Driver assistance

Reference projects

Measurement system

Data from different sensors are synchronouly recorded in ADAF. Simultaneously, multiple applications can be executed to process the data on-the-fly. Hardware can be selected depending on performance reqirements (laptop, industrial PC, embedded PC) .

Generation and management of ground-truth data

ADAF Annotation Toolbox:

  • Data annotation to support application development
  • Data annotation for quality assurance in series production
  • Client-side onfiguration of annotation categories (globally, intervals, ROIs) in XML dialect
  • Automatic creation of database layout (e.g. Oracle or PostgreSQL)
  • Remote access to annotation database
  • Automated offline annotation
  • Tracking algorithms for increased performance
  • Versioning
  • Synchronisation of data from different sensors
  • Support for customer-specific file formats

Speed limit recognition

Setup and provision of the recording and measurement system to investigate fusion of video-based speed limit recognition and map data information. In parallel, HD video (1920x1080, 25 Hz), CAN data from three CAN-buses, and DGPS signals are synchronized, recorded and visualized using ADAF.

Driver assistance for motorcycles

Comparison of detection rates of different objects in recordings obtained from an instrumented car with those obtained from a motorcycle. During the project, video, CAN, and GPS data was recorded for both the car and the motorcycle. The annotation of relevant objects in the recordings was done using ADAF Annotation Framework in semi-automatic mode. The annotation data were stored in a SQL database enabling systematic assessment of the automated annotation. The automatic evaluation of the system performance ( lane detection and speed limit recognition) was conducted using ADAF Validation Framework.

Gesture recognition

The gesture recognition system for control of a head-up-displays was based on TOF cameras by PMDTec . The reliable recognition of gestures in varying lighting situations was achieved with data of the TOF cameras und specifically developed processing plugins for ADAF.

Driver in the Loop

Studies of driver behavior were performed using NISYS TRS. Data of the ego-vehicle and the surrounding vehicles were recorded for subsequent evaluation. The simulator was also used in research projects to study the behavior of multiple interactive drivers sharing the same road in simulated environment (synchronisation of multiple simulator instances).

SIL tests of driver assistance systems

Software-in-the-loop (SIL) tests allow performance estimation of the future driver assistance applications already during the development process. SIL testing are an essential building block in the release for series production of a new product, especially for intervening assistance systems. TRS and ADAF provide the required functionality.

HIL testing of driver assistence systems

The evaluation of ADAS prototypes requires Hardware-in-the-loop (HIL) tests. Performance evaluation and long-term tests are easily and cost-efficiently realizable with TRS.

NISYS GmbH© 2013