A new architecture is suggested for a numerical estimator using Artificial Neural Network (ANN) technology. The method is based on proven technology already implemented in hardware, and is extended for improved numerical accuracy combined with reduced resource usage. The architecture is suitable for embedded hardware solutions.
Improvements to Function Approximation Using a Hardware-Accelerated Artificial Neural Network
Sae Technical Papers
SAE 2006 World Congress & Exhibition ; 2006
2006-04-03
Conference paper
English
Improvements to function approximation using a hardware-accelerated artificial neural network
Automotive engineering | 2006
|British Library Conference Proceedings | 2006
|Approximation of Battery Transfer Function Using Neural Network
TIBKAT | 2020
|HARDWARE ACCELERATED NETWORK INTERFACE FOR AN AUTONOMOUS VEHICLE SWITCHED-NETWORK
European Patent Office | 2021
|Hardware accelerated multichannel receiver
IEEE | 2009
|