Current Research Projects:

• Active Noise Control for MRI noises: Active Method is used to control the noises generated by MRI scanning to eliminate the negative influence for medicla diagnosis and patients' hearing.
• Music Injection for Subjective Speech Enhancement: We aim to improve the speech from the perspective of human being's sensation by inserting music into noisy speech processed by the spectrum subtraction method.
• Blind Dereverberation: Reverberation degrades the performance of ANC. Linear Prediction method is adopted to eliminate the reverberation based on SIMO model.

Previous Research Projects:

• Digital Video Processing System:
• Direction Finding System:

Development:

• Software Development for Cell Phones in Motorola

Active Noise Control for MRI Noises

Active Method is used to control the noises generated by MRI scanning to eliminate the negative influence for medicla diagnosis and patients' hearing.

Music Injection for Subjective Speech Enhancement

Problem:

Conventional noise reduction methods improve the SNR of speech signal. However, the residual noise makes the speech annoying.

Our Idea:

• Music Injection is adopted to reduce the negative effect of residual noise and improve the pleasantness of speech.
• Volume of the injected music is adjusted adaptively.
• An objective pleasantness model is utilized to evaluate the performance quantitatively.
• Subjective test is also carried out to verify the validity of the above pleasantness model.

Flow Chart:

Blind Dereverberation

Room reverberation degrades the performance of ANC. We will add dereverberation into the system to improve ANC performance.

Model:

Algorithm flowchart:

Digital Video Processing System

This system, funded by China 613 Research Institute with $70,000, aims to improve the image quality of the airplane’s back cockpit. In especial, the instrument board is recognized by color with Ostu's method. I'm responsible for algorithm designing and software developing.

Software: MFC and VFW is utilized to process real-time video signals from capturing board.

Hardware: the system is implemented with a 12-layer board based on AD21160 DSP.

Direction Finding System

Direction Finding is important for Radio Supervision Department. Analog product is found instable in practical use. In this project under the contract with an electronic compnay, we aim to development a digital system based on DSP techniques to fullfill the Direction Finding fast and accuratedly. I'm responsible for algorithm designing, software developing and board debugging. It has now been put into mass production.

Algorithm: Digital phase detecting algorithm is developed based on the 90 degree phase-shift of Hilbert transform to detect the phase difference of two coherent signals.

Software: TMSC54X assembly language

Hardware: A PCI interface board is developed, including DSP, CPLD, PCI bridge, A/D converter

Software Development for Cell Phones in Motorola

Cell phones I have ever been working on:

...

E380(GSM), V500(GSM), V600(GSM), V265(CDMA),A860(Worldphone, dual mode:GSM+CDMA),E815(CDMA: EVDO) V3c(Razor, CDMA), etc.

Certificate of Excellence: