Comparison of Performance of Virtual Coupling Schemes for Haptic Collaboration using Real and Emulated Internet Connections

Abstract

Networked haptic virtual environments (NHVEs) are those in which multiple users collaborate and experience force feedback at the same time. The robustness of such systems needs to be tested under various network conditions that closely mirror the Internet. Previously, we had proposed three virtual coupling schemes to maintain position coherency in a NHVE, which were tested using constant and then time-varying delays using the actual Internet through UDP packet reflectors. In this paper we present the results of comparing performance of the virtual coupling schemes for a time varying delay emulated using the popular network emulator NIST Net, with delay conditions that existed during our real Internet experiment to Italy. UDP was used for haptic data communication because of the high transmission rate requirements for NHVEs. Experiments were conducted for three fixed packet transmission rates of 1000, 500 and 100 Hz, and their performance compared using an independent-samples t-test to the data obtained using the Internet. Locally, the haptic update rate was maintained at 1000 Hz during the experiments. Our results show that the NIST Net was a suitable emulator for testing with lower packet transmission rates. At the transmission rate of 1000 Hz the performance of the virtual coupling schemes were significantly different from that of the actual Internet experiment.

Measurement and Emulation of Time Varying Packet Delay with Applications to Network Haptic Virtual Environments

“…While it was lower for a packet transmission rate of 250 Hz, at this low packet transmission rate the networked haptic application may not perform well or may be unstable. We have further extended this work by testing a NHVE using NIST Net and compared its performance to that of using the real Internet [14].…”

Evaluation of objective speech transmission quality measurements in packet-based networks

“…We proved by several experiments [8,9] that NISTNet suits these requirements satisfactorily. It was also used in other experiments [12]. The measured samples were adjusted in Adobe Audition 3.0 (converting stereo → mono) and then tested using the POLQA (ITU-T P.863) PESQ (ITU-T P.862) and 3SQM (ITU-T P.563) algorithms [7].…”

Impact of jitter and jitter buffer on the final quality of the transferred voice

“…The results depend on accuracy and repeatability of the network simulation. We proved by several experiments that NISTNet suits to these requirements satisfactorily and is also used in other laboratories [8], [9]. The measured samples were adjusted in Adobe Audition 3.0 (converting stereo -> mono) and then tested using the algorithms POLQA (ITU-T P.863) and PESQ (ITU-T P.862).…”