Chin-Tien Wu

Ping-Kong Huang, Chien-Wu Lan, Chin-Tien Wu et al.

During the development of wearable exoskeletons, evaluations involving human subjects pose inherent safety risks. Therefore, systematic testing is often conducted using robots that emulate human motion. However, reproducing human movements is challenging due to differences in robot structure and actuator characteristics. This study proposes a three-stage offline control strategy that uses motion-capture data and robot-specific properties to generate control commands for accurate motion replication. First, an optimal torque trajectory is generated via a State-Dependent Riccati Equation (SDRE) controller based on the dynamic model of the bipedal system. Second, joint velocity and acceleration command sequences are synthesized through parameterized optimization under actuator constraints. Finally, a data-driven PID-LQR offline controller refines these commands by minimizing the tracking error between the desired and executed motions. Experimental validation is performed on a suspended bipedal robot platform designed for the evaluation of gravity-counteracting exoskeletons. Motion-capture data collected from squatting and walking tasks are used for system assessment. The experimental results demonstrate high tracking fidelity, with an average root mean square error (RMSE) below 3 degrees. These results verify the effectiveness of the proposed three-stage control strategy for robot-based systematic testing of exoskeletons.

Yu-Hsi Chen, Chin-Tien Wu

Optical flow is a fundamental technique for motion estimation, widely applied in video stabilization, interpolation, and object tracking. Traditional optical flow estimation methods rely on restrictive assumptions like brightness constancy and slow motion constraints. Recent deep learning-based flow estimations require extensive training on large domain-specific datasets, making them computationally demanding. Also, artificial intelligence (AI) advances have enabled deep learning models to take advantage of optical flow as an important feature for object tracking and motion analysis. Since optical flow is commonly encoded in HSV for visualization, its conversion to RGB for neural network processing is nonlinear and may introduce perceptual distortions. These transformations amplify the sensitivity to estimation errors, potentially affecting the predictive accuracy of the networks. To address these challenges that are influential to the performance of downstream network models, we propose Reynolds flow, a novel training-free flow estimation inspired by the Reynolds transport theorem, offering a principled approach to modeling complex motion dynamics. In addition to conventional HSV-based visualization of Reynolds flow, we also introduce an RGB-encoded representation of Reynolds flow designed to improve flow visualization and feature enhancement for neural networks. We evaluated the effectiveness of Reynolds flow in video-based tasks. Experimental results on three benchmarks, tiny object detection on UAVDB, infrared object detection on Anti-UAV, and pose estimation on GolfDB, demonstrate that networks trained with RGB-encoded Reynolds flow achieve SOTA performance, exhibiting improved robustness and efficiency across all tasks.

Li-Wei Chen, Wei-Chen Chiu, Chin-Tien Wu

It is well known that semantic segmentation neural networks (SSNNs) produce dense segmentation maps to resolve the objects' boundaries while restrict the prediction on down-sampled grids to alleviate the computational cost. A striking balance between the accuracy and the training cost of the SSNNs such as U-Net exists. We propose a spectral analysis to investigate the correlations among the resolution of the down sampled grid, the loss function and the accuracy of the SSNNs. By analyzing the network back-propagation process in frequency domain, we discover that the traditional loss function, cross-entropy, and the key features of CNN are mainly affected by the low-frequency components of segmentation labels. Our discoveries can be applied to SSNNs in several ways including (i) determining an efficient low resolution grid for resolving the segmentation maps (ii) pruning the networks by truncating the high frequency decoder features for saving computation costs, and (iii) using block-wise weak annotation for saving the labeling time. Experimental results shown in this paper agree with our spectral analysis for the networks such as DeepLab V3+ and Deep Aggregation Net (DAN).