1. General questions about the software

1.1. Which navigation principles are supported?

openTCS works independently from specific navigation implementations, so any kind of navigation principle may be used. Locating and navigating are tasks usually performed on the vehicle side: The vehicle simply reports its current state, including its position, to the control system, and the control system orders the vehicle to move to a different position.

Note that openTCS focuses on the dispatching of vehicles for transport orders. This task should not be confused with lower-level vehicle-side tasks like accelerating, decelerating and steering the vehicle, which are out of openTCS’s scope.

1.2. What are the requirements to manage a vehicle with openTCS?

The requirements for this are minimal:

  1. Communication with the vehicle must be possible. (This implies that the vehicle’s communication interface must be specified and accessible.) How this communication is implemented exactly is not really relevant as long as the communication hardware exists and can be used from Java software. In many cases, standard Wi-Fi hardware is used.

  2. The vehicle must be able to report its current position/state.

  3. The vehicle must be able to perform movement orders from its current position to an adjacent position in the driving course/environment given by openTCS.

1.3. How many vehicles can be managed at the same time?

There is virtually no limit imposed by the design of the software. The hardware equipment used (CPU, RAM, communication bandwidth etc.) can, however, limit the system’s effective performance.

1.4. Which interfaces for transport orders exist?

Transport orders can be created interactively by a user via the graphical user interface.

To create transport orders from other systems, e.g. a warehouse management system, interfaces based on Java Remote Method Invocation (RMI) and webservices are included. These are documented in the developer’s guide. Additional, custom interfaces for software that cannot use these interfaces can be added easily.

2. Plant models

2.1. How can I migrate plant models created with openTCS before version 3.0?

  1. Backup the contents of the kernel directory’s data/ subdirectory. It contains subdirectories with a plant model in a model.xml file each.

  2. Clear the kernel’s data/ directory by removing all its subdirectories.

  3. Copy one of the model.xml files from the backup directories back to the data/ directory, so that it contains only this single plant model.

    Example: Assuming your data/ directory contains three model subdirectories modelA, modelB and modelC. After step one, the three model directories should have been copied to another location. After step two, the data/ directory should be empty and after step three, it should contain only the model.xml from directory modelA.

  4. Start the kernel and have it load your model. Then start the Model Editor application and select File  Load current kernel model from its menu to read the model data from the kernel.

  5. In the Model Editor application, select File  Save Model or File  Save Model As from the menu. The Model Editor application will persist the model data in a file with the given name and the extension .xml.

    Example: Following the previous example a file with the name modelA.xml should exist now.

  6. Delete the model.xml file you just moved to the data/ directory of the kernel. The migration of this plant model is finished.

  7. Shut down the kernel, go back to step three and repeat the procedure for the remaining models that you want to migrate.

    Example: Follow steps 3 - 7 with modelB and modelC instead of modelA.

2.2. Why are all transport orders marked UNROUTABLE when I only have reporting points in my model?

Vehicles are not allowed to stop at reporting points. Hence at least the starting point and the endpoint (usually linked to a location) of a route must be halt points to make routing possible.

2.3. How can I create curved paths between two points?

Select Bezier from the path tools and connect two points by clicking on the first point, dragging the mouse to the second point and releasing the mouse button there. Then activate the selection tool and click on the previously created bezier path. Two blue control points will appear. Drag the control points to change the shape of the path.

2.4. Why do points have two sets of coordinates (model and layout)?

The openTCS kernel itself works with a logical driving course model - geometric attributes of e.g. points and paths are not relevant for its core functionality. It may, however, have to provide (real/physical) coordinates of a destination point to a vehicle, depending on the way the vehicle’s navigation works. With openTCS, these are called the model coordinates.

The layout coordinates, on the other hand, are coordinates that are used merely for visualizing the driving course in the Model Editor and Operations Desk applications. These coordinates will probably be the same as the model coordinates in most cases, but they may differ, e.g. in cases where the driving course is supposed to be modelled/displayed in a distorted way.

3. Transport orders

3.1. How can I set priorities for transport orders?

With openTCS, transport orders do not have a priority attribute. That is because a transport order’s priority may change over time or when other transport orders are added.

In the end, a single transport order’s effective priority depends on the dispatcher implementation used. With openTCS’s default dispatcher, transport orders' deadline attributes are intended to be used for prioritizing - the sooner the deadline, the higher an order’s effective priority. To give a transport order a higher priority from the beginning, you can set its deadline to something earlier than all other orders' deadlines, e.g. to "right now" or a point of time in the past.

4. Vehicle drivers

4.1. Which side (driver or vehicle) should be the client/server?

This should be decided based on project-specific requirements. You are free to implement it either way.

5. Networking

5.1. How do I enable access to the RMI interface for clients when the kernel is running on a host with multiple IP addresses?

5.2. Why does the communication between the openTCS kernel application and client applications take an extraordinary amount of time when SSL is enabled?

This is probably because the kernel is running on a host with multiple IP addresses. See https://docs.oracle.com/javase/8/docs/technotes/guides/rmi/faq.html#netmultihomed.