proposed monitoring system has two components: Images analyzer and MCA associative memory. The images analyzer has two phases: training phase and recognition phase which are used the associative memory.

3.1. Multi-connect architecture associative memory

This associative memory is a simple single-layer neural network; it can learn a set of pattern pairs (or associations). An efficient associative memory can store a large set of patterns as memories. During recall, the memory is excited with a key pattern (also called the search argument) containing a portion of information about a particular member of a stored pattern set. This particular stored prototype can be recalled through association of the key pattern and the information memorized [27], [37]. Experimentally, associative memory has critical problems because most of the conventional associative memories use Hebbian learning [38]. Since the learning algorithm is based on the vector outer products, it does not guarantee the recall of all training data, unless the training vectors are orthogonal. As a result, associative memories learned by Hebbian learning suffer from extremely low memory capacity. To improve the associative memory and avoid most of the associative memory limitation,[11]and[12] developed a new associative memory named Multi-Connect Architecture (MCA) associative memory. This improving was done by proposed new algorithms for learning process and convergence process. Thus for both , the pattern (pattern: It means a sequence of 1's and 0's) will divide into a number of parts with size three, to be considered as a vector V (each three element of the pattern will be one vector). Each one of these vectors needs to create its learning weight matrix during learning process or need to find the convergence pattern during the convergence process (see figure 1) [11][12].

Figure 1. The data (pattern) divided into a number of vectors with size three, which it need to create its learning weight matrix Because of this new process, MCA can deal with any pattern size and the associative memory capacity became unlimited, and it could remember even the correlation patterns. Additionally, because the size of the vectors is three, there are no more than eight possible vectors (see table 1), this means there are no more than eight weight matrices W will be built during learning process. These matrices are symmetric , zero diagonal and with size 3*3 [13][11].

- 31 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

Table 1. Illustrated the eight possibilities of the binary vector with length three. The architecture of MCA is illustrated in (figure 2). It shows each path represents one learning weight matrix (1< m<8), thus, all the vectors in the pattern will be replacing with a number, which represents the number of the path in the net, by this number, we can call the path again [11][12].

Figure 2. The architecture of MCA associative memory Matrix W, which is sometimes called the connectivity matrix, is an n × n matrix containing network weights arranged in rows of vectors. The weight matrix is symmetric, i.e., wij= wji, and with diagonal entries equal explicitly to zero, i.e., wii= 0. In other words, no connection exists from any neuron back to itself. Physically, this condition is equivalent to the lack of the self-feedback in the nonlinear dynamical system. If the diagonal elements were not 0, the net would tend to reproduce the input vector rather than a stored vector [11], [27]. A network can define various patterns; one can find them by different start vectors in the iteration. Corresponding to the spin-glass theory of solid state physics, such as equilibrium functions in this net are characterized by the fact, that the total energy (Hamilton function) becomes minimum. This leads here to a "Lyapunov function or Energy function" (see equation (1)), which becomes exactly minimum, when one creates a pattern. This energy function can be defined as follows [53].

Where:

n is the number of elements in the vector v. Wij is the weight from the input of neuron i to the output of neuron j. t limiting (threshold) value, which equal to zero in this net.

One can calculate the energy function E for every input vector, which can be created in the network. If one calculates the function E for all the possible input vectors, an energy landscape with maximums and minimums can be obtained. The point is that the minimum is taken when the input is a pattern [12].

3.2. Images Analyzer

The proposed monitor approach uses one video camera in every street in the intersection. The video cameras will be installed in a suitable position regardless its distance from the zebra crossing to obtain the best possible view. Each camera monitors the queue of cars in the street then it will pass a stream of

