Recent research on universal object detection aims to introduce language in a SoTA closed-set detector and then generalize the open-set concepts by constructing large-scale (text-region) datasets for training. However, these methods face two main challenges: (i) how to efficiently use the prior information in the prompts to genericise objects and (ii) how to reduce alignment bias in the downstream tasks, both leading to sub-optimal performance in some scenarios beyond pre-training. To address these challenges, we propose a strong universal detection foundation model called CP-DETR, which is competitive in almost all scenarios, with only one pre-training weight. Specifically, we design an efficient prompt visual hybrid encoder that enhances the information interaction between prompt and visual through scale-by-scale and multi-scale fusion modules. Then, the hybrid encoder is facilitated to fully utilize the prompted information by prompt multi-label loss and auxiliary detection head. In addition to text prompts, we have designed two practical concept prompt generation methods, visual prompt and optimized prompt, to extract abstract concepts through concrete visual examples and stably reduce alignment bias in downstream tasks. With these effective designs, CP-DETR demonstrates superior universal detection performance in a broad spectrum of scenarios. For example, our Swin-T backbone model achieves 47.6 zero-shot AP on LVIS, and the Swin-L backbone model achieves 32.2 zero-shot AP on ODinW35. Furthermore, our visual prompt generation method achieves 68.4 AP on COCO val by interactive detection, and the optimized prompt achieves 73.1 fully-shot AP on ODinW13.

Integrating orthogonal time frequency space (OTFS) modulation into low-Earth-orbit (LEO) satellite communication systems can significantly enhance the robustness against severe Doppler effects caused by fast time-varying channels. However, since the high-dynamic characteristic of terrestrial-to-satellite links considerably constrains the OTFS frame duration, the inevitable fractional Doppler shifts will result in the emergence of inter-Doppler interference (IDI), thereby degrading channel estimation performance. To tackle this challenge, we develop an efficient correlation information-aided channel estimation scheme based on Zadoff-Chu (ZC) sequences in the presence of fractional Doppler shifts. The proposed scheme enables a fast coarse estimation of channel parameters by leveraging the magnitude distribution regularity of the periodical correlation results of ZC pilot under IDI, and remodels channel estimation as a non-convex constrained least squares optimization problem to obtain the fine channel parameters for further resistance of the inter-path interference. Complexity analysis and simulation results validate that our scheme can attain a high estimation accuracy while achieving substantial reductions in both the peak-to-average power ratio and computational complexity, in comparison to the state-of-the-art ones.

The long-term navigation capability of Underwater Internet of Things (UIoT) mobile nodes is essential for building the smart ocean. Currently, the terrain-aided navigation (TAN) method using a single-beam echo sounder (SBES) has shown its potential for achieving long-term underwater navigation, but the robustness of the SBES-TAN positioning results is hindered by the limited measurement information of the SBES. Incorporating terrain matching result confidence assessment into the SBES-TAN method is an effective way to enhance its application and improve the accuracy of navigation results. This article proposed a Markov process-based matching result confidence assessment (MPMRCA) method for the SBES-TAN system. To address the accumulated positioning error in the filter theory, both the positioning error and its variation are considered in the matching result confidence assessment. Additionally, a two-stage confidence assessment strategy is introduced to improve the accuracy of the assessment. Finally, an SBES-TAN method incorporating the MPMRCA is proposed to improve the robustness of navigation results. The validity of the SBES-TAN method, enhanced by the MPMRCA, is verified using experimental data from an autonomous underwater vehicle in a lake environment. The results demonstrate that the proposed method can effectively provide robust and accurate navigation results, benefiting long-term underwater missions for UIoT mobile nodes.

Text prompts are crucial for generalizing pre-trained open-set object detection models to new categories. However, current methods for text prompts are limited as they require manual feedback when generalizing to new categories, which restricts their ability to model complex scenes, often leading to incorrect detection results. To address this limitation, we propose a novel visual prompt method that learns new category knowledge from a few labeled images, which generalizes the pre-trained detection model to the new category. To allow visual prompts to represent new categories adequately, we propose a statistical-based prompt construction module that is not limited by predefined vocabulary lengths, thus allowing more vectors to be used when representing categories. We further utilize the category dictionaries in the pre-training dataset to design task-specific similarity dictionaries, which make visual prompts more discriminative. We evaluate the method on the ODinW [1] dataset and show that it outperforms existing prompt learning methods and performs more consistently in combinatorial inference.