In this Task, a knowledge-driven monitoring and orchestration system (KDMOS) for an assembly line is built from scratch. The system is knowledge-driven, i.e. it uses either Knowledge base (OWL) to make decisions on operations. The operations may include, for example, decisions where the pallet should go next, what operations the robot should execute essentially mapping product needs and the production system capacity. The Knowledge base (KB) or Expert base (KB) also, for example, helps refining how to reach next destination in terms of the services that have to be invoked.
- Language: JavaScript / nodeJs
- IDE: Webstorm
- Software Design: Visual Paradigm
The FASTory assembly line is used to assemble electronic devices. FASTory research environment is located in Tampere University of Technology, FAST-Lab., Finland. The FASTory line contains 10 identical workstations, which draw the main parts of three models of a mobile phone (WS 2-6 and 8-12). In addition, there is workstation, which is used to load raw materials (empty papers) and unload products (paper with mobile phone drawing) to/from pallets (WS 1). Finally, the twelfth workstation (WS7) is use for loading/ unloading pallets to the line
Before being called the FASTory line, the line was used for assembling real mobile phones. The pallets were equipped with a special jig for holding the mobile phones during the production process. See Figure 2. The line then was transformed for education and R&D (research and development) purposes. The modification covered the pallets, end-effectors and sensors.
The FASTory line requires an air pressure to work properly. The air pressure is used to actuate the track selector in the conveyor and the end effector in the robots. Thus, each work cell has a manual valve, pressure relief valve and solenoid valve
Each workstation in the FASTory line consist of two conveyors; main and bypass. The main conveyor is used if the pallet requires service from the work cell. Meanwhile, the bypass is used if the work cell is in busy state (another pallet(s) [max 2 pallets] are in the cell) to bypass the pallet to the next work cell.
There are five different types of zones. In each zone, there is one presence sensor that is used for detecting the pallet presence. In addition, there is one stopper to stop the pallet when the conveyor is transferring other pallets. An RFID reader is located in each zone 1 for each workstation. This reader reads the pallet’s tags for identifying the entering pallet
The factory line uses SONY SCARA robots for production. Each robot is represented as an RTU in the line. The robot has 4 DOF(X, Y, Z and Rz). The robot uses custom-made end-effector for grapping the pen. In this manner, a custom-made jig holds three different pens allowing the robot to pick the needed one
As highlighted in Conveyor Section, each zone in the conveyor has a presence magnetic proximity sensor to detect the presence of pallets.
In order to identify the pallets in the line, The FASTory line is equipped with RFID reader (in each Z1) that can read the attached tags beneath the pallets. The reading mechanism uses the RS-232 serial communication by the conveyor RTU.
In the FASTory line, there are two pneumatic actuators; stoppers and path selector. The stopper stops the pallet in each zone. Meanwhile, the path selector switches between main and bypass conveyors.
The FASTory is equipped with INICO S1000 Remote Terminal Units (RTUs). INICO S1000 is a programmable RTU device, which offers process control capabilities, as well as a Web-based Human-Machine Interface (HMI), and it supports REST and DPWS Web Services. Each Robot and Conveyor is controlled by an RTU. The RTUs are connected to the FASTory local network
To develop a knowledge driven, monitoring and supervisory control system of the FASTory Line using knowledge base (OWL) to make operational decisions for carrying out the required tasks. The implementation of the system must be economically feasible and take the form of iterative and incremental development process following the Unified Process methodology (UP) in the following sequential steps; Inception, Elaboration, Construction, and Transition.
During the inception phase, a case for the system is established and the boundaries of the project scope is determined.
- Analyze Project Requirement: The requirements of the project is first studied.
- Establish System Vision: The purpose: justify strategic importance of new system
- Prepare the Work Breakdown Structure: WBS is subsequently prepared using a Top-Down approach using the scheduling tool Project Libre and the dependencies between the tasks are studied. Time required to carry out each task is studied and entered alongside each task. Key metrics such as Critical path, Slack time and Milestones are identified and a software generated Gantt chart is prepared
- Initial Risk Assessment: Risks involved with the project are identified and its technological feasibility (technological proficiency of group members), schedule feasibility (course workload for the semester).
