Prototype Workshop


In the in-house workshop, we develop and manufacture customized special components and can equip complete prototype vehicles for the testing of new systems. Our specialist know-how ranges from individual measurement solutions via special mechanical concepts for vehicle mountings to the complete cabling and commissioning of prototype vehicles


Our services

 Measurement technology concepts.

 Assembly of individual measuring technology concepts.

 Construction and cabling of vehicle prototypes.

 Standardframeworks for test automation.

 PODBOX for rapid prototyping and as a diagnostic platform.

 Individual solutions for laboratory and vehicle testing.

 Special solutions for brackets, for example camera shades.

 Construction of individual crash targets.


Project Examples

 PODBOX - diagnostic platform in prototypes; Use as a reporting tool in the vehicle.

 Development and construction of small series such as measuring techniques, brackets, seat wedge, among others.

 Development of customized special solutions, eg camera aging.

 HIL test stands - design and construction.

DEVELOPMENT OF EMBEDDED SOFTWARE


As experts in the field of embedded software, our specialist know-how means we can support our customers with the most complex project requirements.

We develop embedded software systems for prototype and standard-production applications. We bring extensive experience in the development tool landscape and processes. Our team is also up-to-date when it comes to the development of current microcontrollers and DSPs

Another team focuses on the right branch of the V-model from verification through to validation of the individual components. Here too we are familiar with the popular tools for white-box testing, static code analysis and test automation.

Topics

 Automotive Embedded Software for series applications

 Automotive Embedded Software for various microcontrollers and DSPs

 Embedded C

 Various debugging environments

 Integration of AUTOSAR Stacks

 Communication interfaces (CAN, FlexRay, …)

 Communication protocols (KWP, UDS, ..)

 Strategies for Functional Safety (ISO 26262)

 Test automotion (CAPL programming)


HIL


The hardware-in-the-loop (HIL) test is an increasingly important validation method for environment sensors and other control units. With HIL, the real hardware components from the vehicle integrate into a simulated sensor/actuator environment for testing. The system is connected via inputs and outputs to a simulator that emulates a real environment.


The EC Mobility HIL test bench comprises the following components:

 Hardware under test (DUT)

 Assembly of individual measuring technology concepts.

 Output and analysis unit


We provides the following HIL services:

 Planning complete HIL systems

 Specifying the HIL environment and tools

 Managing/updating HIL systems

 Conducting HIL tests, for example radar sensor (77 GHz)

 Automating HIL system testing

 Developing tools, for example:

SIL


With software-in-the-loop simulations, the software is simulated independently of the target hardware. Typically a framework is available for re-simulating individual software components and algorithms. Software-in-the-loop tests are flexible and do not require complex hardware equipment. The system models used can be re-used for subsequent HIL tests, thereby permitting a comparison of the different test runs.​

DATA LABELING


In the development of driver assistance systems (especially in camera-based processes), algorithms of machine learning are often used. Both for the so-called training of these algorithms, as well as for the validation and testing of existing classification methods, a very large quantity of labeled data is required

In the case of such databases, the sections which are required for further use are already marked and marked with a label. Thus, for example, all traffic signs which are occurring are identified for road sign detection and provided with the information as to which sign it is. The shield information is referred to as a label in this context.

The quality Of the training and test data is essential for the development of various ADAS systems. The manual creation of this data is extremely time-consuming and therefore cost-intensive. EC Mobility offers all the services related to the topic of labeling, including quality control. In addition, we develop our own tools, which can be used to carry out labeling and which are used to optimize costs.

CMORE Label Tools:​​


The EC Mobility Tagging Tool (CTT) makes it possible to optimize the creation of a labeled database by greatly reducing the amount of data that needs to be considered. CTT offers the possibility to tag (tag) sequences that contain relevant objects for labeling during the recording via a user interface. The advantages are, on the one hand, the significant reduction in the effort and, on the other hand, the simultaneous storage of metadata.

A further approach to the optimization of labeling is investigated within the framework of the project Cost Sensitive Learning (CSL). In doing so, a tool is developed with which it is possible to automatically generate a large part of the labeled database by training a classifier during the labeling process. For more information, see CSL Project.

C.LABEL - THE INNOVATION OF LABELING BY ACTIVE LEARNING


Get high-quality labeled data with reduced cost and time input EC

The quality of training and test data is essential to the development of algorithms and functions for various automotive and industrial applications.

