ADAS
Get Your ADAS Projects on Track Quickly and Reliably

Advanced Driver Assistance Systems (ADAS)

Driver Assistance Systems  

  • support the driver

  • increase safety

  • improve comfort and economy

The decisive success factor for supporting and relieving the driver is the complete perception of the environment with all relevant objects. This also includes the reliable estimation of the future development of a traffic situation. To this end, various sensors record the vehicle environment, the sensor data are analyzed and merged in the control units. 

Whether you develop, verify and test your ADAS applications in a real or virtual environment: Vector offers you comprehensive solutions in the form of software and hardware tools as well as embedded components. 

Advantages

  • Complete and consistent ADAS tool chain consisting of tools, software components, algorithm framework and hardware 

  • Fast development, debugging and testing of multisensor applications from the algorithm prototype to the production ECU  

  • Logging of multisensor data in the vehicle 

  • Efficient sensor and control unit connection for data acquisition 

  • Rapid prototyping of driving functions with real measured data 

  • Recording of test drives in the laboratory 

  • By object overlay you verify your sensor objects in the video image of the environment and in the bird's eye view  

  • Accelerated development of sensor manufacturer-independent data fusion or as a redundancy path for own environment models with BASELABS Create Embedded 

Application Development

The short way from application development to the control unit

vADASdeveloper provides the infrastructure for developing algorithms for driver assistance systems (ADAS) and automated driving. The runtime environment captures sensor data, records it and can replay it for stimulation. vADASdeveloper accelerates the development process, increases the quality of the systems and offers the ECU developer a high degree of flexibility. 

Adaptive MICROSAR is the Vector solution for vehicles based on the AUTOSAR Adaptive Platform. It includes the AUTOSAR Runtime for Adaptive Applications (ARA) and comprises an efficient development environment integrated in Eclipse. Adaptive MICROSAR is particularly suitable for high-performance ECUs such as in-vehicle application servers, ADAS ECUs and infotainment systems. Vector offers you a secure solution up to ASIL D, from the hypervisor to the POSIX operating system to ARA. 

Code

Data fusion development for automated driving functions up to series production

BASELABS Create Embedded is a software solution for the fast and efficient development of data fusion systems for automated driving functions in embedded systems. The tool contains fusion algorithms that combine data from radar, camera and lidar sensors. The resulting object fusion provides a uniform object list of the vehicle environment and serves as a basis for path planning and decision algorithms. BASELABS Create Embedded makes it possible for the first time to apply the developed sensor data fusion directly to series control units.

Executable Platform

Depending on the stage of development, you may want to use your algorithm on a virtual or real platform and test it there. 

DYNA4 is an open simulation environment for virtual driving tests with cars and trucks. The 3D environment simulation with road infrastructure and traffic provides a test environment for assisted and automated driving. The environment is captured by sensor models in different pre-processing stages from raw data to idealized object lists. 

Closed-loop simulations with DYNA4 run on the PC in multiple real-time, e.g. for use in early development phases (MIL, SIL), and can be executed on HIL systems as soon as control units are available in hardware. 

vVIRTUALtarget is a software for generating virtual ECUs for all typical AUTOSAR projects. It is therefore ideally suited for the development of fusion ECUs. vVIRTUALtarget runs on all typical Windows PCs. The application is executed in real time or independently. 

 

ECU / Sensor Results

Sensor and other vehicle data are collected in very different ways

XCP is the ASAM measurement and adjustment protocol for accessing ECU internal data. Either you use an XCP protocol driver in your application or you use the VX1000 solution. In this case, a small circuit board - a so-called "Plug On Device" (POD) - is connected directly in the ECU to the debugging or trace interface. The serialized data is transferred to the VX1000 base module and from there it is passed on as XCP data stream to the PC or logger. 

You can access the bus data of CAN (FD), FlexRay, LIN and Automotive Ethernet via the VN product family.  

