In the field of computer vision, DEtection TRansformer (DETR) has received significant recognition for its ability to streamline the design process of object detectors through the concept of set prediction. However, its exceptional performance comes at the cost of a high parameter count and significant computational requirements. Moreover, its ability to detect small objects is compromised, making it less suitable for analyzing high-altitude Unmanned Aerial Vehicle (UAV) images. This paper proposes UAV-DETR, a DETR architecture specifically designed for detecting UAV images captured at high altitudes, which achieves a trade-off between parameter count and precision. UAV-DETR is built in two steps: first, inverted residual structures are used to preserve low-dimensional image features, followed by a carefully designed cascaded linear attention mechanism to mitigate parameter redundancy. Through observation and analysis of the attention diffusion issue in the encoder, a cross-channel dynamic sampling mechanism is proposed, which effectively expands the model's receptive field while maintaining accuracy. In addition, the loss function is redesigned by incorporating the Wasserstein distance, which is insensitive to bounding boxes, in order to significantly enhance the convergence speed of the model. Extensive experimental results on two major benchmarks, i.e. VisDrone and UAVDT, validate the simplicity and efficiency of our model. Specifically, on the VisDrone2021 public test set, UAV-DETR exhibits superior performance with only 14 million parameters compared to YOLOv8$_{m}$, reducing the model's parameter count and complexity by 44% and 10% respectively, while achieving a 16.6% improvement in accuracy, without any data augmentation or post-processing procedures.

The ultraviolet differential optical absorption spectroscopy (UV-DOAS) technique is widely used for detecting the concentration of atmospheric pollutants. However, its concentration inversion model usually relies on specific measuring instrumentation such as spectrometers and light sources. When the instrumentation is replaced, differences in instrumentation characteristics, environmental conditions, and measurement parameters mean that the model trained with data from the original instrumentation cannot be directly applied to the new instrumentation. To address this issue, this paper proposes a cross-instrumentation gas concentration detection method based on domain adaptation (DA). The research objects are nitric oxide (NO) and sulfur dioxide (SO2). The goal is to achieve the reuse of source domain model on new instrumentation. This method only requires a small amount of target domain data to achieve DA through Gaussian process regression (GPR). The results show that the method performs excellently under different spectrometer and light source conditions. The coefficient of determination (R²) for NO and SO₂ both exceed 0.999. The linear error is lower than 2.00%. This method is expected to be widely applied in cross-instrumentation gas concentration detection.

The design and optimization of magnetorheological elastomer (MRE) isolators fundamentally dictate their ultimate performance and application potential. Current methods are limited by over-reliance on designer’s personal experience or inability to concurrently optimize structural configuration and performance parameters. To address these limitations, this study proposes a synergistic design method for MRE isolator, combining extended-variable tri-phase topology optimization and multi-objective parameter optimization. The tri-phase method expands design variables from [0, 1] to [−1, 1], enabling simultaneous optimization of the excitation coil, magnetic circuit and air region without increasing dimensionality, thus yielding a lightweight configuration of MRE isolator. It overcomes the drawback of traditional optimization methods that require predefined of structural configurations. Based on the topological results, a parametric model is constructed, and a multi-objective optimization method (coupling NSGA-II genetic algorithm with finite element simulation) is developed to derive the Pareto frontier. This frontier resolves design trade-offs among the ‘magnetic field–energy consumption–lightweighting’ of MRE isolator. An optimal design is selected from the feasible solution space, achieving minimal mass while maintaining an 89th-percentile magnetic energy ratio. Experimental validation confirms the design’s adaptive vibration mitigation performance and manufacturability. This work establishes a systematic and scalable methodology that bridges empirical design and data-driven optimization, applicable to MRE isolators and other electromagnetic smart devices.

Frequency-dependent stiffness and damping are critical properties of magnetorheological elastomer (MRE) isolators, significantly influencing their vibration response, modeling and control applications. Existing methods are limited to acquiring dynamic characteristics at a single-frequency per test, making it challenging to obtain continuous frequency-dependent properties of MRE isolators. This study proposes a novel method to determine the continuous frequency-dependent stiffness and damping of MRE isolator via one-time test. Based on a theoretical analysis of the MRE isolator-mass system model, the continuous frequency-dependent stiffness and damping are derived and analyzed, through a series of real and imaginary component transformations on the reciprocal of transfer function. The proposed method identifies the continuous frequency-dependent dynamic characteristics using only a single sweep test result, without additional tooling and test system. The results reveal that the dynamic stiffness of MRE isolator generally increases with frequency across various current levels, while the damping consistently decreases. Furthermore, the trough and inflection point in the frequency-dependent stiffness and damping curves effectively capture the resonance information of MRE isolation system. Beyond enhancing the practical application and performance analysis of MRE isolators, this method is also universal for the frequency-dependent analysis of other actuators, such as MR dampers, rubber isolators, etc.

Traditional magnetorheological elastomers (MREs) often suffer from limited modulus tunability and insufficient energy dissipation, which restrict their applications. This study prepared a novel composite material by an MR gel (MRG) embedded within the MRE, called the MRE encapsulating MRG, to synergistically enhance these properties. Annular and radial MRE encapsulating MRG configurations were fabricated using 3D-printed molds, and their dynamic mechanical performance was characterized under varying magnetic fields (0–1 T) via a rheometer. The results revealed that the composite materials demonstrated significantly improved magnetic-induced modulus and magnetorheological (MR) effects compared to conventional MREs. Specifically, the annular MRE encapsulating MRG exhibited a 238.47% increase in the MR effect and a 51.35% enhancement in the magnetic-induced modulus compared to traditional MREs. Correspondingly, the radial configuration showed respective improvements of 168.19% and 27.03%. Furthermore, both the annular and radial composites displayed superior energy dissipation capabilities, with loss factors 2.68 and 2.03 times greater than those of pure MREs, respectively. Dynamic response tests indicated that composite materials, particularly the annular MRE encapsulating MRG, achieve faster response times. These advancements highlight the composite’s potential for high-precision damping systems, vibration isolation, and adaptive control applications.