A Research Of Hemiplegia Rehabilitation Monitoring System Based On Inertial Sensors

Hemiplegia is a common cerebrovascular sequela. Hemiplegia patients suffer from body movement disorders, and at the same time, they are often accompanied by a slight feeling, language, cognitive disorders. Nowadays, the current hemiplegia rehabilitation training is mainly assisted by physicians or by the mechanistic rehabilitation equipment, but ignores the training process all by theirselves in spare time. With the gradually mature of MEMS sensor technology, the inertial capture technology is more and more used in people’s daily life. This paper, combining the independent recovery training process of hemiplegia patients and the inertial sensor technology, puts forward a new method of tracking and evaluation of the hemiplegia movement recovery process based on inertial sensor, that is mainly used in attitude tracking and athletic ability evaluation of the whole process that hemiplegia patients do recovery movement by theirselves. The specific design solution is shown as follow:First of all, research the domestic and foreign development present situation and background significance based on the subject of this paper and determine the feasibility of this research topic. Then, design the structure and framework of the research topic overall, and built the inertial measurement unit(IMU) based on MSP430. Combining the theory of hemiplegia patients independent rehabilitation training with the concrete practice of the inertial navigation, establish the platform experiment system.Secondly, this paper selects many sets of hemiplegia recovery training actions, and monitors the hemiplegia patients trained by these actions. We build the corresponding motion model based on different actions, put forward the corresponding solving methods, and finish its’ integrity evaluation. In the process of data fusion, combine the traditional Kalman filter with the concrete movement model, raise a kind of strapdown attitude algorithm based on plane constraints. This methond optimize the data fusion algorithm and improves the precision.Thirdly, according to the traditional sensor error model, this paper introduces artificial fish swarm algorithm(AFSA) based on the traditional static correction algorithm. This method avoids the complicated calculation of mathematical model, simplify the whole sampling process, and is more suitable for embedded systems.Last of all, design the corresponding experiment of the proposed algorithm. We make an attitude calculation and contrastive analysis respectively by the original kalman filter and the Plane-Constraint Strapdown Algorithm, and determine the effectiveness of the proposed algorithm. Then, simulate the moving track of the shoulder joint and elbow, calculate the evaluation angles, and provide the action integrity evaluation.