Since the very beginning days of developing intelligent robots, achieving higher speed in data processing has been one of the most problematic challenges for designers. Intelligent robots have always been coupled with heavy and complex calculations, which are vital for them to be computed in the least time, achieving the least delay.
Regular processors known as CPUs leave designers with no option but to replace accurate and expensive algorithms with inaccurate and easy to compute ones to achieve the ultimate goal which is real-time processing.
This paper suggests a method to boost data processing speed by executing some result-independent functions on the GPU. This method results a dramatic improvement in processing speed, promising emersion of some powerful but expensive, thus abandoned algorithms.
Here is our paper: A Novel Approach to Real-Time Processing: Implementing Complex Calculations on GPU