Our Expertise
We offer consulting and subcontracting services in the areas described below.
Spectrum of work starts with consulting for a few hours and address a local problem up to Turnkey Projects,
Writing code, Optimizing DSP code, Drivers and more....
Team up with more experts to meet the engineering and time needs of the project. Other disciplines involved like Optics, Mechanical Design, Board Layout etc might be integrated by us.
Image and video processing including motion detection algorithm optimization.
Adapting standard libraries to specific application and optimizing code for performance. Implementing algorithm written in Matlab, C or Python.
Working with sensors inputs that need filtering or other signal conditioning. Low frequency sensors as well as voice and sound inputs.
Control loops as in Power Supply or Motor Control.
Implementation and performance optimization on CPUs or MCUs.
As algorithms become more complex there is a need for optimization.
The DSP has unique features and instructions that standard C code might not exploit. There is also a need for system design for performance.
Optimization sometimes requires using Intrinsics or Linear Assembly. Extensive benchmarking is used during the optimization process.
Image Processing
Signal Processing
DSP Code Optimization
System architecture requires deep knowledge of the components details. The system design includs components selection including CPU or SOC, memory, peripherals, adding external processing power like FPGA etc.
System design also includes internal software topology, memory allocations and many more details.
Embedded Linux is used extensively in SOCs. The Linux system contains almost all the needs for the devices. The user application must be designed carefully and occasionally adaptations of drivers or moduls is needed.
More and more SOCs contain more then one CPU. Mix of CPU kinds is also popular for example having a DSP and an ARM on the same chip.
Working with multicore devices requires solving load balance issues, inter processor communications and synchronization.
Multicore systems
Embedded Linux
System Architecture
ADAS systems have become common in vehicles. Algorithms use Image Processing tools and Neural Networks. (NN , CNN)
Examples in the videos below :
Lane detection.
Vehicle detection.
Both are implemented in Python using CV and other Python libraries.
ADAS systems algorithms
