Open search

DSP for camera-aided driver assistance systems

Analog Devices (USA) has introduced the Blackfin BF609 and BF608 digital signal processors (DSP). They are designed for driver assistance units requiring image processing, and optimized for embedded vision applications. In particular, active pedestrian detection and protection systems may benefit from these processors. The dual-core chips provide CAN connectivity to be linked to other automotive electronics.

THE ADSP-BF60x SERIES OF DSPs with A 2 x 500-MHz core performance and a dedicated vision accelerator enable broad adoption of multi-function analytics into embedded vision applications. The Blackfin BF608 and BF609 dual-core chips feature a video analytics accelerator, called the Pipelined Vision Processor (PVP). The PVP is comprised of a set of configurable processing blocks designed to accelerate up to five concurrent image algorithms, enabling a very high level of analytics performance.
These processors are dedicated for applications such as automotive advanced driver assistance systems (ADAS), industrial machine vision, and security/surveillance systems.

The CAN interface allows an integration into the in-vehicle networks. Data already available in other ECUs (electronic control units) can be shared via the in-vehicle network. The DSPs also comprise 112 general-purpose I/O, one USB port, two UARTs as well as other preipherals.

“With over 25 billion mathematical operations per second available from the PVP, this accelerator combined with two Blackfin cores provides the basis for a very powerful and flexible processor” said Colin Duggan, Analog Devices’ director of marketing. The DSPs are supported by the CrossCore Embedded Studio. This Eclipse-based IDE enables support for both proprietary and open source tools and technologies including Analog Devices’ C/C++ compiler, Micriµm’s µC/OS-III, Linux and GCC. The American company works closely with Micriµm to integrate the µC/OS-III real-time operating system, USB drivers, and file system into the Embedded Studio. The µC/CAN protocol framework enables the implementation of CAN communication paths. It is a source code library. Developers need only an understanding of signals, messages and bus configurations. Different abstraction layers can be used independently. The software supports multiple CAN ports. The bus management layer organizes the CAN networks and distributes messages to different device drivers. The device driver layer buffers the CAN messages to be sent and received. All hardware dependencies are capsulated in this layer. An embedded target can also have multiple, different CAN controllers. A validation suite provides all documentation necessary to support the use of µC/CAN in safety-critical systems.