Transcript ppt
RoboTechTronix Ryan Fonnesbeck (CS and CE) Brian Clay (CE) Justin Hansen (CE) Project Description Voice Interactive Robot Recognize voice commands. Recognize trained keywords or expressions. Respond to voice commands. Verbal feedback and interaction with the user. Non-verbal feedback such as navigation. The Use of Nueral Networks Neural Networks use a set of processing elements loosely analogous to neurons in the brain. These nodes are interconnected in a network that can then identify patterns in data as it is exposed to the data. After first processing audible data as discrete sound frequencies, data is pumped through a series of Fourier transformations and then sent to a neural network for processing. This complex and sophisticated process allows us to successfully perform speech recognition. Project Motivations Useful natural interface Push Limits of Technology real time natural processing Cool. Project Overview Voice Interactive Robot JStamp – Real-time native execution Java hardware and software micro-controller. VoiceDirect 364 – State-of-the-art speech recognition technology. ChipCorder 2560 – Single-chip record/playback technology. Motorola 6210 – Two-way radio. LM4863 – Audio amplifier. Functional Partition Implemented a serial protocol using the JStamp I/O pins and VoiceDirect 364 I/O pins. Utilized JStamp I/O pins and VoiceDirect 364 I/O pins to control the ChipCorder 2560. Software Flow Chart Packaged software to allow others to easily re-use and implement their own voice interactive robot. Software modes allow for run-time hardware configuration. Problems Encountered Operational range for the VoiceDirect 364 Baby Monitor Walkie-talkies Voice recognition accuracy Noise filtering VoiceDirect 364 configuration Speaker dependent vs. Speaker Independent Number of words vs. accuracy VoiceDirect 364 Serial Protocol Latency Time constraints Conclusions Plan Initial project plan Limitation of Current Voice Recognition technology Importance of documentation Demonstration Questions ???