Company type | Limited liability company (Gesellschaft mit beschränkter Haftung) |
---|---|
Industry | Automotive, aerospace,[1] software, hardware, engineering |
Founded | 1988 |
Headquarters | |
Key people | Carsten Hoff |
Revenue | € 400 million (2023)[2] |
Number of employees | 2600 (2023) |
Website | www.dspace.com |
dSPACE GmbH (digital signal processing and control engineering), located in Paderborn, Germany (North Rhine-Westphalia), is one of the world's leading providers of tools for developing electronic control units.[3][4]
dSPACE GmbH has Project Centers in Pfaffenhofen (near Munich), Böblingen (near Stuttgart), and Wolfsburg, and cooperates with the autonomous local dSPACE companies situated in the US, UK, France, Japan, China, Korea, Croatia and India. Various distributors represent dSPACE in other overseas markets.
dSPACE provides tools for developing, testing and calibrating electronic control units (ECUs) in the automotive, aerospace[5] and medical engineering industries, as well as in industrial automation[6] and mechatronics.[7] In most cases, the process of developing and testing ECUs is based on the five phases of the V-cycle. dSPACE's hardware and software cover four of these five phases, but not the first phase, control design.
The control design phase involves developing the control algorithms that will run on an ECU, usually by modeling them graphically. This process can be performed with Simulink, modeling software from MathWorks, and is outside dSPACE's application fields.
In rapid control prototyping, control algorithms are taken from a mathematical model and implemented as a real-time application so that the control strategies can be tested with the actual controlled system, such as a car or a robot. Simulink is used as the input and simulation tool, and Simulink Coder, also from MathWorks, is used as the code generator. dSPACE provides the necessary hardware platform consisting of a processor and interfaces for sensors and actuators, plus the Simulink blocks needed to integrate the interfaces into the Simulink model (Real-Time Interface, RTI).
In a development process based on mathematical models, the models are designed with graphical software, and then automatic production code generators are used to translate the models directly into code for ECUs/controllers. When a model's behavior has been validated, the code generator has to reliably transfer it to the target processor, whose resources are usually designed for the greatest possible cost-efficiency. In other words, the final production ECU generally has less memory and processing power than the RCP system on which the algorithm was developed and tested. As a result, the C code (production code) generated for the target processor has to meet stringent requirements regarding execution time and efficiency. In 1999, dSPACE introduced the production code generator TargetLink,[8] which is integrated into Simulink, the environment for model-based development. In addition to performing the actual autocoding, including code generation for AUTOSAR software components, TargetLink also makes it possible for developers to compare the behavior of the generated code with that of the original Simulink model (by means of software-in-the-loop (SIL) and processor-in-the-loop (PIL) simulation).
In HIL simulation,[9][10] a simulator mimics the environment in which an ECU will function: a car, an airplane, a robot, etc. First the ECU's inputs and outputs are connected to the simulator's inputs and outputs. In the next step, the simulator executes a real-time model of the ECU's working environment, which can consist of Automotive Simulation Models (ASMs) from dSPACE or of models from other vendors. This method provides a way to test new functions reproducibly in a safe environment, before a prototype of the product has even been produced. As with rapid control prototyping, Simulink models are the foundation. The advantage of HIL simulation in comparison with ECU tests in real prototype vehicles is that the tests on the control unit can be performed already during the development process. Errors are detected and eliminated very early and cost-efficiently.
Optimizing the control functions so that they fit specific applications is an integral part of ECU and controller development. To achieve this, the parameters of the ECUs are adjusted during ECU calibration. dSPACE offers software and hardware for this task.