An LSTM-based Bilateral Active Estimation Model for Robotic Teleoperation with Varying Time Delay

Network Latency in Teleoperation of Connected and Autonomous Vehicles: A Review of Trends, Challenges, and Mitigation Strategies

“…where T 1 and T 2 are periods of trigonometric terms and ω(0, 1) is normally distributed noise. Likewise, Zhou et al [69] model delays as a sine wave distribution. Bacha et al…”

“…The major drawback of these methods is that they require knowledge of system dynamics. Moreover, their performance degrades due to model errors and complicated conditions with variable time delay, large disturbances, or extensions to multilateral systems [69]. On the other hand, model-free approaches based on data-driven models have been developed.…”

“…Likewise, Zhou et al [ 69 ] model delays as a sine wave distribution. Bacha et al [ 70 ] use a random time delay based on a Gaussian distribution in the range [0.1, 1] s. Hatori and Uchimura [ 71 ] and Nagakura et al [ 72 ] vary the delay between 3.5 and 4.0 s and 3.0 and 4.0 s. Other studies that implemented time-varying delays include [ 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 , 81 , 82 , 83 , 84 ].…”

“…On the one hand, earlier researchers focused on mathematical models for accurately predicting and compensating for undelayed controller signals [ 45 ]. These methods are well known to be model-based approaches [ 69 ]. They are usually modelled according to control theorems, out of which the most extensively researched compensation methods are based on the perspective of passivity theory [ 44 , 51 , 114 ].…”

“…They ensure the stability and transparency of the bilateral teleoperation systems. A few examples include wave variable-based methods, scattering signals, time-domain passivity analysis (TDPA)-based methods [ 48 , 77 , 79 , 80 ], and model-mediated teleoperation (MMT) [ 69 , 80 , 115 ]. The major drawback of these methods is that they require knowledge of system dynamics.…”