témavezető: Vásárhelyi József
helyszín (magyar oldal): Institute of Automation and Infocommunication helyszín rövidítés: AUT
A kutatási téma leírása:
With the era of smart phones, “smart world” appliances, self driving cars, behavioral control of embedded systems speech recognition become more and more important. Language specific speech recognition is particular for each language. A speech recognition system has to analyze, to filter, and digitize the sounds. After that speech interpretation and the control action depends on embedded system application. Digitization also means to transform the resulted digitized speech in a format the can be read and then analyzing it for meaning. The processed information, the based on several algorithms and previous input it can make highly accurate guess about what one said. If the speech recognition system is used by only one person the system can be trained specifically for that person.
The problem become more complex when an embedded system (eventually smart device) is used by multiple persons or more then that on different markets around the world, which means different cultures, different languages.
The research should be focused on the speech recognition using re-configurable technology such as Field Programmable Arrays (FPGA), All Programmable System on Chip (SOC), Adaptive Compute Acceleration Platforms (ACAP) for implementing speech recognition system for embedded systems and find applications on robotics (self driving robots), behavior control and etc. The research should focus on using Artificial Intelligence methods in the speech recognition and these methods should use the above mentioned hardware platforms: “the most dynamic processing technology in the industry, enabling rapid innovation with its adaptable, intelligent computing ” (Xilinx).
Bejelentkezés
