A new approach to automated highway systems: Puppet master

In this research, a novel system is designed inscope of automated highways. The aim of this system is toprovide safe, secure and fast transportation on highways.This system uses both road sensors and GNSS (GlobalNavigation Satellite System).
Nội dung trích xuất từ tài liệu:
A new approach to automated highway systems: Puppet masterJournal of Automation and Control Engineering, Vol. 1, No. 2, June 2013A New Approach to Automated HighwaySystems: Puppet MasterOnur Cagirici and Mehmet S. UnluturkDepartment of Software Engineering, Izmir University of Economics, Izmir, TurkeyEmail: {onur.cagirici, suleyman.unluturk}@ieu.edu.trAbstract—In this research, a novel system is designed inscope of automated highways. The aim of this system is toprovide safe, secure and fast transportation on highways.This system uses both road sensors and GNSS (GlobalNavigation Satellite System). Since no human interaction isallowed in this system, the driver mistakes are reduced tozero. Based on the road sensors and the GNSS, the system isable to adjust the speeds and the following distancesbetween vehicles. The messaging system is also designed forthis puppet master architecture, which allows vehicles tocommunicate with the central components of the system.IndexTerms—Automatedtransportation systemsI.highway,never been eliminated [11]. The effect of human driversis much more different than the effect of otherautonomous cars. While controlled by a centralizedsystem, the autonomy of the autonomous car is somehowrestricted. The controller separates the shared data interms of location and sends the agents in a certain region.By sharing the location, the vehicles are able to makeadjustments to their velocity, acceleration or directionaccordingly. They are also able to communicate witheach other to negotiate their priority. This human-likebehavior will lead the vehicles to predict others’behaviors and avoid a possible accident. As a result ofintelligent agent-like approach, we can observe thatimposing emotions on autonomous cars reduces the riskof accidents [6]. In this centralized system, each car isdriven by the centralized system and they are unable tointerfere with the human. Since every single car is givenorders by only one decision mechanism, the cars will notcollide.The organization of the paper is as follows. In Section2, we present the related work and comparison of therelated work with our system. Section 3 describes thestructure of the system. Section 4 introduces themessaging interface. Section 5 includes an example forthe messaging phase. Section 6 clarifies the localizationprocess of the vehicles. Section 7 concludes the paper.intelligentINTRODUCTIONAfter it is estimated that 90% of the traffic accidentsare caused by human factors [1], the researchers aremostly focused on a system that does not need humaninterference. Studies have been made so far dealt withautonomy of a single car. Considering only a single carbrings up safety problems. In 1993, Varaiya made aresearch about automated transportation system andstressed on control problems [2]. In 1997, Thorpe et al.made a demonstration of a system that uses autonomouscars and proved that the system is technically feasible [3].After one year, in 1998, McMillin and Sanford made awide research about intelligent transportation systems andautomated highways [1]. Those were the first approachesfor the autonomy of a system, but it was not designed fora single car. However, there were a number of studiesthat were focused on autonomous vehicles [4]-[6]. Thecars have been developed so far, used sensor systems toobserve the environment [5], [7], [8]. Whereas,previously studied, using sensors to determine theenvironment is a successful, but not sufficient enough foran autonomous car to move safely [9]. Another algorithmthat has been developed to have the autonomous carfollowed the other car which is also, at one point,dependent on human driving [10]. Problem with thisalgorithm is, if the followed car fails, depending oncurrent condition, it might be unable for the other car toreact on time. Even though an efficient method has beendeveloped for a single car to observe the environment andreact dynamically, the effects of human drivers haveII.As described in the previous section, there are numberof studies on autonomous vehicles and autonomousvehicle systems. In this section, we are going to list thestudies and compare those studies with our work.Thorpe’s system used in his demonstration used centralsystem similar to ours [1]. However, the system usesvideo cameras and has a speed limit of 88kmh. The carsmove as clusters through the highway, taking the leadercar as reference to follow. Drawbacks of the system alsoinvestigated in the paper. The main drawback is, thesystem is dependent on one car. In our system, all cars areautonomous, yet controlled by a central systemcontinuously. When one car fails, central controller canhandle it because it controls the other cars. When centralcontroller fails, the cars will halt safely since they are alsoautonomous. The car-like robot developed in year 2003 isalso an autonomous vehicle [12]. The vehicle useslandmark to navigate and has a speed limit 30kmh. Mainobject for this robot is to move among pedestrians,Manuscript received September 11, 2012; revised December 21,2012.©2013 Engineering and Technology Publishingdoi: 10.12720/joace.1.2.82-85RELATED WORK82Journal of Automation and Control Engineering, Vol. 1, No. 2, June 2013therefore it is more reflexive but less robust. Thetechnique used in this system is not applicable to realmotor vehicles for this reason. Autopia is another systemthat has been developed in year 2009 [13]. The systemuses GPS, which has drawbacks against GNSS. Also, thesystem is only tested on empty roads, whereas our systemis applicable to highway traffic. CyberC3 is the systemwhich is the most similar to our system with respect tothe usage of central control [14]. However, the system isfully centralized, which is a drawback compared to oursystem. Once the central controller fails, autonomousvehicles will not be ...