A Smart Hybrid System for Parking Space Reservation in VANET
محورهای موضوعی : Vehicular NetworksFarhad Rad 1 , hadi pazhokhzadeh 2 , hamid parvin 3
1 - yasooj branch,islamic azad university
2 - islamic azad university
3 - islamic azad university
کلید واژه: ANFIS, reservation, VANET, multi-objective genetic algorithm,
چکیده مقاله :
Nowadays, developed and developing countries using smart systems to solve their transportation problems. Parking guidance intelligent systems for finding an available parking space, are considered one of the architectural requirements in transportation. In this paper, we present a parking space reservation method based on adaptive neuro-fuzzy system(ANFIS) and multi-objective genetic algorithm. In modeling of this system, final destination, searching time and cost of parking space have been used. Also, we use the vehicle ad-hoc network (VANET) and time series, for traffic flow predict and choose the best path. The benefits of the proposed system are declining searching time, average the walking and travel time. Evaluations have been performed by the MATLAB and we can see that the proposed method makes a good sum of best cost which is useful and meaningful in a parking space reserved for drivers and facility managers. The simulation results show that the performance and accuracy of the method have been significantly improved compared to previous works.
[1]D. Teodorović and P. Lučić, 2006. Intelligent parking systems. European Journal of Operational Research, vol. 175, pp. 1666-1681.
[2]H. Wang and W. He, A reservation-based smart parking system. in Computer Communications Workshops (INFOCOM WKSHPS), pp. 690-695. IEEE.
[3]H. Zhao, L. Lu, C. Song, and Y. Wu, 2012. IPARK: Location-aware-based intelligent parking guidance over infrastructureless VANETs, International Journal of Distributed Sensor Networks, vol. 2012.
[4]X. Zhang, D. Li, and J. Chen, 2014. Parking Space Reservation based on VANETs, International Journal of Advances in Management Science.
[5]Y. Ma, M. Chowdhury, A. Sadek, and M. Jeihani, 2012. Integrated traffic and communication performance evaluation of an intelligent vehicle infrastructure integration (VII) system for online travel-time prediction, IEEE Transactions on Intelligent Transportation Systems, vol. 13, pp. 1369-1382.
[6]R. P. D. Nath, H.-J. Lee, N. K. Chowdhury, and J.-W. Chang, 2010. Modified K-means clustering for travel time prediction based on historical traffic data, International Conference on Knowledge-Based and Intelligent Information and Engineering Systems, , pp. 511-521.
[7]R. Bauza and J. Gozálvez, 2013. Traffic congestion detection in large-scale scenarios using vehicle-to-vehicle communications, Journal of Network and Computer Applications, vol. 36, pp. 1295-1307.
[1] D. Teodorović and P. Lučić, 2006. Intelligent parking systems. European Journal of Operational Research, vol. 175, pp. 1666-1681.
[2] H. Wang and W. He, A reservation-based smart parking system. in Computer Communications Workshops (INFOCOM WKSHPS), pp. 690-695. IEEE.
[3] H. Zhao, L. Lu, C. Song, and Y. Wu, 2012. IPARK: Location-aware-based intelligent parking guidance over infrastructureless VANETs, International Journal of Distributed Sensor Networks, vol. 2012.
[4] X. Zhang, D. Li, and J. Chen, 2014. Parking Space Reservation based on VANETs, International Journal of Advances in Management Science.
[5] Y. Ma, M. Chowdhury, A. Sadek, and M. Jeihani, 2012. Integrated traffic and communication performance evaluation of an intelligent vehicle infrastructure integration (VII) system for online travel-time prediction, IEEE Transactions on Intelligent Transportation Systems, vol. 13, pp. 1369-1382.
[6] R. P. D. Nath, H.-J. Lee, N. K. Chowdhury, and J.-W. Chang, 2010. Modified K-means clustering for travel time prediction based on historical traffic data, International Conference on Knowledge-Based and Intelligent Information and Engineering Systems, , pp. 511-521.
[7] R. Bauza and J. Gozálvez, 2013. Traffic congestion detection in large-scale scenarios using vehicle-to-vehicle communications, Journal of Network and Computer Applications, vol. 36, pp. 1295-1307.
