Detection of standing and sitting variations based on in-socket piezoelectric sensors for transfemoral amputees / Tawfik Yahya Mohammed Alnusairi

Tawfik Yahya, Alnusairi (2020) Detection of standing and sitting variations based on in-socket piezoelectric sensors for transfemoral amputees / Tawfik Yahya Mohammed Alnusairi. Masters thesis, University of Malaya.

[img] PDF (The Candidate’s Agreement)
Restricted to Repository staff only

Download (700Kb)
    PDF (Thesis M.A)
    Download (10Mb) | Preview


      A transfemoral prosthesis is required to assist amputees to perform activities of daily living (ADL). The purely mechanical or passive prosthesis has some drawbacks such as consumption of high metabolic energy and limitations in mimicking normal dynamics and kinematics of gait pattern. In contrast, the active prosthesis offers better performance and consumes less metabolic energy. However, recent active prosthesis uses surface electromyography as its sensory system which requires massive preparation work, causes sores to the patient by its electrodes and requires a lot of computation to extract meaningful features. This thesis focuses on developing signal conditioning circuitry to classify six different activities related to sit-to-stand of a transfemoral amputee using piezoelectric sensors as an in-socket sensory system. Also, it determines the optimal classifier by evaluating fifteen time-domain and frequency-domain features and selecting effective feature sets, and it investigates the effects of window size on the classification accuracy. Fifteen piezoelectric film sensors were embedded in the inner socket wall adjacent to the most active regions of the agonist and antagonist knee extensor and flexor muscles, an i.e. region with the highest level of muscle contractions of the quadriceps and hamstring. A male transfemoral amputee wore the instrumented socket and was instructed to perform several sitting and standing variations using an armless chair. Data were collected from the fifteen sensors and went through signal conditioning circuits. The overlapped technique was used to segment the data using different window lengths. Fifteen time-domain and frequency-domain features were extracted and new feature sets were obtained based on the feature performance. Eight of the common pattern recognition iv multiclass classifiers were evaluated and compared. Regression analysis was used to investigate the impact of the number of features and the window lengths on classifiers’ accuracies. The classification accuracy was first calculated using k-fold cross-validation method, and 20% of data set was held out for testing the optimal classifier. It was shown that the integration of the developed signal conditioning circuitry, the experimental protocol, and the data collection method could generate a consistent and distinguish signal pattern for each sit-to-stand and stand-to-sit related activity. The results showed that 2-feature set consisting of the root mean square (RMS) and the number of peaks achieved the highest classification accuracy with most of the classifiers. Also, it showed that varying a segment length from 150 ms to 600 ms had no significant effects on support vector machine (SVM) classifiers using the 2-feature set. SVM with cubic kernel was suggested to be the optimal classifier, and a classification accuracy of 98.33 % was achieved using the test data set. In conclusion, this work demonstrates the use of in-socket piezoelectric sensors to classify activities of a transfemoral amputee using pattern recognition. Different variations of sitting and standing activities were accurately classified using two timedomain features and SVM with cubic kernel.

      Item Type: Thesis (Masters)
      Additional Information: Dissertation (M.A.) - Faculty of Engineering, University of Malaya, 2020.
      Uncontrolled Keywords: Transfemoral amputee; Transfemoral prosthesis; Piezoelectric sensor; Insocket sensors; Classification
      Subjects: T Technology > TA Engineering (General). Civil engineering (General)
      Divisions: Faculty of Engineering
      Depositing User: Mrs Rafidah Abu Othman
      Date Deposited: 09 Mar 2021 01:42
      Last Modified: 09 Mar 2021 01:42

      Actions (For repository staff only : Login required)

      View Item