Scripting languages such as Lua are relatively quick to learn, you don't have to use a dedicated compiler and the source code is platform-independent - so it can be used on Windows, Mac OS X and Linux. The deliberate simplicity can also be seen in the scope of the language: While C++ comprises around 2000 language elements and C around 850, Lua has just 250 commands with which it can be programmed efficiently. Admittedly: In terms of performance, Lua lags behind programming languages such as C and C++; however, this speed disadvantage can be compensated for with the help of suitable DLL and real-time libraries.

Figure 1: Lua commands can be instantiated on real-time machines via RPC using simple script commands.

© Profimatics/eclipse.org

So what is so special about scripting languages such as Lua for programming real-time applications in the field of industrial automation? To answer this question, it is worth taking a look at a possible combination of Lua and a real-time operating system, as is the case with the RT-Lua development package from Profimatics in Barmstedt. Here, direct access to the intimate real-time operating system environment of Tenasys is possible by means of prepared interfaces; this allows the direct use of real-time functionality from a scripting language (Figure 1). This includes inter-process communication, real-time access to the existing hardware (I/O, PCI bus etc.) and also the network functions (TCP/IP etc.).

The connection between the scripting language and the real-time platform can be established in several ways. A common approach is an RPC connection, normally based on the TCP/IP protocol. This allows the developer to access the real-time environment using standardized interfaces without having to use special or proprietary protocols. However, this RPC-based access to a real-time OS machine can also be realized in the form of a client-server implementation. This allows Lua script commands to be executed directly on the server, variables to be created and new functions to be loaded onto the server. In addition, Intime's RPC library is also available for other operating systems such as Linux and Mac OS X. This means that functions can also be executed from these operating systems in real time on a corresponding runtime system.

Lua for rapid prototyping and testing

Figure 2: The code snippet shows a function that sends a CAN message as soon as the memory area 'can_mem' has been displayed.

© Profimatics/eclipse.org

Against this background, the Lua scripting language can now also be used for rapid prototyping and testing of real-time systems, making it possible, for example, to commission special real-time hardware with just a few script commands. This does not even require the integration of special drivers, which are programmed in C or C++ as standard. This is already being implemented in the field of medical applications, for example, where the SJA1000 CAN controller is used. The code excerpt in Figure 2 is intended to show how easy it is to write to the registers of a CAN controller chip in order to send a CAN message.

This driverless addressing of hardware with the help of Lua is possible because privileged commands can be executed natively under Intime, which are otherwise only available under Windows in ring 0 - i.e. at kernel level. In order to make these commands more readily available, they can be combined in a library in Lua. These include, for example, calls for reading and writing IO ports, for direct access to PCI components and for displaying memory areas of a PCI card.

Ideal field of application - robot control

Lua can also be integrated into existing applications, as Bosch Rexroth has done with Indra Motion MLC - an integrated engineering and runtime system for logic, motion, robotics and hydraulics tasks.

RT-Lua consists of modular software components that require no installation and are each the size of a conventional JPEG.

© Profimatics/eclipse.org

This means that complex movements can be executed with simple script commands, which allows simplified programming of robots and other machines without PLC knowledge. Indra Motion MLC uses two features of Lua for this purpose: on the one hand, Lua is extended with the existing functionalities of the target system (extensible), and on the other hand, the scripting language is integrated into the application (embeddable) and made available to the user.
Lua is interesting not least for C and C++ developers, as existing functions or applications can be provided with a Lua API using suitable tools. Wrappers are used to check C/C++ functions for their real-time suitability. This saves the effort of having your own test framework, as debugging can then be carried out using Lua script commands. In this context, applications can be developed interactively; individual commands can be entered within the Lua console, which are then translated into byte code and executed in the real-time environment.

However, it is more convenient to debug Lua real-time applications within the EclipseLDT development environment. Here developers can set breakpoints as usual, view variables, change them and more. Communication between EclipseLDT and the debugger is implemented via a TCP/IP connection.