6
Journal of Advances in Computer Engineering and Technology
A Smart Hybrid System for Parking Space Reservation in VANET
Received (Day Month Year)
Revised (Day Month Year)
Accepted (Day Month Year)
Abstract— Nowadays, developed and developing countries using smart systems to solve their transportation problems. Parking guidance intelligent systems for finding an available parking space, are considered one of the architectural requirements in transportation. In this paper, we present a parking space reservation method based on adaptive neuro-fuzzy system(ANFIS) and multi-objective genetic algorithm. In modeling of this system, the driver’s behavior, final destination, time and cost of parking have been used. Also, we use the vehicle ad-hoc network (VANET) and time series, for traffic flow predication and choose the best path. The benefits of the proposed system are declining searching time, the network traffic and the number of mobile cars to find a parking space. Average and the best value of the walking time, travel time and the cost objective function be shown also. evaluations using the software Matlab has been performed, and we can see the method we proposed make good sum of best cost which is useful and meaningful in parking spaces reservation for drivers and facility managers. 2
Index Terms—VANET, Multi-Objective Genetic Algorithm, ANFIS, Reservation.
I. INTRODUCTION
Finding a parking space by drivers is one of the reasons of increased traffic in major cities. Reserving an empty parking space in a crowded area, especially at peak hours is always time consuming and frustrating for drivers. Lots of strategies based on intelligent transportation technologies have been proposed to solve these problems to improve the traffic and ease congestion. Urban parking management systems can recognize and process any information of the parking in the city to allow drivers to be informed from parking information at any time. However, these systems are not able to guidance and reserve the most parking spaces for each of the drivers. In this regard, offering a new solution for centralized management of parking in city can solving traffic problems and parking space reservation. Also, with real-time traffic information as prior knowledge, we can be improved the effectiveness of path planning which is the same for parking space reservation to inform parking spaces’ managers of the arrival time of their users to increase their utilization rate of parking spaces.
The contributions of our paper are as follow. (1) In this paper, we propose the traffic flow predication model to predict traffic speed. This model is based on adaptive neuro-fuzzy system(ANFIS) which collect traffic information with vehicle ad-hoc network. (2) Reservation the most suitable parking space has been done based on multi-objective genetic algorithm and the prediction model in step 1.
The rest of this article is organized as follow. Section II, an overview of related works has been shown. Section III, ANFIS modeling for traffic flow predication is introduced. Proposed method will be present in section IV. Simulation results are presented in Section V. finally; our conclusions are
drawn in section VI.
II. Related works
An intelligent parking space control system to make “on line” decisions whether to accept or reject a new driver’s request for parking was proposed in[1]. In [2], drivers can base on the budgetary constraints and the final destination taking the most parking spaces available. In this scheme, the price of parking services according to the number of available parking space and density are determined dynamically. The major disadvantage of this approach is that density and traffic conditions are not considered in reserve a parking space and is done by the driver.
A scheme that efficiently has allocated parking spaces in vehicular ad-hoc networks and avoided the competition among the vehicles was proposed in [3]. Real-time traffic information not been seen in this paper. In [4], a parking space reservation method with real-time traffic information based on VANETs will be shown. A Road Side Unit (RSU) has introduced to collect traffic information which utilizes the time-prediction model. Then, a parking space reservation mechanism is proposed according to a dynamic path selection based on real-time traffic information. In [5, 6], a support vector machines (SVM) and K-NN used in predicting travel time. The authors in[7], proposed a novel cooperative technique based on Vehicle-to-Vehicle (V2V) communications and fuzzy logic to detect road traffic congestion.
III. proposed scheme
This paper has proposed a hybrid approach using adaptive neuro-fuzzy system and genetic algorithms for parking space reservation. The proposed scheme predicts traffic flow according to data collection by vehicle traffic on the VANET. All reservation requests are examined in a defined period of time by multi-objective genetic algorithm and fuzzy operators simultaneously. Desired output of algorithm, choose the most suitable parking spaces for all drivers that have reached in the same period so that the time-consuming and the path to the final destination and the cost of parking space for all drivers be minimized. Figure 1, have seen this scheme. In continue, the proposed approach will be explained in details.
