Off-campus UMass Amherst users: To download campus access theses, please use the following link to log into our proxy server with your UMass Amherst user name and password.
Non-UMass Amherst users: Please talk to your librarian about requesting this thesis through interlibrary loan.
Electrical & Computer Engineering
Master of Science (M.S.)
Year Degree Awarded
Month Degree Awarded
force feedback device, haptic, network, tele surgery, media synchronization, collaborative system
Telesurgery relies on fast and reliable data transmission between the surgeon and tele-operator side over lossy and delay constrained networks. Medical data involves audio, video, ECG and Force Feedback data. When these media streams are transmitted through best effort networks, the temporal information gets affected due to network constraints. Major network degradation is due to the Force Feedback device with rendering rate of 1 KHz, hence data is generated every millisecond. In our proposal we concentrate on improving the synchronization of force feedback device on varying networking conditions.
Force feedback data is generated by operating a source (surgical) device which controls the movement of remote device. It has a great potential in improving telemedicine facilities, when included with the support of different multimedia services. The channel imposes delay and packet loss constraints for such devices which require unique solutions, unlike audio or video media, due to its high rendering rate.
Current research supports Force Feedback in fiber optic communication, packet switched networks. However, such schemes are not feasible in supporting surgical telepresence system. While efforts are made to support force feedback media in wireless medium, few works have addressed delay synchronization and loss of data. There exists no previous work which has attempted to provide an efficient integrated solution where video and force feedback information have been supported by the same network. This thesis focuses in providing an integrated architecture that supports the force feedback data over a collaborative network and improves the data synchronization and packet loss prediction in the remote side over a varying network link. The goal will be to evaluate the support of such data types.
We have implemented a Linear Packet Predictor Algorithm which predicts the missing packet value. Data generated from the source device are sent as UDP packets. UDP transmission is unreliable and hence we use an RTP over UDP to make it reliable. Each packet will have the current position of the device and force applied. We use a Microsoft Sidewinder Force Feedback joystick. The handle of the joystick is located at the center of the base. So we record the position of the device on both positive and negative axis moving in a two dimensional space. This device provides rotational movement and hence drastic change in position occurs within milliseconds. Once the packet arrives at the receiver side, the control unit checks for the sequence number of the packet. If continuity is missing then, the control unit passes the packet to the predictor algorithm which predicts the packet else it directly updates the packet to the Virtual Time Rendering Algorithm
Another major issue is the delay jitter. On the source (server) side the intra time difference between two packets will be 1msec. But due to varying delay in the network the data packets arrive at the receiver with fluctuating intra time difference. In order to counter the delay jitter effect, we implement the Virtual Time Rendering algorithm which reads the time stamp value at which the packet was generated at the source and modifies the update time at the receiver side. In our work we do not control another device on the remote side, rather an applet which was developed using a Virtual Reality Markup Language in Matlab.
Another challenge which is imposed when other multimedia is introduced with force feedback is the intra media synchronization. Real time video is captured from the applet side and given as feedback to the server side to improve the interactivity of the application. At every instant in time, different multimedia data produce data to be updated at the remote end. Since all the information are inter dependent with other media in time, efficient intra media synchronization is required. This thesis also focuses in providing an architecture which not only supports force feedback data but have a multiplexed model which allows an efficient transmission of all surgical information in real time. Each data occupies significant part of bandwidth in the network and the effect of multiplexing might affect the synchronization scheme of the force feedback device. Our architecture supports the efficient transmission of all types of multimedia information and also maintain the synchronization of the scheme. This method is unique with its methodical approach to support different multimedia information.
Advisor(s) or Committee Chair