Workload Assessment using Non-Contact Sensors (AHRTLAB): SoSE2022 paper accepted
Hybrid deep learning architecture has been developed for comparative analysis between data from the Bio-Harness as a contact sensor and data from the VPG as a non-contact sensor.
Mobile Manipulator Robot using SLAM
The goal of the project was to autonomously navigate to the goal point based on ORBSLAM2 and perform pick and place of an object using the 5 DOF arm robot placed on top the differential drive robot.
Shape Detection and Path Planning of Mobile Robot
The goal of project is using the webcam to capture a closed loop shape and make the Amigobot to navigate in that shape .
Screen Brightness Control using Electrooculography signal
Classification of the eye movements using the EOG signal. With help of denoising filter and feature extraction, the characteristics of the signals are obtained. SVM and k-NN classification methods are used for classification of the signals. Cursor movement is done using the output of the classifier which can be used to development applications
Music Genre Classification using CNN
A comparative analysis between SVM and CNN is done on the GTZAN data.
Travelling Salesmen Problem using Hopfield Network
Hopfield Network is used for solving TSP for both small scale data (10 nodes) and att48 data. It is observed that the network take considerably larger time for att48