C.LABEL with Active Learning Loop (ALL) makes the user independently aware of objects and thus improves the labeling process enormously, decreases the time taken and also provides an indicator of the quality of the labels. Instead of viewing and labeling the entire database, with ALL only a relatively small volume of data needs to be processed manually.

Active Learning Loop (ALL)​​

Thanks to Active Learning, the user is assisted in the annotation of objects so that he or she is only required to make minimal adaptations. The detector for automatic labeling is updated constantly using corrections provided by the user. This causes the algorithm to learn independently, therefore improving the accuracy of label prediction.


Advantages:

Automatic prediction of labels

Optimization of the detection algorithm

Reduction and optimization of training data set

Increasing label quality

Automatic calculation of KPI is regarding the labeling process

Flexible definition of object categories and properties

Friendly, swift and sleek user interface


Areas:

ADAS, Active Safety, Autonomous Driving, Connected Car,Infotainment, Mobile Machinery, Industrial Automation

Development of Alogrithms


EC Mobility has many years of experience in the development of algorithms. As well as developing algorithms in numerous customer projects for radar and camera systems, we are also using our own algorithms for object detection and pattern recognition within the framework of the Cost Sensitive Learning (CSL)project.

The method is basically the same for all sensors. Today's radar, camera, lidar or ultrasound-based driver assistance systems work by detection and classification algorithms which analyse and interpret the information provided by the sensors. The first step involves preprocessing the raw data, e.g. stripping characteristics which describe images. The second step uses learning methods to train models using the identified characteristics. This is done using methods and algorithms from the area of machine learning. Particular challenges in the development of these algorithms are firstly correctly evaluates the sensor data, some of which may be noisy and contain errors, and secondly being able to perform the calculations required for this in real time.

Optimizing safety-critical driver assistance functions like Autonomous Emergency Braking (AEB) or Lane Keep Assist (LKA) requires a comprehensive environmental model provided by means of sensor data fusion, particularly for autonomous driving. This environmental model can offer the algorithms greater precision and safety as input data, which further increases their reliability.

Data Management


The increasing range of functions with regard to driver assistance systems and active safety systems in the vehicle requires ever more powerful sensors. The steadily growing flow of information and the need to merge the data streams to interpret increasingly complex traffic scenarios increases the need for new concepts for data management for the development and validation of the systems.

In order to capture, manage and store the data, the highest demands are placed on the hardware landscape and the software tooling. BigData is the umbrella term and refers to data sets that are too large, too complex, or too fast to be processed with manual and classical methods of data processing. Therefore new and secure access mechanisms are required.

EC Mobility provides adequate concepts and solutions such as data entry, categorization and labeling, databases, server HW, access mechanisms, data distribution, simulation and analysis.

Test & Validation Process

SYSTEM VALIDATION IN THE VEHIC


Vehicle validation


Vehicle validation is carried out on a system level, usually on a test track and in public road traffic.

In public transport, these are performance tests in which the functions are tested under predefined conditions (road, weather, traffic situation) and with special driving maneuvers. In addition, the function can also be tested without specification of concrete profiles, ie, random scenarios are traced in principle.

On the test track EC Mobility plans and carries out development-accompanying and standardized tests .

Whether testing of sensor system limits, field of views or speed-dependent performance evaluation, we support our customers from test planning to the evaluation of the data packages.

Benchmarking

Often, the current market requirements and therefore the latest state of the art must be recorded for the development of new functions and sensors. This requires extensive studies and tests with serial vehicles. EC Mobility equips this with customer-specific measuring technology systems and conducts the studies. The collected data are summarized in a statistics and presented in a performance comparison.

Worldwide permanent runs


In the protection of the sensors and functions much data must be recorded and evaluated. This requires a stable measuring technology system, experienced drivers and a corresponding concept for managing the data that have been generated.

EC Mobility supports the planning of the concepts, carries out the journeys independently according to specifications and supplies the customer-specific reports.

Continuous feedback loops with the customer increase the efficiency and quality of the data.

With hundreds of thousands of kilometers traveled, EC Mobility has a great deal of experience in this area.

