Fully sparse 3D detectors have recently gained significant attention due to their efficiency in long-range detection. However, sparse 3D detectors extract features only from non-empty voxels, which impairs long-range interactions and causes the center feature missing. The former weakens the feature extraction capability, while the latter hinders network optimization. To address these challenges, we introduce the Fully Sparse Hybrid Network (FSHNet). FSHNet incorporates a proposed SlotFormer block to enhance the long-range feature extraction capability of existing sparse encoders. The SlotFormer divides sparse voxels using a slot partition approach, which, compared to traditional window partition, provides a larger receptive field. Additionally, we propose a dynamic sparse label assignment strategy to deeply optimize the network by providing more high-quality positive samples. To further enhance performance, we introduce a sparse upsampling module to refine downsampled voxels, preserving fine-grained details crucial for detecting small objects. Extensive experiments on the Waymo, nuScenes, and Argoverse2 benchmarks demonstrate the effectiveness of FSHNet. The code is available at https://github.com/Say2L/FSHNet.

With the development and application of swarm drones, some researchers have tried to replicating the migration patterns of geese in drones swarm formation to extend their lifetime. However, the problem of performing appropriate position reconfiguration based on the battery energy still remains an unsolved issue. This paper proposes an efficient position reconfiguration approach that reduces the energy consumption imbalance of the swarm and prolongs the lifetime. The approach includes: (1) a two-step MIP (mixed-integer programming)-based optimization method. (2) a two-step heuristic algorithm that can run in pseudo-polynomial time and without the need for an optimization solver. The approach provides a complete position reconfiguration solution that determines (i) the number of position reconfiguration; (ii) which drones need to exchange positions in every position reconfiguration; (iii) the length of time to maintain each position before next reconfiguration. Finally, the approach is compared with other three methods in experiments which demonstrate the effectiveness of it.

In this paper, we propose a centralized multi-robot stereo SLAM method based on a direct approach, with the aim of achieving Fast, Low-bandwidth, and Semi-dense mapping for collaborative applications, namely CMD-SLAM. In CMD-SLAM, each agent independently runs stereo-DSO as a front-end visual odometry(VO) and shares information with the central server via the TCP/IP protocol. To lower the bandwidth, we have designed a new communication strategy. Additionally, we employ a LiDAR descriptor-based place recognition method in the server’s back-end. This method fully exploits the information from the 3D point cloud structure generated during the direct method odometry, achieving efficient loop closure detection. In optimization, to achieve a balance between accuracy and speed, we adopt a Sim(3) pose graph optimization (PGO) and employ some trick to reduce the time consumption. Finally, we evaluate CMD-SLAM on publicly available datasets and large-scale outdoor environments, comparing it with state-of-the-art approaches to demonstrate its effectiveness.

Visual object tracking plays an essential role in various maritime applications. However, most of the existing tracking methods belong to generative models, which only focus on the features of the object and require the target has significant visual saliency for accurate tracking. While the visual saliency is available in most of the common tracking conditions, these methods may fail when facing challenging situations. In this paper, a deep learning based tracking method is proposed to track maritime ships, namely, SiamFPN. In SiamFPN, a modified Siamese Network is combined with multi-RPNs to build a tracking pipeline. Concretely, A ResNet-50 with an FPN structure is used as the CNN of the detection subnetwork of Siamese, and a template subnetwork is parallel to the detection. In order to strengthen the discriminative ability, three RPNs are deployed to process the output of Siamese Network. Moreover, a historical impacts based proposal selection method is developed for selecting correct target areas. Finally, a dataset is collected for training and testing SiamFPN and validating our excellent performance over the other four recent SOTA trackers. Based on the experimental results, we achieved 74 % on average accuracy with real-time speed.

