In order for the transition to the software-defined car to be successful, it is necessary to have a platform that makes middleware available.
In the automotive industry, the introduction of new features has traditionally followed an approach based on the integration of sensors and related ECUs required for their implementation. This has caused the proliferation of several independent systems, resulting in increased complexity and costs.
The concept of the software–defined car contrasts with this vision and aims at the introduction of software distributed on a limited number of processors, which share sensors and the communication network to enable advanced features like autonomous driving, powertrain, body control and infotainment. In order for the transition to the software–defined car to be successful, it is necessary to have a platform that makes middleware—a software layer that allows communication between the different distributed components—available. Middleware must exhibit characteristics of reliability, real–time behavior, and reduced latency.
TTTECH AUTO, ZETTASCALE COLLABORATION
An important step toward widespread adoption of the software–defined car is the recent collaboration pact signed by TTTech Auto and ZettaScale Technology, which aims to create the first European implementation of a data distribution service (DDS) that is safety–certified under ISO 26262 (ASIL D) for use in series cars. The new product, named MotionWise Cyclone DDS, will ensure safe and quality–guaranteed communication in real time within the entire vehicle, supporting software updates after sales to bolster the vehicle life cycle. DDS is a middleware protocol standardized by the Object Management Group (OMG), providing low–latency communication and a set of built–in quality–of–service (QoS) policies.
Through this collaboration, ZettaScale’s open–source Cyclone DDS network protocol will be combined with TTTech Auto’s MotionWise platform, as well as with time–sensitive networking technologies.
MotionWise is a safety software platform designed for automated driving. It can handle different applications, each of which runs in its own environment. That allows you to create a safe environment where applications with different safety and real–time requirements can coexist and interact.
“We are providing a software platform that allows the orchestration of applications in an automotive architecture beyond ECUs,” said Friedhelm Pickhard, chief growth officer of TTTech Auto. “We have to ensure that each application satisfies the timing, performance, and communication requirements and that this happens in a deterministic manner.”
Real–time orchestration, deterministic behavior, and guaranteed latency (regardless of the system load) are all key factors that make the MotionWise software platform suitable for challenging applications, such as automated driving. That implies a very strict regime because when the application detects an object, you have to ensure you break in a certain time for physical reasons, and the software cannot handle other time–consuming tasks that would otherwise result in the application reacting with unacceptable latency.
As Pickhard points out, “There are two aspects to consider. First of all, we have to ensure that the whole application chain runs and reacts in a certain time, without being blocked by another application. Second, we have to ensure that if a new application is downloaded, this behavior will be maintained.”
This property is very important because it simplifies software testing each time a new application has to be deployed. To apply a service–oriented architecture (SOA) like MotionWise to a vehicle, a proper communication stack is required. DDS is a technology that satisfies these requirements, but it has to be certified for safety.
“This is something we are doing together with ZettaScale,” Pickhard said. “Bringing together DDS and our core technology in a car is really a game–changer for the industry, because then we can enjoy all the benefits without having any degradation in terms of safety.”
ZettaScale was established as a spinoff of Adlink Technologies to address the demand for its two core technologies—Cyclone DDS and Zenoh—coming from the automotive and robotics markets. Zenoh is an innovative protocol designed to provide unified and location–transparent abstractions for data in motion, data at rest (databases, file systems, etc.), and computations. Capable of throughput of over 50 Gbps, Zenoh has a latency of few tens of microseconds, has a minimal wire overhead of 5 bytes, and can run on different types of hardware architectures, from a multi–core processor to a tiny microcontroller.
“Real–time behavior is not necessarily just about timescales; it’s also about real determinism and making sure that you have a schedule that is always met, because if you miss a deadline in a car, then you know something bad might happen,” said Angelo Corsaro, CEO of ZettaScale Technology.
According to Corsaro, the automotive industry is adopting DDS because other markets, such as avionics and military vehicles, have successfully done it before, achieving a high level of modularity, reconfigurability, and seamless integration of the world in real time.
“In DDS, one of the things that our users enjoy a lot is dynamic discovery,” Corsaro said. “This is an essential feature for systems where you can have partial failure, or you want to add things on the fly and expand the capabilities. This is something that you cannot certify and that is not needed on the most critical part of the system, yet it is a game–changer during development and for the less critical portion of the system.”
Users can take advantage of this dynamicity aspect in the non–safety–critical portion of the system. If you need to test the software of a distributed system and you don’t have dynamic discovery, you will need to configure all the communication endpoints, which is a tedious and error–prone activity. However, if you had dynamic discovery, you could run the same system with one node, 10 nodes, 50 nodes, or more. There is no need to change anything in the configuration, as nodes are discovered automatically.
“One of the simplifications that we bring, with respect to other technologies, is that DDS supports by design the quality of service,” Corsaro said . “By describing the quality of service through a set of QoS policies, DDS captures and expresses key non–functional properties and gives the system high flexibility.”
Though an SOA architecture allows you to configure a system dynamically, this is contradictory to safety requirements. Because you have limited resources, you have to ensure that your applications that are safety–relevant are prioritized and that they are running in the lead time and in the order that has been defined.
“We try on one side to make the life of the designer easier by providing the dynamic configuration mechanism, and then on the other side, we ensure with MotionWise that the applications that are safety–relevant have a deterministic behavior,” Pickhard said.
This is a paradigm of SOA, which on one side makes life easier for engineers by reducing the complexity, and on the other side ensures safety and determinism.
“The next step is to win the market with DDS, making it a high–performance safety platform for SOA architectures in a car,” Pickhard added.
This article was originally published on EE Times.
Maurizio Di Paolo Emilio holds a Ph.D. in Physics and is a telecommunication engineer and journalist. He has worked on various international projects in the field of gravitational wave research. He collaborates with research institutions to design data acquisition and control systems for space applications. He is the author of several books published by Springer, as well as numerous scientific and technical publications on electronics design.