UTEM: Research Methodology (MPSW 5013): Suffian Bin Ayob

Optimization Of Vision System Using AI To Improve AGV Navigation In Industrial Working Environment Chapter 1:Introduction: From the first introduction of Automated Guided Vehicle (AGV) in early 1950s by Barrett Electronics of Northbrook, Illinois, up to this day, there seems to be a tremendous efforts and research to upgrade the capability and flexibility of this vehicle system. Throughout this development stage, many definitions have been done to describe and define the AGV terms. Groover (1987) defines Automated Guided Vehicle System (AGVS) as a materials handling system that uses independently operated, self propelled vehicles known as the automated guided vehicle or AGV that moves along defined pathways between delivery points or stations. Muller (1983) as cited in Vis (2004) identifies an automated guided vehicle (AGV) as a driverless transport system used for horizontal movement of materials. The first model of AGV was simply a tow truck that followed a wire in the floor and is lack of flexibility. This AGV model performs minimal capability to quickly and easily change paths or routes. The AGV is navigating based on the detection of a frequency that was induced through the wire, and so be directed through its route. During that specific period, the AGV acts as a kind of dumb-device. The continuous development of this vehicle system later on allowing the AGV to navigate base on the reflected light from paint strip mounted on the floor where the route is preprogrammed in the AGV system The introduction of laser technology in late 1980s further influencing most AGV manufacturers to adapt this recent technology into their vehicle system. Laser Guided Vehicle (LGV) contributions in an automated process, permits the vehicle to communicate (via an off board server) with other robots to ensure smooth movement of the products through the manufacturing areas, warehouse or sent directly to shipping areas. Rapid growth in microelectronics and microcomputers technologies with lower cost, enhance the capability and flexibility of the next generation of AGV systems. The AGVs transforms to be more intelligent, enable them to store instructions about the routes, perform decisions making task and take part in traffic control of the global system. Moreover, this new wireless guidance systems, using lasers or inertial systems permitted AGVs to maneuver without physical guide paths. This free-ranging AGVs technology allows the installation of such systems easier and simplifies the guide path modifications when new stations or flows were added. As a replacement of the navigating method using fixed path, the current AGV capable to navigate freely through various method such as using sensor that enabling the vehicle to calculate the distance it has traveled or using other landmark or equipment as a reference. With reference to the emerging development history of AGV mentioned above and the manufacturing paradigm in the 21st Century, the following sections will discuss the potential areas and gaps to optimize the performance of AGV as the important transporter within the manufacturing environment.

-1-

UTEM: Research Methodology (MPSW 5013): Suffian Bin Ayob

Chapter 2: Literature Review AVG applications: Since their introduction stage, the use of AGVs has grown enormously. The areas of application and variation in types have increased significantly in number. The applications of AGVs covered both inside and outside environments, such as manufacturing, distribution, transshipment and (external) transportation areas. Gotting (2000) mentioned that over 20,000 AGVs are used in industrial applications. The author emphasize that the usage of AGVs will pay off for environments with repeating transportation patterns. Among the AGV applications environments examples are distribution, transshipment and transportation systems. In the distribution areas such as warehouses and cross docking centres , AGVs acts as the internal transport of, for example, pallets between the various departments, such as receiving, storage, sorting and shipment areas. Meanwhile, at transshipment systems, such as container terminals, AGVs take care of the transport of products between the various modes of transport. Berman, Schechtman and Edan (2008) stated that an automated guided vehicles system (AGVS) enables flexible material routing and dispatching, and is especially suited for flexible manufacturing environments in which product mix and priorities may continuously vary. These authors also suggest that the concepts and methodologies from the multi-robot field have migrated to AGVS control. AGVS Classification &Navigation method: In terms of classifications, the AGV can be divided into three basic levels as per mentioned below: 1. Guide path determination a) Static Path or Fixed Path In static guide path, the vehicles use a set of predetermined paths between possible origins and destinations. Variety of guidance mechanisms can be used such as wires embedded in the floor, chemical or optical sensors, dead reckoning and mapping of the paths by using software. Barbera and Perez (2009) cited that due to the fixed paths, a modification of the layout implies stopping the whole system and changing the paths physically. On the other hand, the navigation task is easy, requiring only a sensor to detect the guide on the floor. This path can furthered be divided into unidirectional and bidirectional system. Anh and De Koster (2005) states that AGV can travel a lane in only one direction (unidirectional)or both directions (bidirectional). The unidirectional system considered as the easier to control system as the deadlocking and collision problems can be avoided. Meanwhile, the bidirectional system able to introduce improvement in productivity and minimize vehicle usage, but the control system for this path method is much more complex since multiple AGV shares the same guide path.