1. adaptive neuro-fuzzy system
This paper presents a method to collect data and predict traffic flow using neural networks. The main objective of this step is maximization the amount of information and combine this information to obtain more accurate detailed information. q(t) As the flow of traffic at the time (t-1, t] (t is an integer number). By analyzing historical data traffic can be seen so that the flow of traffic always during the week (every week from Monday to Friday) changes and the weekends are the same. So, three different time series, S1(t) time series of the same daily, S2(t) time series of the same week, S3(t) hourly time series to collect time-series data will be considered.
S1(t), a collection of previously recorded traffic flows at the same time at k1 days ago.
Set S2(t) includes traffic flow recorded at the same time, k2 week before in the same day at the week. For example, to predict traffic flow on a Tuesday, need k2 traffic flow information at the before Tuesday.
S3(t) is previous set of k3 traffic flow, before traffic flow q(t).
Various models can be selected for predicting the three respective time series. Let qi(t) is the predicted value of the time series model i for si(t), i = 1, 2, 3's. In follow, a NN model is used to produce the final prediction.
Where F(.) Is a nonlinear function which be determined by NN trained. There are many ways to apply time series. In this paper, we used MA methods for time series S1(t), S2(t) and S3(t). MA model of rank K is calculated as follow.
Where k is the number of conditions defined for the MA. MA is concerned only with K last period known data. The flowchart in Figure 1 has be shown this step.
Figure 1: Proposed ANFIS
2. Multi-Objective Genetic Algorithm
Multi-objective genetic algorithm performs all of requested parking reservation in a defined period of time simultaneously. Desired output of algorithm, choose the most suitable parking spaces for all applicants that have reached in the same period so that the time-consuming and the path to final destination and the cost of parking space for all drivers is minimized.
3. Fitness Function:
The fitness function should be able to make a good parking model for reservation. A driver will choose the lowest cost and most facilities parking for reservation. Among the facilities listed for each parking, Minimum travel time of the request to place of parking and minimum walking distance from the parking location to the destination. The fitness function of genetic algorithm has three function. Fitness function can be stated as follow.
Where, Pt reflect the cost of parking space, Ta is time interval needed to traverse the path from the request a reservation by the driver until the parking area. Lm Determines the distance between the request for the park to the parking place. Vm shows the average speed in each direction and is measurable according to the shortest route, estimate the speed and density of the paths. Ta, will show the driver walking time from the parking place to the end of destination.
Each gene has been shown with a parking space vehicle and amounts allocated to each gene indicating the number of parking will be reserved for the device. Chromosome length is equal to the number of applicant cars. To create the initial population of chromosomes has been assigned equal weight of the empty parking lots to any of the parking spaces. The flowchart presented in Figure 2 shows how this process is done. flowchart of the Proposed method be shown in figure 8.
Figure 2: The process of making initial population
4. Crossover
For generating the children of the two parental chromosomes, natural crossover operator is used. In this paper, to perform crossover, multi-crossover method is used. After the recombination, the number of parking spaces that will be allocated per child is checked so that the capacity does not exceed. As a further innovation in this article is combined number of points using fuzzy inference system for doing. The current generation number and variable changes in the best solution of the previous iteration is inputs in the fuzzy inference system. The output of this system is fuzzy numerical that determine the number of points in a predetermined range. The general model of proposed fuzzy inference system for crossover operation is shown in Figure 3. Membership functions for each input and output variables of this system is shown in Figure 4.
Figure 3: Crossover operation
Figure 4: Membership functions
5. Mutation
In this paper, to create random answers of the entire search space available, a random mutation operator is used. Randomly, selected a set of genes from chromosomes and changed parking numbers can be allocated also. For most operations, speed mutation or mutation operations, will not be considered fix. This variable is calculated using fuzzy system designed. The input fuzzy system is the current generation number and variable changes in the previous iteration of the best solution. The output is a fuzzy number that determines the number of operations mutation. In figure 5, the membership functions for each input and output variables is shown.
