Axel Techmer

A 100-GOPS programmable processor for vehicle vision systems

New generations of automobiles will include driver assistance systems requiring powerful, low-cost processors to handle video/camera applications and to enable fast, convenient application development. Shrinking feature sizes on processors already in development will bring substantial increases in system speed and functionality.

Contour-based motion estimation and object tracking for real-time applications

A new approach for fast motion estimation and object tracking is presented. Motion is estimated along contours, which are tracked over time. Contour tracking can easily be extended to object tracking. To track contours a transformation is estimated which allows an optimal match of contour parts. This is measured by distance values. Using the distance transformation of contour images, contour tracking can be formulated as a correlation approach. Contour tracking works robustly in real-world image sequences. Furthermore it allows real-time implementations of motion estimation and object tracking on a Pentium III processor. This is essential for using motion information in real-time applications.

Implementing FlexRay on Silicon

FlexRay^1, a high speed, time triggered and fault tolerant communication protocol, was specified to fulfil the requirements of safety-critical automotive applications. The maturity of the FlexRay specification already allows the implementation on silicon. The CZC-310^2 device is a FlexRay standalone communication controller from Infineon Technologies. It includes the widely spread E-Ray^3 from Bosch. A complete communication node for FlexRay requires additional devices for the physical layer and the application part. CZC-310 can communicate with a host controller via three different interfaces (micro link interface, serial synchronous interface, external bus). Its physical layer interface corresponds to the FlexRay specification. CZC-310 provides features like intelligent move engines to maximize the achievable data rate as well as to minimize the workload of the host. Therefore, CIC-310 allows a very flexible and efficient way to build up FlexRay nodes.

Embedded vision platform for video surveillance systems

Automatic video surveillance systems are used in many different fields of applications like intelligent vehicles, intelligent highways or in security tasks. An automatic and correct interpretation of video sequences is based on the detection, tracking and classification of various objects under highly diverse conditions. This requires highly sophisticated algorithms as well as high computational performance to fulfil real-time constraints. Here, Infineon Technologies develops a fully programmable, scalable multi-processor architecture optimized for video processing, which provides a processing performance similar to actual PCs but at much lower costs, lower power consumption and lower physical size. This architecture supports data-parallelism for low-level image processing, parallelism of tasks in the medium- and high-level as well as control oriented processing. A cycle accurate, virtual prototype of the architecture is available. A library of optimized image processing functions supports a comfortable application development and the reuse of existing application software. Beside the implementation of standard low-level operators new efficient approaches for motion estimation, object detection and tracking are developed and tested in applications for intelligent vehicles and intelligent highway scenarios. The integration of these application specific tasks into the image processing library results in a powerful embedded vision platform for video surveillance systems, which supports a comfortable application development.

100 GOPS vision processor for automotive applications

Car vision systems for the improvement of drivers comfort and safety will be introduced in the next car generation, with an exponential market growth in the next years. The Corporate Research division of lnfineon Technologies AG has developed a fully programmable vision processor, which fulfills the requirements of low power, low system costs, and high computational performance (e.g., robust approaches for object detection based on stereo processing need a performance of more than 15 GIPS).The prototype currently available achieves a peak performance of 53 GOPS. With a redesign in 0.13 μm CMOS technology, the performance increases to more than 100 GOPS, while the power consumption is less than 500 mW and the die size reduces to 70 mm2.