Tool Development


EC Mobility offers standardized tools in the area of measurement technology for developing and validating vehicle components, for example via the PODBOX. These tools can be used directly in the vehicle are in the laboratory.

In addition, we can also develop custom tools for a customer's specific needs. We have a multitude of already implemented interfaces and dialogs we can use for this.

In the case of such databases, the sections which are required for further use are already marked and marked with a label. Thus, for example, all traffic signs which are occurring are identified for road sign detection and provided with the information as to which sign it is. The shield information is referred to as a label in this context.

The quality of training and test data is essential for the development of various ADAS systems. The manual creation of this data is extremely time-consuming and therefore cost-intensive. EC Mobility offers all the services related to the topic of labeling, including quality control. In addition, we develop our own tools, which can be used to carry out labeling and which are used to optimize costs.

Overview:​​​


Development in C/C++, Python, Node.js, etc.

Prototyping in Matlab

PC-based tools for optimising the validation process

ALL (Active Learning Loop) labelling tools

Test automation

HIL tools

SIL tools

Tools for data management

THE NEW SOLUTION FOR REMOTE PROTOTYPING AND VALIDATION


Use your project time effectively with remote access to your prototype


In the development of driver assistance systems (especially in camera-based processes), algorithms of machine learning are often used. Both for the so-called training of these algorithms, as well as for the validation and testing of existing classification methods, a very large quantity of labeled data is required.

Wherever your prototype may be, whether on the move or simply located elsewhere, you can easily access data, solve system issues and monitor it directly and instantly from your laptop, tablet or smart phone. PODBOX (Persistent Onboard Diagnostic Box) as a multifunctional platform offers custom measurement technology solutions for use in the vehicle or in the laboratory and additionally for prototyping algorithms and functions.

Basic functions:

 PB DataLog - data logger for CAN (can be extended for LIN and other interfaces)

 PB InVehicleReport - automated reporting for road trials

 PB InVehicleDiagnostic - central diagnostic unit for prototypes

 PB InVehicleTest - automated testing of all systems in vehicle prototypes

 PB Smart Meas Application - measuring technology applications (e.g. CTTView, TableView, CANTrace, etc.)


New functionalities:

 Storage of configurations

 Display of real-time data in multiple widges

 New user interface

Telematics solution for intelligent online data analysis​​​

The telematics solution utilizes the intelligent PODBOX server – C.LYTIX.SERVER – as well as an advanced online analysis toolchain and new protocols such as CCP and XCP. Further vehicle interfaces can be implemented on request. This feature enables online rapid prototyping and saves valuable time in all validation phases.

Advantages:

 Intelligent online data analysis during test drives

 Instant online statistics and KPIs

 Online pre-analysis

 Online monitoring of test fleets and track

 Online troubleshooting

General features:

 Easy system configuration

 Standalone operation

 Wireless LAN functionality

 Variety of logging options, e.g. CAN, LIN, GPS, temperature, etc.

Interfaces/connections:

 2x CAN

 2x Lemo or 15-pin D-sub for power supply via the vehicle's OBD connectors

 2x Ethernet with 100 Mbps

 WLAN

 2x USB 2.0

 1x RS 232

 12 V & 24 V power supply

Technical specification:

 ARM Cortex-A8 processor

 (800 MHz)

 External SD card memory

 Voltage range 8 V – 30 V

 Temperature range of the entire system -25 °C to 80 °C (without WiFi and battery)

PROTECTION BY STANDARDS


The complexity and specific requirements on the way to autonomous driving require new paths for the protection of the components and systems. From this, new standards are emerging, which are incorporated into the development processes and the performance evaluation of sensors and functions.

In the area of ​​test & validation, we carry out appropriate tests according to Euro NCAP or NHTSA specifications. This includes the planning and overall execution of the tests on the test tracks as well as the evaluation of the data that has been created. Customer-specific requirements in the field of tooling (measuring technology for data acquisition, reference systems, ...) are planned and implemented if required.

Currently, the standard tests use different targets that simulate real vehicles and people in order to determine the performance of the camera and radar sensors.

A challenge in the future will be the so-called VRUs (vulnerable road users), on which the future standardization will concentrate. This is essentially the most difficult to recognize for today's environment sensors, very mobile participants in road traffic such as bicycles, motorcycles or even wheelchair users. We are working on a research project to offer optimized targets for the simulation of real situations.