The Research Of Neural Network Control System For Cervical Traction Device

Traction therapy is an important method for treatment of cervical disease,according to the characteristics of cervical physiological and pathological, give thecervical period loading and load shedding can get better therapeutic effect in theprocess of traction. We should precise control of the traction in order to ensure thesafety and efficacy of traction therapy. Due to the larger scope of traction, load is aperiodic piecewise functions, the patient have muscle resistance and other reasons inthe process of traction, experiments show that the traditional PID control method isdifficult to meet the accuracy requirements, Therefore, this thesis studies a cervicaltraction control system based on fuzzy expert and neural network. Complete thefollowing tasks:1In order to eliminating the systematic errors of PID neural network controlalgorithm, we have introduced the variables C of systematic deviation, greatlyimprove the control precision, the PID neural network algorithm has been improved,so that it can meet control requirements of the equipment. Found the linearrelationship between C and traction on the study of variation of C.2The greater change rate on the piecewise point lead to the error has beenincrease, due to the load is a periodic piecewise functions. In order to improve thecontrol precision, we have used the control strategy combine expert system and PIDneural network, measured the weight of W and C under various operating conditionsget a fuzzy expert system.3Development the traction control system based on the algorithm of theimproved PID neural network and fuzzy expert system, using Visual C#2008assoftware development tool and Windows XP as the operation platform. The systemachieve the expected targets， realizes the traction control and patient informationmanagement.