In this thesis, we present a novel network transport protocol that we call the trinomial protocol for Internet-based teleoperation and e-service robotic systems. The trinomial protocol provides minimized delays and delay jitter. In the steady state of the network, its transmission rate is smooth, however, when network bandwidth varies, it adapts to the change quickly. The trinomial protocol satisfies all the constraints on transport protocols:​ it is responsive, inter-protocol fairness convergent, intra-protocol fairness convergent and efficiency convergent.

The implementation issues of the trinomial protocol are examined. In particular, the estimation of roundtrip time (RTT), which is an integrated component of the implementation of the trinomial protocol, is explored. It is revealed that the adjacent and near-adjacent observations in the delay time series are linearly dependent rather than random. It is also indicated that there are only weak or no non-linear correlations among RTT observations. Based on these findings, we present a novel algorithm based on the maximum entropy principle (MEP) for RTT estimation. Compared to traditional autoregressive moving average (ARMA) method, the MEP method is adaptive and capable of tracking RTT dynamics very well.

In the last part of the thesis, based on the trinomial protocol, we develop and implement a new modular platform for mobile e-service robotic systems. In the experiments, the users successfully guided a mobile robot remotely through a clustered laboratory environment over the Internet by using a Java-enabled web browser on an ordinary PC.