Figure 5: Membership functions
IV. simulation
The speed of traffic flow contains 3000 records in 15-minute intervals from a city region is considered for testing system. Data is collected on http://data.gov.uk/dataset Dataset.
In Figure 6, the test result of proposed ANFIS fuzzy neural systems using training data is shown. In the experiment, test data as training data. While the error rate is zero, we expected the output is fall on the training data.
Figure 6: The proposed ANFIS system testing using training data
1. Performance evaluation of the proposed ANFIS system using dataset
In this section for a better review, the proposed neural fuzzy systems using training data will be evaluated. In Figure 7, the result is shown. As can be seen in this figure, in most cases, the predictions of adaptive neuro-fuzzy systems that are shown in green color has been adapted on the data recorded that are shown in blue.
Figure 7: The proposed ANFIS system testing using training data
2. Parking reservation system efficiency
to evaluate the effectiveness of the proposed system, we test proposed system using the information that has been predicted using ANFIS system. To check the road network system area of 2km by 2km is covered. The system is including six intersections in each of the vertical and horizontal directions. In this study, the initial population is 100, mutation rate is 0.4 and maximum fuzzy crossover is 0.8. Figures 9-12 have been shown to optimize the tree objective function in different generations of the algorithm. In figure12, the performance of fuzzy-genetic operators has been shown. For each of the above modes, this algorithm run 10 iterations and its mean will be shown. The result of proposed system with fuzzy operators show that in 71 generation, the optimal response be achieved while without using of fuzzy operator after 95 generation can be achieved. This result indicates that the use of fuzzy operator increases the speed of achieve optimal response and causes improved the performance of the algorithm. In the remainder of this section (Figure 13), we compare the proposed system with blind search system and reserved system proposed in [4]. The system proposed in [4] not consider traffic information and this has led to inefficiency of the system.
Figure 8: flowchart of the Proposed method
Figure 9: Average and the best value of the walking time
Figure 10: Average and the best value of the travel time
Figure 11: Average and the best value of the cost objective function
Figure 12: Compare the performance of the proposed system
V. conclusion
The main objective of this paper is to provide a parking space reservation system based on optimal selection method. The proposed system seeks the most suitable space based on the information collected from the VANET and using an adaptive neuro-fuzzy system and genetic algorithms to solve this problem. In this paper, at first, we proposed that the information collected in VANET classified to form a series of hourly, daily and weekly times. After the stage, classified information will be delivered to ANFIS system. With this series of time, the proposed ANFIS system trained and will be able to quickly predict traffic flows with acceptable accuracy. The simulation results show that the performance and accuracy of the method have been significantly improved compared to previous works.
Figure 13: Searching time in proposed method
References
[1] D. Teodorović and P. Lučić, "Intelligent parking systems," European Journal of Operational Research, vol. 175, pp. 1666-1681, 2006.
[2] H. Wang and W. He, "A reservation-based smart parking system," in Computer Communications Workshops (INFOCOM WKSHPS), 2011 IEEE Conference on, 2011, pp. 690-695.
[3] H. Zhao, L. Lu, C. Song, and Y. Wu, "IPARK: Location-aware-based intelligent parking guidance over infrastructureless VANETs," International Journal of Distributed Sensor Networks, vol. 2012, 2012.
[4] X. Zhang, D. Li, and J. Chen, "Parking Space Reservation based on VANETs," International Journal of Advances in Management Science, 2014.
[5] Y. Ma, M. Chowdhury, A. Sadek, and M. Jeihani, "Integrated traffic and communication performance evaluation of an intelligent vehicle infrastructure integration (VII) system for online travel-time prediction," IEEE Transactions on Intelligent Transportation Systems, vol. 13, pp. 1369-1382, 2012.
[6] R. P. D. Nath, H.-J. Lee, N. K. Chowdhury, and J.-W. Chang, "Modified K-means clustering for travel time prediction based on historical traffic data," in International Conference on Knowledge-Based and Intelligent Information and Engineering Systems, 2010, pp. 511-521.
[7] R. Bauza and J. Gozálvez, "Traffic congestion detection in large-scale scenarios using vehicle-to-vehicle communications," Journal of Network and Computer Applications, vol. 36, pp. 1295-1307, 2013.
