The thermal runaway hazards pose a serious threat to the application and transport of lithium-ion batteries on the aircraft. Hence, the researches of thermal safety in flight condition are necessary. In this study, the tests were conducted in a dynamic pressure chamber to study the effects of ambient pressure and heating power on the thermal runaway characteristics. The results show that the fierce behaviors of jet fire, deflagration, and explosion only were observed in high ambient pressure with high heating power. The open time of the safety valve is advanced as pressure from 95 kPa to 20 kPa. The parameters of heat release rate (HRR), total heat release (THR), cell surface temperature, peak concentration of CO₂, and mass loss decrease as the descend of external pressure or heating power. The peak values of hydrocarbon (CH_x) and CO increase with the descent of pressure but decrease as the reduction of heating power. The effects of ambient pressure on the thermal runaway (TR) fire behaviors mainly attribute to the low oxygen density. The time of heating and smoking may account for the difference of TR behaviors with various heating power. It is revealed that the fire risk and the hazards of toxic/flammable gas emissions are tightly relative to the TR behaviors. These results provide valuable proposals and inspiration for the safety warning and hazard reduction under low pressure.

To clarify the reasons for inaccurate fire detection in aircraft cargo holds, this article depicts research from the perspective of a single type of sensor detection. In terms of fire smoke, we select dual-wavelength photoelectric smoke sensors for fire-data collection and a genetic algorithm to optimize the classification and detection of random forest fires. From the perspective of fire CO concentration, we use PSO-LSTM to train a CO concentration compensation model to reduce sensor measurement errors. Research is then conducted from the perspective of various types of sensor detection, using the improved BP-AdaBoost algorithm to train a fire-detection model and achieve the high-precision identification of complex environments and fire situations.

The fire hazard of lithium-ion batteries (LIBs) poses a serious threat to their transportation and use. The purpose of this study is to investigate the efficiency of low-pressure twin-fluid water mist (TFWM) on suppressing lithium-ion battery fires. Experiments were executed to research the effect of working pressures and release stages on extinguishing the fire. Aqueous vermiculite dispersion (AVD), a commercial agent that was specifically designed to extinguish battery fires, was chosen to compare with the fire suppression performance of TFWM under the same conditions. The results indicate that the type 21,700 LIB fires could be controlled by applying the water mist within 10 s. The cooling ability at various working pressures (0.4, 0.8, 1.0, and 1.2 MPa) demonstrated an increase in inhibitory effects as the working pressure increased, and the optimal pressure was 1.2 MPa. The results further show that the extinguishing ability of the TFWM was better than the AVD agent. When the water mist was applied at the optimal working pressure, the surface temperature, flame temperature and concentration of CO reduced more significantly, compared with the AVD agent. Therefore, the TFWM shows considerable merit as a candidate to fight LIB fires.

Bird's-eye view (BEV) radar-camera fusion is an economical and robust solution for 3D perception in autonomous driving. However, existing methods struggle with spatial misalignment between modalities and lack the ability to dynamically adjust the contribution of each modality across different spatial regions in complex scenarios. To address these issues, we propose SDEF-BEV, a novel spatial-aware dual-expert fusion network. At its core, SDEF-BEV adopts a parallel dual-path fusion architecture. The Spatial-Aware Dual-Expert Fusion (SDEF) module adaptively learns optimal feature fusion strategies via two specialized experts (semantic-enhanced and geometry-enhanced) and a spatial gating network, generating location-wise fusion weights. Meanwhile, a parallel Channel and Spatial Fusion (CSF) path preserves rich spatial context through convolutional inductive biases. The outputs of these two paths are aggregated to form unified and information-rich fused features. Experiments on the nuScenes dataset demonstrate that SDEF-BEV achieves an NDS of 57.1% and an mAP of 45.7%, showing competitive performance. Extensive ablation studies further validate the effectiveness of the parallel SDEF architecture, and the overall approach exhibits strong robustness, particularly under challenging adverse weather conditions.

Camera-based bird's-eye-view (BEV) 3D object detection is important for autonomous-driving perception, but the quality of BEV representations is easily affected by depth-estimation errors, object motion, and cross-frame alignment errors. Existing temporal BEV detection methods usually introduce historical frames to enhance the current representation. However, historical BEV features do not always provide effective complementary value and may introduce misaligned or redundant responses under occlusion changes, dynamic object displacement, and local depth uncertainty. To address this issue, this paper proposes a Depth-Confidence-Guided Complementary Residual Temporal Adapter (DCRT). DCRT is integrated into a baseline temporal BEV detection framework as a bypass residual module. It uses depth-confidence cues to construct complementary relationships between the current and historical frames, and generates historical residual candidates through dynamic local sampling, enabling selective compensation using historical information that provides effective complementary value. In addition, DCRT introduces a Null-aware residual arbitration mechanism to suppress low-quality residual injection when historical information lacks effective complementary value. Experiments on the nuScenes and Waymo validation sets show that DCRT consistently improves the BEVDepth + T temporal baseline under their respective evaluation protocols. Ablation studies, complexity analysis, and visualization results further validate the effectiveness of the proposed method.