- 32 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

images to the monitor system which it will convert these images to binary images using one of the classic filters. These images will be used as a pattern, which reflected all the street cases( different time and weather conditions). The proposed monitor approach will use these images as a training image during training process and save it in MCA associative memory to be remembered during recognition process. This approach has two phases: training Phase and recognition phase. Training Phase: In this phase, training process will be implemented to learn the entire three street cases, the ability of recognize these cases will be acquired. To monitor any street in an intersection, there are three cases: 1. Empty street case: The street is empty; this means (no cars in this street). 2. Crowded street case: The street is crowded; this means (traffic jam). 3. Normal street case: The street is normal, this means (not crowded and not empty but normal). Figure (3) illustrated the flowchart of the training phase. In training phase it will be used training images to learn street cases. For each case, training process use at least one training image. Training images will be converted to binary images (black and white images) using one of Edge detection filters. Training process will be implemented to learn the street cases using MCA associative memory to save all the training images which they reflect all the excepted street cases.

Figure 3. Illustrated the flowchart of the Training phase. Where recognition phase will be implemented to monitor an intersection depending on the previous training process to recognize all the three street. Depending on recognition phase flowchart (see figure 4), each street camera in the intersection sends its own video stream. This video will be converted to stream of JPG images. The same with the previous phase, the system converts these images to black and white images using one of the classic filters to determine the case of each street in the intersection. These determinations done after these images converge with the training images which are learned and saved in the associative memory.

- 33 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

Figure 4. Recognition phase flowchart of proposed system (For vehicles)

4. Experiments and decisions

These experiments are to illustrate this approach and improving it by writing simulation program. This program wrote in Delphi program language. Empirically, in these experiments, we have 2571 JPG images for night and 2734 images of a day, which are extracted from eight video samples. These video samples are for one of Penang Island intersection in Malaysia. The intersection had two types of shooting: day shooting and night shooting. These shooting taking into account all possible weather conditions( i.e. for day shooting have sunny, cloudy and rainy weather conditions and for night shooting have Normal and rainy weather conditions). In addition, it is possible for these images to be noisy images (perhaps some pedestrian, bicycles, motorbike, etc.). As a mentioned before, image representation or coding scheme is one of two important components, thus to the baste representation scheme to the stream of images, which are constructed from video samples, this research used many classic image filters to convert the JPG color images to black and white images. With night shooting just seven filters were succeeding (see table 2 and figure 5).

Table 2. Number of training images corresponding to each classic filter for night shooting with two expected weather conditions (Normal and rainy) .

- 34 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

Figure 5. Number of training images vs. to each classic filter for night shooting with two expected weather conditions (Normal and rainy). However, with day shooting, nine filters were succeeding (see table 3 and figure6). All filters used the same images to make a comparison between the number of training images and all the succeed filters.

Table 3. Number of training images corresponding to each classic filter for day shooting with three expected weather conditions (sunny, cloudy and rainy) .

Figure 6. Number of training images vs. to each classic filter for day shooting with three expected weather conditions (sunny, cloudy and rainy).

- 35 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

Logically, there are two weather conditions that affected the work of this approach efficiently, namely: fog and heavy raining conditions. In this case, the approach will recognize these two weather conditions by using just two training images one for fog and one for heavy raining conditions to warn the traffic light control system that the monitor system can't recognize all cases of the intersection's streets, therefore the controller should be change to the traditional traffic light controller. Depending on the principle of using one filter for all the day period time conditions(day and night period times) and for all the weather conditions the previous analysis shows that Hough Transform filter is the beast filter to be use with this proposed approach because it needs the lowest number of training images in comparison to the other filters(see table 4 and figure 6).

Table 4. Illustrated all the filters with all the expected weather and period times conditions vs. the total number of training images for all these conditions. This research, comparing with the previous computer vision approaches has the same principles except using multi-connect architecture associative memory with the proposed approach. Thus not like other approaches, training process will be learned each training image as one segment, instead of using any Vehicle detection, tracking object or even image segmentation techniques. The associative memory gives the proposed approach the ability of adaptation to any intersection. The observations are pointed out to the importance of the accurate choosing of the training images and its number to exploitive the adaptation feature. Based on the fact that each intersection has its own cases, which are repeated every day, the number of the training images will be limited depending on the number of intersection cases. This limitation gives the proposed approach acquirement to adapt to its environment quickly. Logically, training will be implemented just one time using small associative memory to gain the equitable experience.

