Presentation on theme: "Dwaipayan Biswas University of Southampton, U.K. ESS Open Day."— Presentation transcript:
Dwaipayan Biswas University of Southampton, U.K. ESS Open Day
HIGHLIGHTS OF THE RESEARCH 2 ► Classification of four elementary arm movements that constitute a significant proportion of daily activities. ► 18 healthy subjects repeating 20 trials of each movement with a tri-axial accelerometer and a rate gyroscope located proximal to the wrist. ► Determining the appropriate type of sensor and associated data processing and classification techniques that may allow long term monitoring using a body-worn system.
HIGHLIGHTS OF THE PAPER 3 ► Ten time domain features extracted on individual sensor streams, their modulus and specific fused signals. ► Three classifiers used – LDA, QDA and SVM. ► Evaluated using a ‘leave-one-subject-out’ cross validation strategy. ► The four movements are identified with sensitivities of 83-96%, using 12 features extracted from individual gyroscope data with the low-complexity LDA learning algorithm.
MOTIVATION 4 ► To develop a wireless body area network (WBAN) system that will detect the use of the impaired limb during prescribed exercises and activities of daily living (ADL) and classify the type of movements performed using minimal number of sensors placed at optimal positions on the body. ► Bottleneck ? Significant energy expenditure in the radio front-end for supporting continuous data transmission. Reduced battery life of the sensors ► Solution ? Capturing of vital data by the sensor nodes Feature extraction and relevant data processing carried out on WSN-nodes itself, Storage of vital data and features Transmitting clinically relevant parameters to the patient station at pre-set intervals. Selecting low-complexity data processing algorithms to extend the battery life since computational complexity is directly proportional to the energy consumption
Experimental Protocol 5 Movement selection ► Four tasks representative of natural arm movements performed during ADL Task A: Reach and retrieve object Task B: Lift cup to mouth and return to table Task C: Swing arm in horizontal plane through 90° and return Task D: Rotate wrist through 90° and return ► Tasks involve sufficiently different kinetics to test the sensors ► 20 trials of each task were performed in 4 groups of 5 repetitions ► Each group of trials was separated by approximately 3 minutes (avoids self learning)
Experimental Protocol 6 Inertial sensors ► Shimmer kinematic module – 9 degress of freedom (Tri-axial accelerometers and rate gyroscopes) Provides a wireless, wearable solution with minimal hindrance Provides a means of time stamping individual events ► Magnetometers can be affected by ferromagnetic materials in the home environment and hence excluded ► Data acquisition rate – 50 Hz sampling rate ► Accelerometer range : +/- 1.5 g ► Gyroscope range : +/- 500 o/sec ► Position: Wrist - likely to produce the largest sensor response.
8 Data Sources ► 6 individual sensor streams – (3 x accelerometers and 3 x gyroscopes) ► 2 modulus signals: ► 3 fused signals combining specific accelerometer and gyroscope axes MovementSignal Combination AAccX * GyroY B and CAccY * GyroZ DAccz * GyroY
Data Processing 9 Acquisition & Pre-processing ► Raw sensor data is low-pass filtered (3rd order Butterworth filter, cut- off frequency of 12 Hz) Attenuates the high frequency noise components ► Resultant data is high-pass filtered (3rd order Butterworth filter, cut- off frequency of 0.1 Hz) Attenuates the low frequency artefacts, e.g. drift
Data Processing 10 Feature extraction ► Time domain features chosen from literature review, including: Standard DeviationMaximum Peak Amplitude Root Mean SquareAbsolute Difference (x max – x min ) Information EntropyIndex of Dispersion Jerk MetricKurtosis Peak NumberSkewness
Data Processing 11 Feature normalisation ► Logistic normalisation and standardisation techniques were tested ► Linear normalisation produced the best results Where y i is the normalised value of x i
Data Processing 12 Feature Selection ► Wrapper approach – various feature combination vectors are selected to test for the minimum classification error probability ► Suboptimal searching technique – Sequential forward selection (sfs) ► Selects the m best ranked features out of n ranked features (m <= n) ► Chosen as opposed to other common methods like ReliefF algorithm and clamping technique which are computationally intensive. ► Features are selected from: Individual X, Y, Z axes (3×10 features) for accelerometer and gyroscope, Modulus signals (1×10 features) for accelerometer and gyroscope and Fused signals (3×10 features)
Data Processing 13 Classification ► Choice of Classifiers Linear Discriminant Analysis (LDA) Quadratic Discriminant Analysis (QDA) Library for Support Vector Machines (LIBSVM) ► Cross validation methodology Leave-one-subject-out ► Classification sensitivity of each class (accuracy is not applicable for multiclass classification) Computed from confusion matrix (actual versus predicted class) Diagonal elements represent correctly classified classes The sensitivity of Class i is given by:
Results 14 SignalA (%)B (%)C (%)D (%)Features Acc_mod58 51739 Acc_xyz8591849018 Gyro_mod827839807 Gyro_xyz9683 8812 Fused8174607513 SignalA (%)B (%)C (%)D (%)Features Acc_mod496154724 Acc_xyz8992789115 Gyro_mod827136857 Gyro_xyz9491958912 Fused8672547411 SignalA (%)B (%)C (%)D (%)Features Acc_mod425355705 Acc_xyz898782908 Gyro_mod907435805 Gyro_xyz9785908911 Fused757150699 Summary of sensitivities for each arm movement using the modulus signal (mod), individual sensor signals (xyz) and fused signals applied to the LDA, QDA and SVM classifiers. The number of features required in each case is also shown.
Results 15 Summary of sensitivities for each arm movement using the modulus signal (mod), individual sensor signals (xyz) and fused signals applied to the LDA, QDA and SVM classifiers. The number of features required in each case is also shown.
Conclusion 16 ► The sensitivity for each movement using the individual sensor signals for both the accelerometer and the gyroscope is better than that for the fused and the modulus signals. ► The difference in the recognitions rates between modulus and individual signals is partly due to the fact that bipolar information present in the raw data is retained with individual sensor signals, but lost with modulus signals. ► Hence, using the individual sensor signals provides the classifier an opportunity to select from a larger number of features and hence the recognition rate for the movements is reflected in the higher accuracies achieved. ► Considering LDA with individual sensor signals, the gyroscope recognises the four movements with sensitivities in the range of 83-96%
Conclusion 17 ► The accelerometer also has a similar detection rate with sensitivities in the range of 84-91% across all movements. ► However, the gyroscope uses only 12 features as compared to the 18 used by the accelerometer out of a total of 30 (3×10 features) and hence is the obvious choice with regard to a lower complexity solution. ► The gyroscope results using individual sensor signals with QDA and SVM are marginally higher than LDA, though the number of features required to successfully classify all four arm movements is similar for all three algorithms ► In view of the trade-off between the recognition rate and the complexity involved, LDA is considered to be computationally less complex and hence is the best choice classifier.
18 Thank you for your attention Any Questions ?