A Dual-Channel Framework for Blind Perceptual Quality Assessment in Bilateral Teleoperation
Zican Wang, Xiao Xu, Zhi Jin, Dong Yang, Eckehard Steinbach
AI summary
Problem
Perceptual haptic quality assessment currently relies on time-consuming subjective experiments or requires reference signals that are often unavailable in real-world teleoperation.
Approach
The method uses a dual-channel neural network to jointly analyze signal distortions and semantic interaction states, trained on a large augmented dataset with pseudo-MOS labels.
Key results
- Constructed a large-scale augmented haptic dataset with pseudo-MOS labels
- Developed a dual-channel no-reference network combining distortion analysis and semantic tokenization
- Achieved Spearman’s Rank-Order Correlation scores above 0.85 for QoE prediction
- Validated robustness across real-world teleoperation tasks with diverse haptic devices and robotic arms
Why it matters
Enables reliable, real-time QoE monitoring for teleoperation systems without costly subjective testing or reference signals, benefiting remote surgery, rescue, and space exploration.
Abstract
This paper proposes a perceptual no-reference (blind) haptic quality assessment framework for predicting the Quality of Experience (QoE) in teleoperation systems with force feedback. The proposed approach employs a deep neural network that combines semantic and distortion-based channels. The semantic network generates a semantic vector that charac- terizes the interaction between the robot and its environment. Meanwhile, the distortion network decomposes complex noise introduced by control algorithms and communication artifacts into artificial noise of known types. To train the proposed network, we also construct an augmented dataset for perceptual quality assessment in teleoperation based on the subjective experiments. The dataset augmentation and the model are validated with real-world teleoperation tasks. Our experimental results demonstrate that the performance of our No-Reference (NR) haptic quality assessment model is comparable to or surpasses that of commonly used Full-Reference (FR) methods, achieving Spearman’s Rank-Order Correlation scores above 0.85 for QoE prediction.