Research On Control Methods For Rehabilitation Robot Based On Patient’s Pain

Annually, approximately 15 million people in the world h ave permanent paralysis due to stroke and other cardi ovascular diseases. It bring s great incon venience to the patients’ daily life; also to the f amily and society bring heavy spirit and econom y burden. A growing number of stroke patients need to recuperate to regain limb motor function. Currently, many rehabilitation robots are developed by dom estic and foreign researchers to assis t patients for rehabi litation training. Rehabilitation robots are expected to solve the manual training problems of rehabilitation physician and alleviate physician resources tension. But m ost of robots can only provide m echanical assisted exercise. It can not resh ape the damaged neur al pathways effectively. It also lacks the active movement intentions of patients and it’s difficult to arouse the enthusia sm of patients in rehabilitation traini ng. In addition, the l ack of pain m onitoring in training process can cause secondary damage to patients. To solve these problems, this paper studied on control methods of rehabilitation robot based patients’ pain. The m ovement intentions were identified and the pain intensity was quantif ied throngh detecting various biological feedback si gnals. The a im is to esta blish an ef fective security rehabilitation system for active rehabilitation training.Firstly, this paper brief ly described the domestic and foreign researches of pain assessment, brain-computer interface(BCI), functional electrical stimulation(FES) in the field of rehabilitation robot. And, the main contents of this article were proposed.Secondly, we carried out the study of pain intensity recognition based on multiple physiological signals. In the vi ew of the original features contains a large am ount of irrelevant or redundant features for decreas ing the pain in tensity recognition rate, a feature selection procedure based on gene tic algorithm was designed to find the combination of features asso ciated with pain. The optim al pain intensity recognition models were obtained.Again, the identification method of patients’ active movement intentions based on EEG signal was studied. The appropriate vi sual stimulation paradigm with 12 HZ 、15HZ、20HZ frequency was set up to evoke the EEG signal. The relevant steady-state visual evoked potential signals w ere extracted to estab lish an effective BCI for intentions recognition.Then, the modern control m ethod was introduced in func tional electrical stimulation for upper extrem ity rehabilitation. An iterative learni ng control algorithm with PD feedback was designed to obtain the optimized pulse width sequence of FES for completing different recovery trajectory tracking, in turn to reshape neural pathways and regain motor function.Finally, a preliminary experiment of brain-controlled rehabilitation system based on the feedback of pain was implemented. The BCI was used as the upper rehabilitation control instruction; the integrated rehabilitation system of robot with FES conducted the specific rehabilitation training. At the same time, the pain infor mation was used as a feedback parameter to adjust the rehabilitation strategy.