Algorithm And System Research For Wireless Implantable Brain Computer Interface Based On Information Compression

Brain-Computer Interfaces (BCIs) establish a communication pathway between the nervous system and the external environment, to achieve direct interaction between the peripheral or central nervous system and the devices, to provide the disables with motion control. Brain signal acquisition and real-time signal transmission is the key to the success of brain computer interface. Wireless transmission of the information in BCIs is one of the growing research subjects. Wireless implantable BCI platform can make the animals free when they were in the experiment, to extend the research methods, and promote an invasive BCI development. In this thesis, we focus on t information compression method used for wireless BCIs.Information compression is to process the original data before the wireless transmission of BCI information without losing essential information for post-processing, or to represent the original data using limited data set with data compression method. In this thesis, both signal feature extraction and data compression were studied. For signal feature extraction, we focus on spike detection and classification using nonlinear energy operator and matched filtering to calculate neural spike sequence data for decoding. The experiment on simulation data proves that this method is efficient for spike detection and classification. For data compression, we focus on vector quantization aspects with designing a three-stage strategy framework to keep the original neurological data waveform. The experiment result shows that the strategy worked well on data compression and restoration.The main work of this paper is to study the method of spike detection and classification, and neural signal compression algorithm, and to design a hardware system. The system includes (1) an implantable neural signal acquisition module to collect and amplify the neural signal, (2) digital signal processing unit for information compression, (3) Bluetooth module for wireless transmission. The results confirm that the offline algorithm is feasible and effective. The signal acquisition module is validated in a two-direction system. The compression algorithm was off-line tested on the DSP system. The serial transmission results show the high feasibility of the digital signal processing unit and the wireless transmission module.