Tell us about dSPACE business development – both in 2019 and currently
Martin Goetzeler: 2019 was a good year for dSPACE. We grew our worldwide sales by 11 percent. Therefore, we were able to increase our investments into research and development. Our new strength grew organically, but we also invested in technology and mergers and acquisitions (M&A). In the last two years alone, we grew our team by around 400 employees to a total of 1,800. More than 1,400 of our employees are engineers and computer scientists. The effects of the pandemic are not yet foreseeable for the current year. We see reluctance on the part of our customers. Costs and projects are under scrutiny. We have to continue monitoring this development. However, there are also many important product launches that we want to become a success. To provide our customers with the best possible support, we accompany them closely during implementation. We also make sure that we are able to deliver and work consistently on the continuous development of our portfolio. We have the means and the will to do so.
How does dSPACE manage to maintain operations in times of the pandemic?
Martin Goetzeler: A key measure was the immediate introduction of more flexible working conditions. For a while, more than 80% of our workforce was working from home. Communication is very important, which is why we are providing our employees with a regular update on developments in the company as we continuously reassess the overall situation. We are currently working on effectively taking advantage of many new opportunities offered by the new reality of working.
How are you adjusting to the current radical changes in technology?
Martin Goetzeler: Mobility trends such as connectivity, autonomous driving or new mobility services, naturally have a significant impact on development and validation. Software is the key component in the digitalized and intelligent vehicle. Vehicles are ‘mobile devices on wheels’. For our customers, this means not only applying the agile development methods that are established in the IT sector, but also adopting the data-driven development and validation of the AI-based functions of an autonomous vehicle. Additionally, comprehensive expertise in cloud computing is required to handle the resulting data volumes. Most companies have set up their own divisions or even new companies with their own culture to meet these challenges.
However, there are other mobility trends that also require new ways of thinking. For example, electric mobility is based on a new ecosystem. After all, we are not only dealing with a new vehicle technology, such as batteries or power electronics. We also must redefine the infrastructure and information systems, as well as service and maintenance. Even though the electrification of the powertrain is based on tried and tested methods, especially model-based development, the building blocks must be integrated, developed, and tested in the ecosystem using software.
The greatest challenge with all trends is to make complexity manageable and break new ground in the process. dSPACE is contributing to this.
How will dSPACE contribute?
Martin Goetzeler: We have reviewed our strategy in light of the challenges and complexity. The result: A new dSPACE. Today, we offer our customers a comprehensive validation solution that features the right architecture for model- and data-based development, software and hardware simulations, and our integrated, scalable tool chain. With this end-to-end solution, we enable customers to benefit from efficient processes and structures, short time-to-market, and cost optimization. We have been working on this for the past two years, and partially have had to reinvent ourselves.
Our strengths are prototyping, hardware-in-the-loop testing, and embedded software development, as well as bus systems – where we have been working on innovation projects with key customers for many years. Together, they form an important basis for our work. Our software-in-the-loop solutions are also essential components of our end-to-end solution. Yet, there was still room for improvement.
We have, therefore, drastically improved our competencies in the areas of software simulation, cloud computing, and artificial intelligence (AI). Developing AI competencies has two dimensions for us. On the one hand, we bring our tools to a level that they can optimally handle the requirements of our customers‘ neural networks. On the other hand, AI helps us become more efficient. Highly automated annotation, for example, is made possible only by artificial intelligence.
We have now bundled our software activities into a separate business unit. In this unit, we integrate data-driven simulation and validation solutions, as well as our expertise in software coding, test automation, and data management. This means that we speak the language of many of our new customer contacts and can respond flexibly and quickly to their requirements. We also ensure that we can provide our customers with ideal support by means of agile development methods and interdisciplinary teams. This makes us a competent partner for everything related to software development in the vehicle.
Something that is really important to me during the restructuring of the company: Our traditional business with test solutions and our software business are of central importance, both individually and in combination. We are strengthening both areas to tackle complexity.
Go to the next page to learn how dSPACE intends to provide the complete solution chain, read about the importance of electromobility for the company, the major technical challenges it faces, and how dSPACE is moving forward.
What exactly does this mean for the users?
Martin Goetzeler: Our end-to-end range of solutions in the various applications, which extend to electric mobility and autonomous driving, covers the spectrum of simulation and validation. Our customers appreciate this comprehensive approach because it accelerates their processes and reduces costs. In addition, customers can integrate their tools into our solutions at any time. We are also prepared to invest with the customer. Moreover, we have set up dSPACE Consulting, a separate unit, to advise customers on developing and implementing their test strategies, all the way to homologation. This is also done in cooperation with technical services such as TÜV.
How can dSPACE work on the complete solution chain?
Martin Goetzeler: Of course, we cannot cover the entire range of products alone. We, therefore, regularly review acquisitions to close technological gaps. And today, we increasingly work with partners. These cooperations can involve technological, infrastructure, or implementation tasks. For technological topics such as sensors or software tools, these partners include NXP, Intempora, BTC, and MES. We cooperate with them in the area of data-driven development or code generation. We pursue similar approaches in cloud computing or engineering. This enables us to support specific solutions and services depending on the customer.
In the current climate, are you noticing that manufacturers are reassessing their development budgets?
Martin Goetzeler: We can see that some manufacturers and suppliers are focusing their development projects on autonomous driving, while others are implementing their roadmaps at almost the same speed. Safety- and comfort-driven assistance functions, based on software and sensor technology, are both advancing. The pace of development in the field of electrification also remains high. In China, we are seeing clear signs that the order situation is improving again. Solutions for testing battery management systems (BMS) continue to play a major role. Manufacturers of trucks and buses, in particular, are increasingly asking for HIL test systems for fuel cell technology. Despite the corona pandemic, we still receive inquiries from the USA. They have a demand for test solutions for battery management systems, complex motor models, and high-voltage motor emulators.
How important is electromobility for dSPACE?
Martin Goetzeler: Electromobility is not a new topic for us, but again we are faced with complex challenges. Our focus currently is on power electronics and the charging infrastructure. The onboard chargers in the vehicles must be compatible with the worldwide variety of different charging infrastructures. For efficiency reasons, power electronics are operated at ever higher switching frequencies, so we also have to make our simulation models faster.
What support do you provide specifically?
Martin Goetzeler: With our new Smart Charging Solution, our customers can test onboard chargers for all relevant charging standards. On request, we combine this with real power electronics. This lets us reproduce the many different charging stations. In the field of motors or power electronics, we use our modeling approaches to perform highly-accurate and fast simulations on powerful FPGAs. We are also seeing great demand for our battery and motor emulators. For these, we use patented high-voltage systems for the precise reproduction of voltages and currents. This allows for the power electronics to be tested at full electrical power. At the same time, the customer benefits from continuity in testing – from the signal level to the power level and highly dynamic mechatronic test benches.
With our scalable end-to-end solution, we are also making an important contribution to electromobility by mastering the complexity of the interoperation of electric motors, battery management, power electronics, and charging infrastructures. We support our customers worldwide in this task and therefore contribute significantly to their development productivity.
Let‘s have a look at autonomous driving: How convinced are you that autonomous driving works?
Martin Goetzeler: Autonomous driving is revolutionizing the entire world of mobility – whether in individual transport, in service-oriented mobility concepts, or in the commercial vehicle sector. At dSPACE, we are certainly convinced that autonomous driving works. In my view, the important question is always for which use cases a business case can be presented and how long the timeline is. My perception is that many manufacturers are continuing to drive forward development in the robotic taxi and passenger car sectors, albeit partly delayed by the pandemic. Applications for the commercial vehicle sector are also economically attractive – for example, using autonomous vehicles on construction sites or in other confined areas. An essential success factor in the introduction of Level 3, 4, and potentially Level 5 systems is validation and homologation. After all, the question of whether we will succeed in making cars drive autonomously and safely from a purely technical point of view is as important as the question of how to set up the development chains and processes in such a way that the vehicles are eventually approved for the road.
Find out more about the major technical challenges and (radar) sensor technology in the last part of the interview – including information on how dSPACE is moving forward.
Where do you see the major technical challenges and how do you support mastering these challenges?