5. Conclusion

A review and discussion of the research is presented. The developments presented in this thesis are considered improvements to previous work by adding a component to the main concepts. This study focused on the development of an intelligent vision traffic light monitoring system via associative memory in order to demonstrate an improvement in traffic light configurations. Improving traffic light

- 36 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

systems has huge economic value via reducing wasted fuel due to unnecessary waiting times at intersections and the wasted time and lost lives of vehicle users.

6. References

[1] [2] Akanegawa, M., Y. Tanaka and M. Nakagawa. 2001. "Basic study on traffic information system using LED traffic lights." Intelligent Transportation Systems, IEEE Transactions on 2(4):197-203. Alejandro Lopez, J., R. Garcia, A. Garcia Blanco and I. A. Zuniga Felix. 2007. "Traffic Lights Fuzzy Control Proposals to Improve Vehicular Flow." In Electronics, Robotics and Automotive Mechanics Conference, 2007. CERMA 2007. Alvarez, I., A. Poznyak and A. Malo. 2008. "Urban traffic control problem a game theory approach." In Decision and Control, 2008. CDC 2008. 47th IEEE Conference on.

[3]

[4] Ayad. M. Turky, M.S. Ahmad, M.Z.M. Yusoff and N.R. Sabar, Genetic Algorithm Application for Traffic Light Control, Information Systems: Modeling, Development, and Integration book, Springer Berlin Heidelberg, part2 ,vol. 20, ISBN:978-3-642-01111-5 (Print) 978-3-642-01112-2 (Online) , 2009.

Chao, Kuei-Hsiang, Ren-Hao Lee and Meng-Hui Wang. 2009. "An Intelligent Traffic Light Control Based on Extension Neural Network." In Knowledge-Based Intelligent Information and Engineering Systems. [6] de la Escalera, A., J. M. Armingol and M. Mata. 2003. "Traffic sign recognition and analysis for intelligent vehicles." Image and Vision Computing 21(3):247-258. [7] De Schutter, B. 1999. "Optimal traffic light control for a single intersection." In American Control Conference, 1999. Proceedings of the 1999. [8] De Schutter, Bart. 2002. "Optimizing acyclic traffic signal switching sequences through an Extended Linear Complementarity Problem formulation." European Journal of Operational Research 139(2):400-415. [9] Di Febbraro, A., D. Giglio and N. Sacco. 2002. "On applying Petri nets to determine optimal offsets for coordinated traffic light timings." In Intelligent Transportation Systems, 2002. Proceedings. The IEEE 5th International Conference on. [10] Di Febbraro, A., D. Giglio and N. Sacco. 2004. "Urban traffic control structure based on hybrid Petri nets." Intelligent Transportation Systems, IEEE Transactions on 5(4):224-237. [5]

[11] Emad Issa Abdul Kaream, Alternative Hopfiled Neural Network With Multi-Connect Architecture, journal of College of Education ,Computer Department , Al-mustansiryah university, Baghdad, Iraq, 2004. [12] Emad Issa Abdul Kaream, k. N. M., Hussein A. Moussa, Gray Image Recognition Using Hopfield Neural Network With Multi- Bitplane and Multi-Connect Architecture, Proceedings of the international Conference on Computer Graphics, Imaging and Visualisation (CGIV'06) IEEE, 2006. [13] Emad Issa Abdul Kaream and Aman Jantan, Intelligent Traffic Light Control Using Neural Network with Multi-Connect Architecture, National Conference on Information Retrieval and Knowledge Management (CAMP08) in Kuala Lumpur,Malaysia, 2008. [14] Feizhou Zhang, Dongkai Yang, Xuejun Cao and Jia Chen, Multi-Agent Design of Urban Oriented Traffic Integration Control System, Frontiers of High Performance Computing and Networking ISPA 2007 Workshops book, Springer Berlin / Heidelberg, Vol. 4743 , ISBN: 978- 3-540-747666, 2007.