Fusion control units are increasingly based on AUTOSAR Adaptive. SOME/IP is used as the protocol. CANape and CANape log can both passively observe and actively participate in the SOME/IP protocol and thus access the data of the fusion ECU. 

The easiest way to acquire ground-truth data is to use our Lidar and GPS integration in CANape. The interface used for this purpose is open for the integration of other systems, so we can also seamlessly integrate your special products and protocols.  

In addition, you can access ECU and vehicle diagnostic data via the KWP2000, UDS and OBD diagnostic protocols.  

Measured variables, such as humidity, temperatures, voltages and currents are recorded via measuring modules, e.g. by CSM.  

ADAS Logging

The collection of ADAS data in the vehicle is complex and ranges from the physical connection of sensors to the data management of many terabytes per vehicle in a cloud.  

Physical connection of the sensors: 

Depending on the sensors used, up to three different data streams are generally available from one sensor. 

Raw data, e.g. radar reflections or video images from video sensors are collected either via the VX1000 hardware (Aurora interface, FPD Link, ..) or via Automotive Ethernet due to the amount of data. 

Data objects are recognized objects, e.g. other road users, which are transmitted in the form of signals via vehicle buses (e.g. CAN FD or Automotive Ethernet).  

Internal data from sensors and ECUs are transmitted via XCP over the VX1000 hardware or also via Ethernet. 

Recording of data in the vehicle: 

Multiple sensor data must be recorded synchronously. The CANape software used (or CANape log Smart Logger) has a scalable system of different recorders. A so-called DHPR (Distributed High Performance Recorder) is used for each type of high-performance sensor or manufacturer. It receives the sensor data and stores them in one or more files. The recordings are orchestrated via CANape or CANape log. 

Access to fusion control units: 

Fusion ECUs have a very powerful computer architecture and increasingly use AUTOSAR Adaptive as middleware and Ethernet as the physical interface. Data acquisition from a fusion control unit can be done in various ways:  

XCP-on-Ethernet: In one or more applications on the fusion ECU, XCP is integrated as a measurement and calibration protocol. CANape uses this method to directly access the data in the algorithms. 

PCIe measurement technology: At very high bandwidths, measurement data can be acquired via PCIe in the ECU. The POD (Plug-On-Device) first configures the measurement via Ethernet, then retrieves the measurement data via active DMA (Direct Memory Access) with minimal impact on the processor and sends it to CANape via a secure channel. 

SOME/IP: CANape actively participates in SOME/IP communication and can use this method to acquire data from the ECU. 

Controller Interfaces: In fusion ECUs with a combination of controller and processors the controller interfaces, such as Aurora or HSCT, are integrated into the logger via the VX1000 measurement technology. 

 

Robust logger solution for vehicle use: CANape log 

In addition to the sensor data, further vehicle data is of course required. Besides vehicle buses, such as CAN FD, Automotive Ethernet and FlexRay, ground truth data, coordinates and reference cameras are also recorded. CANape log provides you with a stand-alone logger for all vehicle data. CANape configurations are used directly in the logger. 

 

Data logistics - How does the data get into the cloud? 

Using the vMDM Client, as well as the CANape log data transfer tool in combination with the readout station for removable SSD media, there are various ways to transfer recorded data from the vehicle to the cloud, depending on the volume of data and the available network connection of the SmartLogger. The measurement data recorded as MDF and video files of the context cameras in AVI format are automatically uploaded to vMDM and pre-processed there according to user-definable criteria for efficient use. 

Sometimes, however, the solution is not only found in the product, but in the project-specific process. For example, for extremely large data volumes - especially raw sensor data - the connection of a user-defined storage solution to vMDM may be desired. In this case, network messages, ECU and A/D measurement data are stored in the Vector Cloud, and the externally stored raw sensor data is integrated with a project-specific link to the MDF measurement data. In this way, the raw sensor data of a measurement can be easily located at any time from vMDM with the metadata stored there and can be unerringly fed into further analyses and resimulation chains. 

The Vector application team will be happy to advise you on this type of storage system integration with vMDM - simply contact us with your specific requirements. 

Measurement Files

Due to its standardization and the high performance when writing and reading measurement data, the MDF format is not only ideally suited for classical measurement tasks, but also plays out its advantages when storing extremely large data volumes in the ADAS area.  

From the recording of simulated sensor and environmental data in the virtual DYNA4 road test as OSI3 objects, to the high-bandwidth logging of raw sensor data with CANape log in the test vehicle and the reuse of this data in the HIL test with CANoe, MDF provides the basis for efficient data storage and the smooth use of the recorded data - within the entire Vector ADAS tool chain and also in combination with user applications. 

Preprocessing

When logging ADAS data in the vehicle, extremely large amounts of data are generated. In order to sort out unsuitable measurement data before the actual data analysis or tests, the quality of the data is checked in a preprocessing step. The measurement files suitable for further work are enriched with metadata and - in the case of image data - pixelated if necessary. 

For example, the number of lost image data of a video sensor is a quality feature. If too much image data is missing in a recording sequence due to a sensor malfunction, this sequence cannot be used for resimulation purposes. CANape and vSignalyzer allow you to automatically identify and sort out such video clips. If a MobilEye SoC (System-on-Chip) is used in the camera control unit, the measurement system can also check this quality feature during recording and thus mark or sort out the corresponding video clip during logging. 

In order to use recorded data in a re-simulation in a meaningful way, they often have to be "labelled". This means that the recorded measurement files are enriched with information such as the number and type of existing road users, weather data and much more.  

Video sensors in particular inevitably record specially protected information with the environmental data: Faces and vehicle license plates. For data protection reasons, these must be pixelated. 

For the labeling of the data in MDF format as well as for pixelation with the help of trained AIs, the library MDFLib provides suitable interfaces for the connection of these algorithms - e.g. in Python. Thus, these steps can also be quickly and efficiently integrated into the post-processing chain. 

Test & Test Design

CANoe is the versatile tool for developing, testing and analyzing entire ECU networks as well as individual ECUs. It supports network designers, development and test engineers at OEMs and suppliers in the complete development process - from planning to commissioning of complete distributed systems or individual control units. This means that the same configuration including all models and tests can be used consistently in all phases. From MIL/SIL operation in purely simulated environments to the HIL test bench including all interfaces to communication networks and other I/Os. Interaction with DYNA4 for virtual road tests works seamlessly via the direct integration of the generated models.  This means that tests for ADAS functions can also be carried out easily. Appropriate display windows, such as Birds-Eye-View for analysis, are also available. 

vTESTstudio is a powerful development environment for creating automated tests. To increase the efficiency of test design and to simplify reuse, you can use 

  • programming language based, 

  • tabular and 

  • graphical test notations and test design methods are available

vTESTstudio can be used in all development phases: from model tests to system validation. Thanks to open interfaces, vTESTstudio can also be flexibly integrated into existing tool landscapes, e.g. for requirements management. 

Data Analysis

For data analysis vSignalyzer is available with the Driver Assistance option. You can use it to view and automatically analyze sensor data.   

vMDM is a powerful database solution as a cloud application for storing and managing measurement files. This can be operated completely as Software-as-a-Service in the Vector Cloud with minimal IT effort and fast availability - and if your organization operates its own cloud with AWS or Microsoft Azure - also there.  

For integration into your own environment we recommend MDFlib, which allows you to access measurement data conveniently and efficiently without having to worry about the measurement data format. 

Technical Article with Useful Know-How

Graphic: car with five different suppliers and five logging systems for ADAS sensor data
Example of an ADAS Level 3 setup with sensors from five different suppliers and five logging systems

Data Recording for ADAS Development

The automotive industry is working extremely hard on technologies for autonomous driving. To replace human perception of the environment, the use of driver assistance systems with high-resolution radar and video sensors is essential. Massive quantities of data that need to both be transported via the communication networks as well as processed in real-time are thereby generated. This creates challenges in an unprecedented dimension on recorder solutions as well. When selecting a suitable system, a number of important aspects must therefore be taken into account.

