This letter presents Switch-SLAM, switching-based LiDAR-inertial-visual SLAM for degenerate environments, designed to tackle the challenges in degenerate environments for LiDAR and visual SLAM. Switch-SLAM achieves high robustness and accuracy by utilizing a switching structure that transitions from LiDAR to visual odometry when degeneration of LiDAR odometry is detected. To efficiently detect degeneration, Switch-SLAM incorporates a non-heuristic degeneracy detection method that does not require heuristic tuning and demonstrates generalizability across various environments. Switch-SLAM is evaluated on diverse datasets containing both LiDAR and visual odometry degeneracy scenarios. The experimental results highlight the accurate and robust localization by the proposed method in multiple challenging environments with either LiDAR or visual SLAM degeneracy.

Sit-to-stand (STS) motion is an important daily activity, and many post-stroke patients have difficulty performing STS motion. Previous studies found that there are four muscle synergies (synchronized muscle activations) in the STS motion of healthy adults. However, for post-stroke patients, it is unclear whether muscle synergies change and which features primarily reflect motor impairment. Here, we use a machine learning method to demonstrate that temporal features in two muscle synergies that contribute to hip rising and balance maintenance motion reflect the motor impairment of post-stroke patients. Analyzing the muscle synergies of age-matched healthy elderly people (n=12) and post-stroke patients (n=33), we found that the same four muscle synergies could account for the muscle activity of post-stroke patients. Also, we were able to distinguish post-stroke patients from healthy people on the basis of the temporal features of these muscle synergies. Furthermore, these temporal features were found to correlate with motor impairment of post-stroke patients. We conclude that post-stroke patients can still utilize the same number of muscle synergies as healthy people, but the temporal structure of muscle synergies changes as a result of motor impairment. This could lead to a new rehabilitation strategy for post-stroke patients that focuses on activation timing of muscle synergies.

A novel method to estimate the missing dimension in 2D acoustic images for 3D reconstruction is proposed in this paper. Acoustic cameras can acquire high resolution 2D images in underwater environment insusceptible to water turbidity and light condition. However, the formulation of acoustic images leads to the missing dimension problem. Estimating the unknown elevation angle dimension is a difficult task which has recently drawn the attention of researchers. The non-bijective characteristic between 3D points and 2D pixels increases the complexity of the problem. In this paper, a novel elevation angle estimation method is proposed. The method transfers the acoustic view to pseudo front view using a deep neural network. The proposed network can estimate the missing dimension and resolve the non-bijection problem of the 2D-3D correspondence. Because of the difficulty of acquiring depth information in underwater environments, the network is trained using simulated images. To mitigate the sim-real gap, a neural style transfer method is implemented to generate a realistic image dataset for training. Simulation experiments were carried out for evaluation and real data proved the feasibility of the proposed method.

In this paper, we propose a novel method to generate realistic acoustic datasets for forward-looking sonars. A forward-looking sonar, which is also known as an acoustic camera, outperforms other imaging sensor when applied in underwater tasks as it can provide more accurate and detailed information about the environment, even in dark or turbid water. However, the difficulty and high cost of acquiring acoustic images in real experiments encourage researchers to consider the generation of simulated acoustic image datasets. In particular, deep learning-based methods demonstrated high performance in computer vision tasks, such as in object detection. However, a large dataset is necessary in most cases. In the proposed method, we first build a novel user-friendly acoustic image simulator based on 3D modeling software. Then, the CycleGAN is applied to generate realistic acoustic images based on the generated dataset from the simulator. The experimental results demonstrate that our method can generate effective and realistic acoustic datasets with relatively simple operations. GRAPHICAL ABSTRACT

In this article, a novel dense underwater 3-D mapping paradigm based on pose graph simultaneous localization and mapping (SLAM) using an acoustic camera mounted on a rotator is proposed. The demands of underwater tasks, such as unmanned construction using robots, are growing rapidly. In recent years, the acoustic camera, which is a state-of-the-art forward-looking imaging sonar, has been gradually applied in underwater exploration. However, distinctive imaging principles make it difficult to gain an intuitive perception of an underwater environment. In this study, an acoustic camera with a rotator was used for dense 3-D mapping of the underwater environment. The proposed method first applies a 3-D occupancy mapping framework based on the acoustic camera rotating around the acoustic axis using a rotator at a stationary position to generate 3-D local maps. Then, scan matching of adjacent local maps is implemented to calculate odometry without involving internal sensors, and an approximate dense global map is built in real time. Finally, based on a graph optimization scheme, offline refinement is performed to generate a final dense global map. Our experimental results demonstrate that our 3-D mapping framework for an acoustic camera can achieve dense 3-D mapping of underwater environments robustly and accurately.