[15] Fathy, M. and M. Y. Siyal. 1995. "Real-time image processing approach to measure traffic queue parameters." Vision, Image and Signal Processing, IEE Proceedings - 142(5):297-303. [16] Findler, Nicholas V., Sudeep Surender, Ziya Ma and Serban Catrava. 1997. "Distributed intelligent control of street and highway ramp traffic signals." Engineering Applications of Artificial Intelligence 10(3):281292. [17] GiYoung, Lim, Kang JeongJin and Hong YouSik. 2001. "The optimization of traffic signal light using artificial intelligence." In Fuzzy Systems, 2001. The 10th IEEE International Conference on. [18] Gradinescu, V., C. Gorgorin, R. Diaconescu, V. Cristea and L. Iftode. 2007. "Adaptive Traffic Lights Using Car-to-Car Communication." In Vehicular Technology Conference, 2007. VTC2007-Spring. IEEE 65th. [19] Helbing, D. and A. Mazloumian. 2009. "Operation regimes and slower-is-faster effect in the controlof traffic intersections." The European Physical Journal B - Condensed Matter and Complex Systems 70(2):257-274. [20] Hirankitti, Visit and Jaturapith Krohkaew. 2007. "An Agent Approach for Intelligent Traffic-Light Control." In Modelling & Simulation, 2007. AMS '07. First Asia International Conference on. [21] Hong, Wei, Yong Wang, Xuanqin Mu and Yan Wu. 2001. "A cooperative fuzzy control method for traffic lights." In Intelligent Transportation Systems, 2001. Proceedings. 2001 IEEE.

- 37 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

[22] Hong, You-Sik, Geuk Lee, Cheonshik Kim and Jong Kim. 2007. "Traffic Signal Planning Using a Smart Agent System." In Agent and Multi-Agent Systems: Technologies and Applications. [23] Hoyer, R. and U. Jumar. 1994a. "An advanced fuzzy controller for traffic lights." Annual Review in Automatic Programming 19:67-72. [24] Hoyer, R. and U. Jumar. 1994b. "Fuzzy control of traffic lights." In Fuzzy Systems, 1994. IEEE World Congress on Computational Intelligence., Proceedings of the Third IEEE Conference on. [25] Ibrahim, A. H., M. Ismail, T. S. Kiong and Z. B. K. Mastan. 2005. "Development of software planning tools for an intelligent traffic light wireless communication link using 5.8 GHz WLAN." In Applied Electromagnetics, 2005. APACE 2005. Asia-Pacific Conference on. [26] Ibrahim, D. and M. Beasley. 1998. "The benefits of LED traffic lights in London and the pilot test sites." In Road Transport Information and Control, 1998. 9th International Conference on (Conf. Publ. No. 454).

[27] Jacek M. Zurada Ed, Introduction to Artificial Neural Systems . Jaico publishing house, 1996. [28] Javier Sánchez, Manuel Galán, and Enrique Rubio, Applying a Traffic Lights Evolutionary Optimization Technique to a Real Case: "Las Ramblas" Area in Santa Cruz de Tenerife, IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION, VOL. 12, NO. 1, 2008.