By using recorder solutions with consistently scalable architecture, developers of driver assistance systems obtain very flexible systems that can be expanded to a nearly unlimited extent. The data rates of 500 to 700 MByte/s required in current autonomous driving projects for the storing of radar, video and ECU data can still be managed at the present time with just a single PC. As requirements increase, investment security is ensured by the VX1000 hardware – which can be used for both ECUs as well as for raw sensor data – and by the scalability of the complete solution.

Photo Autonomous Driving
The complex hardware and software environment for autonomous driving raises the question of how release processes for this systems can be designed.

Combining Forces for ADAS Testing – Skillful Use of the Toolbox

When heading at full speed into the new ADAS world, it is better if there is no risk to any of the parties involved. How do we handle the current IT paradigm change in a way that makes the testing of all the functions truly reliable? Do we need more complex or completely different tools? Or would the skillful use of existing tools be a suitable approach?

Case Study

Graphic Model-in-the-Loop test system for automated evaluation of driver assistance functions
Model-in-the-Loop test system for automated evaluation of driver assistance functions (schematic)

Virtual Model-In-The-Loop Test for the Development of Driver Assistance Systems

Due to the ever-increasing complexity of driver assistance functions, combined with the integration and fusion of different sensors, it is necessary to evaluate the performance of those systems as early as possible in the development cycle. Ideally, this is done automatically and already during the modeling phase. If vehicle dynamics influences the detection performance of the sensor system, or limiting cases need to be taken into account, the test system also needs to be able to test the Device Under Test (DUT) in a closed-loop. Having appropriate, application-specific metrics – especially for the validation of data fusion – is another requirement. The case study shows a solution based on tools from TASS, BASELABS and Vector.

The Challenge: Evaluating data fusion and driving functions early in the development cycle

The Solution: Automated testing of the models of algorithms and functions in virtual environments

The Advantages: Early objective evaluation of driver assistance functions and data fusion by means of complex virtual scenarios

Tools & Hardware

The following products from Vector assist you in validating object data and testing image processing algorithms:

This measurement and calibration tool lets you acquire sensor and ECU data, optimize an algorithm and stimulate real or virtual ECUs.
This optional CANape extension lets you verify the results of the ADAS algorithms. Detected objects from the vehicle are shown from a bird's eye perspective and superimposed on the video image.
vADASdeveloper provides an infrastructure for developing algorithms for advanced driver assistance systems and automated driving. The tool reduces your workload in developing sensor data fusion applications.
Algorithm framework for the data fusion of camera, radar and LiDAR and the efficient development of data fusion applications for series production. BASELABS Create Embedded enables dynamic object fusion for automated driving functions – from prototyping to series production.
This measurement and calibration hardware utilizes the standardized XCP-on-Ethernet protocol for high-performance data transfer between the ECU's controller interface and the PC.

Embedded Software

In the area of embedded software, the MICROSAR product gives you a complete solution for basic software according to AUTOSAR. The following MICROSAR functions and properties are especially relevant to ADAS projects:

Ethernet
MICROSAR ETH gives you an AUTOSAR-based communications stack for exchanging data in automotive Ethernet networks. MICROSAR.ETH provides the necessary functionalities and flexibility for new types of ADAS applications.
AVB
You can use MICROSAR AVB to transfer your Ethernet-based audio and video data that is necessary for assistance systems.
Safety
The extensive lineup of MICROSAR Safe products lets you implement safety-related applications according to ISO 26262.
Security
Protect your applications against external attacks with MICROSAR Security. In Car2x radio communications, you can use the cryptographic algorithms it contains – such as Message Injection and Man-In-The-Middle Attacks.
Time synchronization
MICROSAR SYS lets you time-synchronize distributed applications on a wide variety of networks (CAN, FlexRay, Ethernet). For instance, a common time base makes it possible to merge measurement data from different sources.

Downloads