-2-

UTEM: Research Methodology (MPSW 5013): Suffian Bin Ayob

b) Dynamic path or open path navigation This system use fully autonomous vehicles, which are capable of determining its path through obstacle detection and avoidance systems. The vehicle then determines its path form its current position to the desired position through its internal navigation scheme (Peters et al). Barbera and Perez (2009) cited that in open path navigation the AGV can, at least theoretically, take any guide path to navigate between points. Thus, in order to navigate in this environment ,the AGV needs a map and a method to know its own location. 2. Vehicle capacity a) Single unit load Briefly, in a single load system, an empty vehicle will travel from its current position to a station to pick up the load and then travel to the desired position to drop off the load. b) Multiple loads In multiple load system, the task of the vehicle is more complicated where the vehicle may be interrupted while performing its task. It may stop to another station to pick up another load. 3. Vehicle addressing mechanism a) Direct address In direct address system, the routes are not determined in advanced and the vehicles are not restricted to serve any particular station. Any vehicle is allowed to visit any stations available in the same system. Comparing to the real-world, this system is much alike the taxi service. This requirement creates complicated planning action since the location of the vehicle is not known initially but only changes upon system changes.

b) Indirect address In indirect address system, vehicles will stop at stations in a fixed sequence. Comparing to the real-world, this system is much alike a bus service. The routes are predetermined as part of the system design, thus the dispatching in this system is straightforward. As the route of the vehicle is predetermined, it will pick up and drop off loads when it reach each stations in its route (Peters et al). A typical AGV will consist of the frame, batteries, electrical system, drive unit, steering, on board controller and work platform Problems and proposed solutions According to Freitag and Reiter (2007), the evolution of manufacturing paradigms from Flexible Manufacturing Systems over Reconfigurable Systems towards Autonomous Manufacturing Systems. These paradigm shifts are driven by changing market demands and the corresponding requirements to the manufacturing systems. To -3-

UTEM: Research Methodology (MPSW 5013): Suffian Bin Ayob

remain competitive, one of the major characteristics of future manufacturing companies is their changeability. This will lead to self-x characteristics such as selfadaptation, self-optimisation,and self-organisation of the system. Weber (2008) suggested that even todays AGVs have the ability to adapt somewhat to changing conditions, manufacturers are looking for machines that are even more flexible and more autonomous. The industry is moving toward vehicles equipped with much more sophisticated navigation skills, speed and precision. Based on the above statement, indirectly its shows that AGV as one of the material transporter within the Autonomous Manufacturing System need to adapt to the self-xcharacteristics requirement. In order to remains relevant, AVG need to develop to become the most efficient, flexible and able to react to dynamic and uncertain environments. Barbera and Perez(2009) stated that at this moment, AGV systems are widely chosen by manufacturers to implement truly flexible material handling systems (MHS),which are often necessary for the highly automated manufacture model used today. These authors found that although most of the key issues to obtain autonomous navigation systems have been broadly addressed in different scenarios by the autonomous systems community, developing systems that are able to plan and react to dynamic and uncertain environments, there are not many successful examples of fully working service robots in industrial applications, like AVGs. In addition, these authors suggested that in order to remove the dependence on guidance infrastructure, which is not suitable or not possible in some applications, some AGVs prototypes have incorporated vision systems for operating in infrastructure-free environments. With reference to Barbera and Perez(2009) above suggestion, it shows the needs to incorporate AGV with vision system in order to obtain an optimal guide-path and flexibility of the system to adapt to the manufacturing future demands and environment variations. Furthermore, Vis (2005) proposed that in controlling and designing AGV systems the problem of prevention of AGV collisions and deadlocks should be addressed. These authors recommended attaching sensors on AGVs, to avoid physical collisions. An AGV should have the ability to avoid obstacles and the ability to return to its original path without any collisions. Golnabi (2002) cited that vision systems can be considered intelligent sensor systems for complex measuring or inspection tasks .Several tasks such as image acquisition, processing, segmentation, and pattern recognition are conceivable. The role of imageacquisition sub-system in a vision system is to transform optical image data into an array of numerical data, which may be manipulated by a computer. Jia, Balasuriya and Challa(2006) suggested that further experiments can be done on autonomous vehicles to conduct the real time target tracking tasks. In this research, the solution approach is to equip the AGV with smart navigating capability is by incorporating onboard cameras to control the vehicles movement path. The AGV moves autonomously using 2 DC motors controlled by a remote PC using fuzzy logic system. A camera is connected to the computer to serve as the eyes of the system. All decisions or commands coming from the computer for appropriate AGV actions are dependent to the images captured by the camera. From this image, information regarding the target point, the travelling routes, and the AGV location will be extracted and analyzed to come up with navigation strategies. These strategies dictate the AGV real time action in performing its task like positioning, navigating and avoiding obstacles. -4-

UTEM: Research Methodology (MPSW 5013): Suffian Bin Ayob

REFERENCES Anh,T.L., and De Koster,M.B.M.,2005,A review of design and control of automated guided vehicle systems,European Journal Of Operation Research,171(2006),pp.1-23. Barbera, H.M., and Perez ,D.H.,2009,Autonomous navigation of an automated guided vehicle in industrial environment,Robotics and Computer Integrated Manufacturing,26(2010),pp.296-311. Berman,S.,Schechtman,E., and Edan,Y.,2008,Evaluation of automatic guided vehicle systems, Robotics and Computer Integrated Manufacturing,25(2009),pp.522-528. Freitag,M.,Scholz-Reiter,B.,Autonomous Process in Assembly Systems,Annals of the CIRP,Vol.56/2/2007,pp.712-729. Golnabi,H.,2002,Role of laser sensor systems in automation and flexible manufacturing, Robotics and Computer Integrated Manufacturing,19(2003),pp.201210. Vis,I.F.A.,2004,Survey of research in the design and control of automated guided vehicle systems, European Journal Of Operation Research,170(2006),pp.677-709. Weber.A.,2008, Next Generation AGV Technology.[online] Available at: http://www.assemblymag.com/Articles/Web_Exclusive/BNP_GUID_9-52006_A_10000000000000449251 [Accessed on 29 November 2011].

-5-