| The China Manned Space(CMS)program is entering the era of space station,and the manned lunar landing as well as other interplanetary exploration missions might also be conducted in the forward plan,thus requiring the astronauts to suffer long-term stay in space for complicated works.However,the astronauts will be confronted with great challenges due to the significant environment differences between the space and the earth along with the restrictions brought by the spacesuits.Hence there is an urgent need for human-machine interaction systems with higher performance that can assist the astronauts in working safely,healthily,and efficiently.Brain-computer interfaces(BCIs)could transform mind intentions into operation commands and realize natural interaction between the human and the machines,which have the potential to become the"third hand" of the astronauts.The steady-state visual BCIs have made great progress in recent years,whereas the applying of such systems in astronautic interactive operation is facing two problems,which lie in the uncertainty of the applicability in space,and the big gap between the current perfomance of codec technologies and the requirements of astronautic applications.To this end,this thesis conducted a series of researches on the verification of on-orbit applicability,the promotion of interaction speed,and the extention of information dimension for steady-state visual BCIs.Aiming at the problem that there is no data support for the applicability of steady-state visual BCIs and the experimental conditions are restricted in space,this thesis developed an experimental platform for on-orbit brain-computer interaction,which overcame the constraints of space environment by optimizing the design of hardware and software as well as the operation and procedure of the experiment.The astronauts could collect the data of multiple basic BCI paradigms in 30 minutes with an electroencephalography(EEG)acquisition device of 481g,and completed the onorbit experiments for three times in the space lab mission.This thesis analyzed the experimental results of the steady-state visual evoked potential(SSVEP)paradigm in different stages of spaceflight,and compared the results with those of the P300 paradigm that comes from the event-related potentials(ERPs).The results showed that the SSVEP paradigm had favorable on-orbit applicability,as the accuracies of the two astronauts reached 87.5%and 100%respectively.Moreover,the SSVEP features could reach stability earlier than those of P300 after entering orbit,which demonstrated the interaction ability of steady-state visual BCIs.On this basis,the thesis conducted the further studies on the interaction speed and the information dimension,aiming at the codec performance requirements of steadystate visual BCIs for future astronautic applications.From the aspects of promoting interaction speed,this thesis built the stimulation interface and the EEG device towards practical use,and introduced the dynamic stopping strategy to enhance the output flexibility.The dynamic recognition effectiveness was improved by optimizing the coding paradigm and decoding algorithm,which realized the maximal information transfer rate as 420.2 bits/min.This work provides technology support for future high-speed BCI system development towards interaction in space.From the aspects of expanding information dimension,this thesis proposed a spatial codec method by eliciting multifocal steady-state visual evoked potentials(mfSSVEPs).A spatially filtering algorithm named inter-task-related component analysis(iTRCA)was designed and enhanced the signal-to-noise ratio for each frequency,which resulted an average accuracy of 80.9%with a peak at 95.3%.This codec method provides new ideas for resolving two-dimensional visual evoked EEG information so as to decode the visual image of astronauts and expand the instruction set.This thesis verified the validity of current steady-state visual BCIs during spaceflight,formed methods about optimizing high-speed SSVEP systems towards future interaction demands,and explored the feasibility of coding and decoding a twodimensional visual BCI.The results of this thesis are promising to provide experience,technology,and data supports for developing novel BCI systems towards future astronautic interactive operation. |