[29] Jin, Lei and U. Ozguner. 1999. "Combined decentralized multi-destination dynamic routing and real-time traffic light control for congested traffic networks." In Decision and Control, 1999. Proceedings of the 38th IEEE Conference on. [30] Jose Alejandro, Lopez, Garcia Rafael, Blanco Alejandro Garcia and A. Zuniga Felix Ismael. 2007. "Traffic Lights Fuzzy Control Proposals to Improve Vehicular Flow." In Proceedings of the Electronics, Robotics and Automotive Mechanics Conference: IEEE Computer Society. [31] Kaur, D. and E. Konga. 1994. "Fuzzy traffic light controller." In Circuits and Systems, 1994., Proceedings of the 37th Midwest Symposium on. [32] Khalid, M., Liang See Chin and R. Yusof. 2004. "Control of a complex traffic junction using fuzzy inference." In Control Conference, 2004. 5th Asian. [33] Krajzewicz, Daniel, Elmar Brockfeld, Jürgen Mikat, Julia Ringel, Christian Rössel, Wolfram Tuchscheerer, Peter Wagner and Richard Wösler. 2005. "Simulation of modern Traffic Lights Control Systems using the open source Traffic Simulation SUMO." Proc. 3rd Industrial Simulation Conf. 2005; Berlin, Germany [34] Kuei-Hsiang, Chao, Lee Ren-Hao and Yen Kun-Lung. 2008. "An intelligent traffic light control method based on extension theory for crossroads." In Machine Learning and Cybernetics, 2008 International Conference on. [35] Kulkarni, G. H. and P. G. Waingankar. 2007. "Fuzzy logic based traffic light controller." In Industrial and Information Systems, 2007. ICIIS 2007. International Conference on. [36] Kun-chan, Lan, Wang Zhe, R. Berriman, T. Moors, M. Hassan, L. Libman, M. Ott, B. Landfeldt, Z. Zaidit and A. Senevirante. 2007. "Implementation of a Wireless Mesh Network Testbed for Traffic Control." In Computer Communications and Networks, 2007. ICCCN 2007. Proceedings of 16th International Conference on.

[37] Laurence Fausett Ed., Fundamental of Neural Networks, Architectures, Algorithms And Applications, Prentice-Hall, 1994. [38] Lior Kuyer, Shimon Whiteson, Bram Bakker, and Nikos Vlassis, Multiagent Reinforcement Learning for Urban Traffic Control Using Coordination Graphs , Machine Learning and Knowledge Discovery in Databases book, pages: 656-671, Springer Berlin / Heidelberg, Vol. 5211, ISBN: 978-3-540-87478-2, 2008.

[39] Leeuwaarden, J. S. H. van. 2006. "Delay Analysis for the Fixed-Cycle Traffic-Light Queue." Transportation Science 40(2):189-199. [40] Li, Lin, Nan Tang, Xiangyang Mu and Fubing Shi. 2003. "Implementation of traffic lights control based on Petri nets." In Intelligent Transportation Systems, 2003. Proceedings. 2003 IEEE. [41] Mayukh, Bit and Beaubouef Theresa. 2009. "A rough set approach for traffic light system with no fixed cycle." J. Comput. Small Coll. 24(4):14-20. [42] Miguel, Sanchez, Cano Juan-Carlos and Kim Dongkyun. 2006. "Predicting Traffic lights to Improve Urban Traffic Fuel Consumption." In ITS Telecommunications Proceedings, 2006 6th International Conference on. [43] Nijhuis, Emil, Stefan Peelen, Roelant Schouten, Merlijn SteingrÄover and Supervised by : Bram Bakker. 2005. "Project Design and Organization of Autonomous Systems : Intelligent Traffic Light Control." http://www.science.uva.nl/~arnoud/education/DOAS/2005/Project2005/IntelligentTrafficLightControl.pdf. [44] Pedraza, L. F., C. A. Hernandez and O. Salcedo. 2008. "Intelligent Model Traffic Light for the City of Bogota." In Electronics, Robotics and Automotive Mechanics Conference, 2008. CERMA '08. [45] Sanchez, J., M. Galan and E. Rubio. 2008. "Applying a Traffic Lights Evolutionary Optimization Technique to a Real Case: "Las Ramblas" Area in Santa Cruz de Tenerife." Evolutionary Computation, IEEE Transactions on 12(1):25-40.

- 38 -

An Intelligent Traffic Light Monitor System using an Adaptive Associative Memory Emad I Abdul Kareem, Aman Jantan International Journal of Information Processing and Management. Volume 2, Number 2, April 2011

[46] Sanchez, J. J., M. Galan and E. Rubio. 2004. "Genetic algorithms and cellular automata: a new architecture for traffic light cycles optimization." In Evolutionary Computation, 2004. CEC2004. Congress on. [47] Serrano, Ángel, Cristina Conde, Licesio Rodríguez-Aragón, Raquel Montes and Enrique Cabello. 2005. "Computer Vision Application: Real Time Smart Traffic Light." In Computer Aided Systems Theory ­ EUROCAST 2005. [48] Shi, Shuo, Tian Hongli and Zhai Yandong. 2009. "Design of Intelligent Traffic Light Controller Based on VHDL." In Knowledge Discovery and Data Mining, 2009. WKDD 2009. Second International Workshop on. [49] Tubaishat, M., Qi Qi, Shang Yi and Shi Hongchi. 2008. "Wireless Sensor-Based Traffic Light Control." In Consumer Communications and Networking Conference, 2008. CCNC 2008. 5th IEEE. [50] Viera, K. Proulx, Raab Jeff and Rasala Richard. 2000. "Traffic light: a pedagogical exploration through a design space." In Proceedings of the fifth annual CCSC northeastern conference on The journal of computing in small colleges. Ramapo College of New Jersey, Mahwah, New Jersey, United States: Consortium for Computing Sciences in Colleges. [51] Wada, M., T. Yendo, T. Fujii and M. Tanimoto. 2005. "Road-to-vehicle communication using LED traffic light." In Intelligent Vehicles Symposium, 2005. Proceedings. IEEE. [52] Wen, W. 2008. "A dynamic and automatic traffic light control expert system for solving the road congestion problem." Expert Systems with Applications 34(4):2370-2381.

[53] Werner Kinnebrock, Neural Network, Fundamentals, Applications, Examples, Galotia publications, 1995

[54] Wiering, Marco, Jelle van Veenen, Jilles Vreeken and Arne Koopman. 2004. "Intelligent Traffic Light Control." Technical report, Dept. of Information and Computing Sciences, Universiteit Utrecht. [55] Wong, Y. K. and W. L. Woon. 2008. "An iterative approach to enhanced traffic signal optimization." Expert Systems with Applications 34(4):2885-2890. [56] Wu, Jia, Abdeljalil Abbas-Turki, Aurelien Correia and Abdellah El Moudni. 2007. "Discrete Intersection Signal Control." In Service Operations and Logistics, and Informatics, 2007. SOLI 2007. IEEE International Conference on. [57] Xiaohua, Zhao and Chen Yangzhou. 2003. "Traffic light control method for a single intersection based on hybrid systems." In Intelligent Transportation Systems, 2003. Proceedings. 2003 IEEE. [58] Yan, F., M. Dridi and A. El Moudni. 2008. "Control of traffic lights in intersection: A new branch and bound approach." In Service Systems and Service Management, 2008 International Conference on. [59] Yi-Sheng, Huang, Lee Shung-Shing and Liu Yung-Kuer. 2007. "A Supervisor of Traffic Light Systems Using Statecharts." In Networking, Sensing and Control, 2007 IEEE International Conference on. [60] Yi-Sheng, Huang, Chung Ta-Hsiang and Lin Ting-Hui. 2006a. "Design and Analysis Urban Traffic Lights Using Timed Colour Petri Nets." In Networking, Sensing and Control, 2006. ICNSC '06. Proceedings of the 2006 IEEE International Conference on. [61] Yi-Sheng, Huang, Chung Ta-Hsiang and Lin Ting-Hui. 2006b. "A New Modeling Methodology of Urban Traffic Lights Based on Timed Coloured Petri Nets." In Systems, Man and Cybernetics, 2006. SMC '06. IEEE International Conference on. [62] Yuan, Erming, Biao Li and Ying Feng. 2008. "Optimal coordinated traffic control of adjacent intersections based on multiple objectives programming techniques." In Control Conference, 2008. CCC 2008. 27th Chinese.

- 39 -

Information

17 pages

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

676897


You might also be interested in

BETA
http://www.whale.to/b/sp/springmeier.html
New Title
SECTION
Microsoft Word - MICROCONTROLLER_EMBEDDED BASED PROJECT LIST.doc