BrainChip Accelerator is an 8-lane, PCI-Express add-in card based on FPGA hardware that increases the speed and accuracy of the object recognition function of BrainChip Studio software by up to a factor of six, while increasing the simultaneous video channels monitored to up to 16 per card.
The unnamed automaker is evaluating BrainChip Accelerator for Advanced Driver Assistance System (ADAS) and Autonomous Vehicle (AV) applications.
The spiking neural network approach does not require the protracted, usually cloud-based training programs, required for convolution neural networks and even allows "one-shot" training and self-learning procedures. The approach would be further improved by a dedicated processor based on the SNN approach could be the next development from BrianChip.
The low-power accelerator card can detect, extract and track objects using a proprietary SNN technology. When FPGA based it provides a 7x improvement in images/second/watt, compared to traditional convolutional neural networks accelerated by GPUs, the company claims.
Related links and articles: