Large Language Models (LLMs) such as ChatGPT are increasingly used by developers to generate test code, yet little is known about their real-world usage patterns, developer satisfaction, and the extent to which generated code is integrated into software projects. This paper presents an empirical study of 407 user-ChatGPT conversations related to test generation from GitHub, including 152 that contain both test and corresponding source code. We systematically categorize the primary use cases of ChatGPT in test generation, employ sentiment analysis to assess developer satisfaction across different testing scenarios, and analyze the adoption between ChatGPT-generated test code and existing project code using PMD’s Copy-Paste Detector (CPD) tool. Our analysis identifies six ChatGPT usage scenarios in test generation: test case creation, framework demonstration, test enhancement, debugging, CI/CD integration, and mocking solution. Among these, test case creation and framework demonstration emerge as the most frequent. Sentiment analysis reveals that, except for framework demonstration, all other scenarios exhibit a higher proportion of negative feedback than positive. Moreover, adoption analysis indicates that only 25.7% of conversations show evidence of code reuse, suggesting developers are generally cautious when integrating LLM-generated test code. Based on these findings, we propose several practical implications for enhancing the effectiveness of LLM-assisted test generation.

Image matching is a fundamental task in computer vision, underpinning applications such as visual localization and structure-from-motion. While deep convolutional neural network (CNN)-based approaches have achieved high detection accuracy, their high computational cost limits their deployment on resource-constrained platforms such as mobile and embedded systems. This paper presents a lightweight image matching network that achieves a favorable trade-off between accuracy and efficiency. The proposed model further enhances robustness to large image rotations, a common challenge in aerial and robotics applications. Extensive experiments demonstrate that our method maintains competitive accuracy while significantly reducing inference time compared to existing CNN-based approaches.

Flexible joints are the integral drive and control components for manipulators in interaction applications. Ensuring optimal performance while implementing torque control is crucial to maintaining stability during interactions. However, the system is susceptible to numerous disturbances, such as motor inaccuracies, nonlinearity caused by friction, and hysteresis. These disturbances are distributed across the motor-side, reduction gear, and load-side, leading to limitations in system performance, including response speed and steady-state accuracy. This paper proposes a disturbance compensation method combined with dual torque sensors and an improved disturbance observer (DOB). Two torque sensors are installed on the fixed side and load-side, respectively. These sensors are utilized to measure the load-side disturbances without considering accuracy models. Furthermore, an improved DOB is designed based on one of the torque sensors to estimate the motor-side disturbances. The disturbance compensator and feedback controller are introduced as the control algorithms for the flexible joint to achieve high-precision torque control. The robust stability of the proposed method is analyzed. Finally, several comparative experiments are conducted under various conditions. The results demonstrate that the proposed method enhances torque control performances and backdrivability compared with the traditional methods. Note to Practitioners—Generally, lots of nonlinear disturbances in the flexible joint system are distributed across low-speed/high-torque and high-speed/low-torque ports. These disturbances will lead to limitations in torque control performance. To address the problem above, this paper focuses on integrating dual torque sensors with an improved DOB. Specifically, two torque sensors are installed on the fixed side of the reduction gear and load-side, respectively. This enables the real-time measurement and calculation of load-side disturbances. On the other hand, considering the torque sensor mounted at the fixed side and motor-side model, an improved DOB is designed to estimate the motor-side disturbances. Subsequently, the closed-loop control architecture is designed based on the torque sensor attached to the load. The complex modeling process can be reduced by installing and applying torque sensors. The model and parameter errors are not considered, and the torque control performance is improved. The research outcome of this paper provides an effective approach that can be used to achieve stable and rapid torque tracking, as well as effortless human-robot interaction. The proposed method still has good torque tracking performance in the case of a large load.

High-resolution surface soil moisture is essential for advancing various applications. The increased synthetic aperture radar (SAR) missions over the past decade present an opportunity to obtain large-scale, high-resolution soil moisture data. Machine learning methods are increasingly used for this purpose, but they generally suffered from the availability of ground-based observations. The real performance in view of a global product is still unclear. Consequently, commonly used machine learning methods were evaluated in this study in simulated global mapping scenarios with few training data, using a global dataset of 209 318 samples from 1021 locations worldwide, and a unique regional dataset with intensive ground and airborne-derived soil moisture from L-band passive microwave observations. Three evaluation scenarios based on the global dataset were involved, with ≤5% samples used for training. The target accuracy of 0.06 m3/m3 was only met in the dependent evaluation scenario, where the training and testing samples were randomly split. In the temporal evaluation scenario and spatial evaluation scenario, where training and testing samples came from different time periods or locations, the best models achieved median root-mean-square errors (RMSEs) of only 0.078 and 0.089 m3/m3, respectively. The evaluation on the regional dataset showed consistently worse accuracy statistics (RMSE > 0.1 m3/m3 and R < 0.41). Moreover, all methods failed to capture the spatial patterns of soil moisture, compared to airborne-derived passive soil moisture maps. These findings, therefore, suggest that current expectations for SAR-based soil moisture estimation using machine learning may be overoptimistic, requiring more robust approaches for scenarios with sparse ground measurements.

This paper proposes an improved stability observer to obtain the interaction state in real-time for physical human-robot-environment interaction (pHREI). Firstly, a stability observer is designed to effectively avoid the misjudgment caused by the phase advance or delay of the filter. Then, the variable admittance control algorithm is designed according to the value of the observer. Finally, a comparative experiment is carried out. The results show that the proposed method can effectively reduce the oscillation when the operator drags the robot into contact with the environment.