- Approximate Resource Utilization: The tasks are allocated to various group members based on the results findings of the assessment in the previous step.
- Estimate on Budget: economic feasibility (Cost/benefit analysis) are studied.
At the end of the inception phase is the first major project milestone: the Project Objectives Milestone. The project may be cancelled or considerably re-thought if it fails to pass this milestone.
The primary goals of Elaboration are to address known risk factors and to establish and validate the system architecture.
- Analyze Use Case models: Prepare Use Case Models for the system
- Finalizing Architecture: The architecture of the system to be developed is finalized.12
- Finalizing software requirements: The integrated development environment for the system is agreed upon. In this case it was WebStorm, Jetbrains. Also, a basic pseudo code is designed
- Finalizing the hardware requirements: The hardware requirement for the project is next finalized. Most of it are them are available in the Fastory Line.
- Analyzing Probability of Risk: Based on the selected architecture, are finalized hardware and software a final risk assessment is carried out before moving to the next phase, i.e Construction.
- Finalize Budget: The budget for the project is finalized at the end of this stage. At the end of the elaboration phase is the second important project milestone, the Project Architecture. The project may be aborted or considerably re-thought if it fails to pass this milestone.
During the construction phase, all remaining components and application features are developed and integrated into the product, and all features are thoroughly tested.
- Software development: The software components of system is being constructed by using suitable web development tools and platforms. This part also involves development of different communication mediums which allows us to communicate between factory line with front end user interface.
- Build user Interface: This project part consists of building front-end UI and data retrieval implementation. The basic front-end UI will be built first and since it is an iterative activity. The modifications will happen later after inclusion of various operational features and experiencing different user experiences.
- Introduce knowledge base: The processing of data and building OWL Knowledge base will be completed in this phase. In order to develop a pre-knowledge based system an informative set of data collection is also included in this phase.
- Prototype Testing: It involves the testing and simulations of different software blocks on the Fastory Simulator.
- Modification and debugging of Prototype: Depending on the simulations and prototype testing results, some necessary modification to pre-developed prototype can be done in this span of time.
- Project Deployment: The software hardware integration and testing phase is planned for this phase. It includes several iterations since it is different to replicate the simulation results on hardware. Finding of the hardware constrains and implementation restrictions for full scale deployment is the major part of this phase. Other tasks include researching suitable methodology for project integration and successful operational deployment.13
- Project Output Assessment: A detailed comparison between planned project proposal and achieved targets is carried out. This can be taken as an experience for future projects to avoid setting unrealistic or unclear goals.
At the end of the construction phase is the third major project milestone -Initial Project Operational Capability Milestone. At this point we decide if the project is ready to go operational, without exposing the it to high risks. This release is often called a “beta” release.
The purpose of the transition phase is to transition the software product to the user community.
- Detailed Analysis of beta version: The pre-release version is analyzed.
- Present Demo for user: A demo is presented to the user/customer.
- Analyze user feedback on Demo: Customer/User feedback is analyzed
- Make further correction / improvements: Any necessary corrections stemming from user feedback is incorporated.
- Submit the final version: Final version is released/submitted.
At the end of the transition phase is the fourth important project milestone, the Product Release Milestone
- Uses Form Validation to not allow users to Place Order without entering vital information such as Name, Address and Phone number
- Uses CSS to remind user of the missing information.
- Enables User to enter multiple orders using the Add Order button
- Allows User to delete and order by ticking a check box and pressing the delete row button in case the user inputs an erroneous order.
- Shows a preview of the Order entered so far.
- Fill User details in the Name, Address and Phone number fields
- Select the Order Variant by the drop down menus
- Add multiple Orders using the Add row Button
- Delete any order if necessary by checking the box adjacent to the order and pressing the “delete order button”
- Press the send button to place the order
The manufactring line requires increased agility and flexibility in its operation which is acheived by incorporation of knowledge expressed by ontologies concerning the location of pallets, the services available at the workstations, pallet routing algorithms etc. The ontology for representing production logic is represented in a computer interpretable way. This knowledge model can be used by automated decision-making to configure the control component that monitors and harmonizes the whole manufacturing system.
