This paper proposes a multimodal fusion 3D target detection algorithm based on the attention mechanism to improve the performance of 3D target detection. The algorithm utilizes point cloud data and information from the camera. For image feature extraction, the ResNet50 + FPN architecture extracts features at four levels. Point cloud feature extraction employs the voxel method and FCN to extract point and voxel features. The fusion of image and point cloud features is achieved through regional point fusion and voxel fusion methods. After information fusion, the Coordinate and SimAM attention mechanisms extract fusion features at a deep level. The algorithm’s performance is evaluated using the DAIR-V2X dataset. The results show that compared to the Part-A2 algorithm; the proposed algorithm improves the mAP value by 7.9% in the BEV view and 7.8% in the 3D view at IOU = 0.5 (cars) and IOU = 0.25 (pedestrians and cyclists). At IOU = 0.7 (cars) and IOU = 0.5 (pedestrians and cyclists), the mAP value of the SECOND algorithm is improved by 5.4% in the BEV view and 4.3% in the 3D view, compared to other comparison algorithms.

To improve the efficiency of the electric vehicle (EV) drive systems and EV performance, the use of multi-speed transmissions and distributed drives has been studied extensively. In addition, to develop efficient and compact drive systems, new clutch solutions are needed. In this paper, we propose an in-wheel two-speed automatic mechanical transmission (IW-AMT) with a selectable one-way clutch (SOWC). The IW-AMT consists of a high-speed motor and a mechanical shift actuator, and it can realize shifting without power interruption, thus effectively reducing the unsprung mass and the technical specifications of the motor. We established a virtual prototype model of the IW-AMT to show the shifting process and evaluate the quality of shifting. The simulation results of the upshifting process indicated that the vehicle torque and velocity changed smoothly, and the maximum jerk is less than 10 m/s${}^3$. Furthermore, to improve the jerk induced by the downshifting process, we analyzed the momentary state of the SOWC struts that are dropped and attempted to improve the jerk from two aspects: improving the wet multi-plate clutch (WMPC) combination curve and improving the SOWC structure. The results indicated that the downshift-induced jerk can be reduced to 13 m/s${}^3$.

This paper proposes a multimodal fusion 3D target detection algorithm based on the attention mechanism to improve the performance of 3D target detection. The algorithm utilizes point cloud data and information from camera. For image feature extraction, the ResNet50+FPN architecture extracts features at four levels. Point cloud feature extraction employs the voxel method and FCN to extract point and voxel features. The fusion of image and point cloud features is achieved through regional point fusion and regional voxel fusion methods. After information fusion, the Coordinate attention mechanism and SimAM attention mechanism extract fusion features at a deep level. The algorithm's performance is evaluated using the DAIR-V2X dataset. The results show that compared to the Part-A2 algorithm, the proposed algorithm improves the mAP value by 7.9% in BEV view and 7.8% in 3D view at IOU=0.5 (cars) and IOU=0.25 (pedestrians and cyclist). At IOU=0.7 (cars) and IOU=0.5 (pedestrians and cyclist), the mAP value of the SECOND algorithm is improved by 5.4% in the BEV view and 4.3% in the 3D view, compared to other comparison algorithms.