Recurrent neural networks (RNNs) are powerful in the tasks oriented to sequential data, such as natural language processing and video recognition. However, because the modern RNNs have complex topologies and expensive space/computation complexity, compressing them becomes a hot and promising topic in recent years. Among plenty of compression methods, tensor decomposition, e.g., tensor train (TT), block term (BT), tensor ring (TR), and hierarchical Tucker (HT), appears to be the most amazing approach because a very high compression ratio might be obtained. Nevertheless, none of these tensor decomposition formats can provide both space and computation efficiency. In this article, we consider to compress RNNs based on a novel Kronecker CANDECOMP/PARAFAC (KCP) decomposition, which is derived from Kronecker tensor (KT) decomposition, by proposing two fast algorithms of multiplication between the input and the tensor-decomposed weight. According to our experiments based on UCF11, Youtube Celebrities Face, UCF50, TIMIT, TED-LIUM, and Spiking Heidelberg digits datasets, it can be verified that the proposed KCP-RNNs have a comparable performance of accuracy with those in other tensor-decomposed formats, and even 2,78,219x compression ratio could be obtained by the low-rank KCP. More importantly, KCP-RNNs are efficient in both space and computation complexity compared with other tensor-decomposed ones. Besides, we find KCP has the best potential of parallel computing to accelerate the calculations in neural networks.

It is essential and forward-thinking to investigate the personalized use of four-wheel driving and steering wire-controlled unmanned chassis. This paper introduces a personalized path-tracking approach designed to adapt the vehicle’s control system to human-like characteristics, enhancing the fit and maximizing the potential of the chassis’ multi-directional driving and steering capabilities. By modifying the classic vehicle motion controller design, this approach aligns with individual driving habits, significantly improving upon traditional path-tracking control methods that rely solely on reference vector fields. First, the classic reference vector field’s logic was expanded upon, and it is shown that a personalized upgrade is feasible. Then, driving behavior data from multiple drivers were collected using a driving simulator. The fuzzy c-means clustering method was used to categorize drivers based on typical states that match vehicle path-tracking performance. Additionally, the random forest algorithm was used as the method for recognizing driving style. Subsequently, a personalized path-tracking control strategy based on the reference vector field was developed and a distributed execution architecture for four-wheel driving and steering wire-controlled unmanned chassis was established. Finally, the proposed personalized path-tracking approach was validated using a driving simulator. The results of the experimental tests demonstrated that the personalized path-tracking control approach not only fits well with various driving styles but also delivers high accuracy in driving style identification, making it highly suitable for application in four-wheel driving and steering wire-controlled chassis.

4D millimeter-wave radar provides high-precision ranging capability and exhibits strong robustness under adverse weather and low-visibility conditions, but its point clouds are relatively sparse and suffer from severe elevation-angle measurement noise. Monocular cameras, by contrast, provide rich semantic information and high recall, yet are fundamentally limited by scale ambiguity. To exploit the complementary characteristics of these two sensors, this paper proposes a radar-camera fusion 3D multi-object tracking framework that does not rely on complex 3D annotated data. First, on the radar signal-processing side, a Gaussian distribution-based adaptive angle compression method and IMU-based velocity compensation are introduced to effectively suppress measurement noise, and an improved DBSCAN clustering scheme with recursive cluster splitting and historical static-box guidance is employed to generate high-quality radar detections. Second, a disparity-domain metric depth recovery method is proposed. This method uses filtered radar points as sparse metric anchors, performs robust fitting with RANSAC, and applies Kalman filtering for temporal smoothing, thereby converting the relative depth output of the visual foundation model Depth Anything V2 into metric depth. Finally, a hierarchical fusion strategy is designed at both the detection and tracking levels to achieve stable cross-modal state association. Experimental results on a self-collected dataset show that the proposed method achieves an overall MOTA of 77.93%, outperforming single-modality baselines and other comparison methods by 11 to 31 percentage points. This study provides an effective solution for low-cost and robust environment perception in complex dynamic scenarios.

Infrared remote sensing is indispensable for aerial target detection due to its all-weather thermal radiation sensing capability. Integrating visible light’s high resolution with infrared’s robustness to environmental interference represents an advanced strategy to improve segmentation accuracy. Current challenges of the strategy include: 1) degradation of infrared images due to background noise; 2) cross-modal feature mismatches caused by temperature disparities between thermal and spectral data; 3) insufficient annotated datasets restricting model generalization. To address these challenges, this paper proposes a temperature-aware dynamic fusion network for few-shot infrared image segmentation algorithm. The method significantly improves segmentation accuracy through two key innovations: temperature-guided dual-modal feature fusion and thermodynamic few-shot transfer mechanisms. Specifically, in temperature-guided dual-modal feature fusion, Dual-stream Pyramid Enhancement Module (DPEM) is designed to suppress noise and boost contrast in infrared image. Building upon this, Temperature-guided Dual-modal Encoder is constructed to dynamically fuse visible light and Infrared features. In thermodynamic few-shot transfer, Thermodynamic Adversarial Alignment can obtain dual-modal features with discriminative information in few-shot transfer. Then, Thermodynamic Pseudo-Labeling Enhancement is created to improve pseudo-label quality by Temperature Confidence Propagation. Numerous experimental results performed on a multi-scenario dataset with 12,800 visible-infrared aircraft images have validated the method’s superiority, achieving SOTA performance with minimum gains of 7.07% mPA, 6.48% mF1, and 7.93% mIoU over existing methods like GANFuse and TransFusion. This framework advances explainable infrared visual understanding in complex environments.