Visual object tracking has achieved remarkable progress in recent years and has been broadly applied in intelligent transportation systems such as autonomous vehicles and drones to monitor and analyze the behavior of specific targets. One typical tracking approach is the discriminative tracker, which branches into two main categories: the correlation filter (CF) and the convolutional neural network (CNN). However, most of the current researches consider both categories as two separate techniques and only rely on one of them. Thus, a dense cooperation between the CF and the CNN still remains less discovered and the question of how to effectively join both techniques to further boost the tracking performance is still open. To address this issue, in this paper, we propose a collaborative architecture which incorporates models constructed with both techniques and dynamically aggregates their response maps for target inference. By an alternating optimization, both models are learned on each other's errors to persistently improve the classification power of the whole tracker. For further efficiency, we present a faster solver for our utilized CF and an analytical solution for dynamic model weighting. Through experiments on standard benchmarks, we reveal the influence of key factors on the joint learning architecture and show that it outperforms the state-of-the-art approaches.

Vehicle detection from short-range aerial image faces challenges including vehicle blocking, irrelevant object interference, motion blurring, color variation etc., leading to the difficulty to achieve high detection accuracy and real-time detection speed. In this paper, benefiting from the recent development in MobileNet family network engineering, we propose a compressed MobileNet which is not only internally resistant to the above listed challenges but also gains the best detection accuracy/speed tradeoff when comparing with the original MobileNet. In a nutshell, we reduce the bottleneck architecture number during the feature map downsampling stage but add more bottlenecks during the feature map plateau stage, neither extra FLOPs nor parameters are thus involved but reduced inference time and better accuracy are expected. We conduct experiment on our collected 5-k short-range aerial images, containing six vehicle categories: truck, car, bus, bicycle, motorcycle, crowded bicycles and crowded motorcycles. Our proposed compressed MobileNet achieves 110 FPS (GPU), 31 FPS (CPU) and 15 FPS (mobile phone), 1.2 times faster and 2% more accurate (mAP) than the original MobileNet.

Detection of the navigable regions for the unmanned surface vehicles (USVs) sailing on the narrow rivers is very important. Existing detection methods mostly depend on the cameras, which is sensitive to environments and cannot provide reliable navigable regions for sailing. In this paper, we propose a scheme to process 3D LiDAR data to achieve an accurate and robust navigable regions detection. We conduct field experiments in a narrow river in different scenarios to prove the performance of the proposed scheme, which reaches on average 93.8% precision and 92.7% recall.

Urban driving requires the autonomous vehicles to drive with smooth control and track the planned path accurately. However, most of the existing path-tracking controllers pay more attention to the tracking errors than the smoothness because of the difficulties to balance them. This paper proposes a learning-based method to achieve the trade-off between the smooth control and the tracking-error control. An Reinforcement Learning algorithm, which is called Proximal Policy Optimization, is used to train a neural model to tune the weights of a designed controller PP_PID (Pure-Pursuit_Proportional Integral Derivative). The successfully trained model will adaptively select the optimal weights for the Pure-Pursuit and PID to guarantee the control smoothness and accuracy. Finally, the proposed controller will be tested in two path tracking scenarios. The results show that the proposed controller can change the weights adaptively to maintain a balance in the tracking error and lateral acceleration under the 35km/h.

Geometric path tracking controllers are very popular to be implemented in the autonomous vehicles because of its simple implementation and validation in the practical applications. However, it also suffered from the sensitivity to the selections of the look-ahead distance. Therefore, reasonable tuning strategy of the look-ahead distance is very important to the application of the geometric controllers. This paper proposed a more robust tuning strategy for all the geometric controllers based on the careful analysis of the previous methods. Instead of utilizing as much as possible elements, we just want to make use of only two variables, the distance from the current location to the reference path and the changing rate of it. Fuzzy logic was utilized to combine these two variables to tune the look-ahead distance. We will show that we just use two variables, but we actually consider another two variables' affections indirectly. Lastly, we will add our strategy to two different classical geometric controllers and compare the relative benefits of previous tuning strategies.