Web Ontology Language (OWL) is Resource Description Framework (RDF) based XMl language, it provides rich definition to describe entites and their internal relationships and introduces concepts of classes, properties and individuals for the purpose od description of structured information. SPARQL Protocol and RDF Query Language (SPARQL)is a useful language for retrieveing data from ontology. OWL model is queried and modified with SPARQL ASK and SELECT queries and modified by SPARQL UPDATE. SPARQL and OWL Language provide semantics to make the knowledge representation of systems up to date in real time with changes and even to control the execution of workflow based on the well-defined knowledge base
The following figures shows some of the generated graph from the knowledge base of the class members. It also showing some connections with other members of the classes and its capabilities.
The structure of the Database is such that there are three Main classes (subclasses of Thing) namely Equipment, Order and Process. Equipment have subclasses Conveyor and Robot. Order has subclass Product(Order is seen as a collection of Products) and Process has subclasses Execute and Transport. Conveyor further has subclasses which are its zones.
Zone_1 Instances are as many as the number of zone 1’s in the system and so on. “zone_1_10” is read as Zone 1 of workstation 10. Each zone has a transitive property “hasNeighbour” that has both domain and ranges as Zones. This is the fundamental property of the system that enables it to have a dynamic character with regards to the transport of pallets in the system. In the case shown in the Figure zone _1_1 has two neighbours zone_2_2 and zone_4_2. The Dynamic Behaviour is explained in the end
Robot have a property “hasExecuteProcess” that have the domain as “Robot” and range as “Process>Execute”. The intuition is that Robot have the ability that enable them to “execute” certain “Process”. What these processes are will be covered shortly.
The properties of the Products are shown in the above figure and are self-explanatory. Please note that this is populated dynamically during run time by the Order Class and the above picture is shown just for illustrative purposes.
The execute class contains service execution Urls for each Robot. The Robot first checks its execution capabilities by its “hasExecutionProcess” property which inturn returs the service execution URLs to the Robot. This is how the robot executes its “Process” operations on the pallets. The basic URL is retrieved from the hasUrl property and is merged with the requirement of the Pallet also retrieved from the Knowledge Base from the Product class during runtime.
The Transport class has instances that can take any instances names. In this why we introduce the dynamic capability of transport of pallets/products within the system. Here the names are just labeled for sanity purposes and for easy debugging during the project development phase. Each transport instance, has atleast two properties, hasLinkDest which has an arbitrary zone number and the Urls that can be used to reach it. To understand better, we can consider the case as in the above figure. “zone_1_1” read as zone 1 of workstation one has to Links that can be used to reach it as shown. One link is from the zone 3 of the previous workstation and the other being from zone 4 of the previous workstation, which in this case is Workstation 12. The Dynamic Transport Mechanism, when each Pallet reaches zone 1 of any given workstation, it first checks if the workstation is busy or not. In the case that it is not, it pallet is assigned a temporary goal to reach the zone 3 of the workstation. In the case that it is busy, the pallet is assigned a temporary goal to reach the zone 1 of the next workstation. Now the Pallet checks for the Neighbours of the current zone to see if it’s the temporary goal assigned, using the “hasNeighbour” property of the zones. This would typically return zones 2 and 4 of the workstation. Next as the goal is not found, the workstation further checks their neighbours (neighbour’s neighbour). At this level, the temp goal for the pallet is found and the workstation queries the Knowledge Base for the link that can be used to reach the neighbor whose neigbour is the temporary goal using the “hasReachLink” property of the . This is not just the case with zone 1. In every zone, the workstation first queries the KB for its neighbouring zone and then queries the Knowledge base for links to reach those zones. The way this is done is that the workstation queries “hasReachLink” Property of a zone that has the “hasLinkDest” of the destination that it wants to reach. In this way irrespective of the name of the Transport Instance the link for it is always returned correctly. Also if there were a robot that manually picks the pallet and puts it in any random place in the system it would carry on from there by checking its immediate neighbours and so on and so forth.