Martin Goetzeler: It is a vital challenge to make the complexity of the overall system manageable. This requires a robust architecture with hardware and software components. All development steps must be coordinated with each other and integrated into a continuous process so that both the individual components and overall systems can be validated. At dSPACE, we have done this by providing a solution component for each step. We offer tools for data collection, storage in data centers, data annotation, simulation, and resimulation systems. We close the gap from data annotation to simulation with a service for generating simulation scenarios from measurement data to transfer real driving situations into simulation. Once the scenarios are virtualized, we can use scenario-based testing to repeatedly simulate the real-life situation in many different variants, thus generating wide test coverage. The AUTERA system, which we recently launched successfully, provides support for the collection of data and resimulation.
What exactly does this mean for development?
Martin Goetzeler: Sensor simulation is a good real-life example: A vehicle has to be tested in a defined environment, for example, in intersection traffic, using many parameters: Road users, speed, weather, etc. This cannot be done in the real world because the multitude of situations to be tested cannot be reproducibly represented. In simulation, however, this is feasible. We offer integrated solutions that allow us to test the traffic and the vehicle with all of its sensors, both in purely software-based simulation and in hardware simulation – for all types of sensors from cameras, radar and lidar to ultrasound.
In our customer projects, we work very closely with sensor manufacturers and sensor specialists, not only to realistically represent the physics, but also to be able to integrate the algorithms or the real hardware of the manufacturers into the test setup. You need both – the most accurate representation of the physics and the integration of the signal processing – to build test systems that can also return reliable results.
For me, this is a good example of how the experience of dSPACE in the area of algorithmics goes hand in hand with our simulation and software know-how, as well as our hardware expertise.
How were you able to acquire knowledge in the area of AD in such a short time?
Martin Goetzeler: Mainly organically by building and rebuilding our team, but also through acquisitions. To strengthen the portfolio, we acquired the start-up understand.ai, or UAI for short. The acquisition enabled us to achieve several goals: First, we were able to complete our solution portfolio for autonomous driving by directly closing the gap between data recording and simulation. With the software from UAI, the captured data is annotated with a high level of automation and in a very cost-efficient manner. Labeling makes the data usable for training and validating algorithms. The automated generation of scenarios lets us reproduce traffic situations with changed parameters as often as we want in our simulation world. Second, the acquisition has strengthened our AI and cloud expertise and created opportunities for the agile development of our corporate culture.
Earlier we touched on the subject of sensor technology. Why has dSPACE strengthened its radar business?
Martin Goetzeler: Radar is a key technology for ADAS and AD functions. Therefore, at the beginning of 2019, we expanded our portfolio through a close partnership with the radar specialists ITS and mirosys. We are consistently developing the existing DARTS (dSPACE Automotive Radar Test Systems) range, which is characterized by precision, large range coverage, and high flexibility. Only recently, we equipped two products with multitarget capabilities for testing automotive radar sensors. Thanks to this, multiple independent radar echoes can now be simulated simultaneously by intelligent software with only one device. This was very well received by the market.
How do you ensure that you are aware of new requirements in sensor development at an early stage?
Martin Goetzeler: By means of direct cooperation with chip manufacturers. In a pilot project with a leading manufacturer, we are developing a target simulator with 5 GHz bandwidth for a new type of high-resolution radar generation. And we are working closely with Uhnder to advance the development of their 4-D radar chips. We want to extend our innovative lead in radar test solutions.After all, the solutions are used throughout the entire value chain, right down to the workshop area. At the same time, we are also working on building up lidar expertise to be able to extend our over-the-air test portfolio to lidar, in addition to radar, camera, and ultrasound. However, it is still a little too early to talk about the specific results.
What‘s next for dSPACE?
Martin Goetzeler: My most important task is to lead a strengthened dSPACE out of this crisis. To continue the close partnership with our customers, in the long term, we and our entire team are consistently pushing for the application of the latest IT technologies, the expansion of our cooperations, our simulation expertise, and our end-to-end solutions for the various domains – always with the aim of mastering complexity in an agile manner and increasing added value for the customer.
Alfred Vollmer
(aok)