One of the key technologies for autonomous driving is localization. Precise localization is inseparable from high-precision maps. Simultaneous Localization And Mapping (SLAM) is technologies for building maps and locating in unknown environments. One of the mainstream directions of current autonomous driving research is how to carry out efficient closed-loop processing over long periods of time, eliminating accumulated error in the odometer. Improving the accuracy of the map and ensuring the accuracy of localization is crucial for ensuring the safety of autonomous driving. Therefore, this paper proposes a featherweight and high-precision method for multi-sensor fusion of laser and GPS for loop detection. The closed-loop results are optimized together with IMU to improve the mapping accuracy.
This paper proposes a teacher-student network for industrial anomaly detection tasks using knowledge distillation techniques, where the feature representation capabilities of the teacher network model are utilized to guide the student model in efficiently identifying anomalous images. In detail, the VGG model is introduced as the backbone network for anomaly detection. Meanwhile, a residual attention mechanism is proposed to improve the feature output capability of the student network. The student network is able to learn good feature extraction abilities due to the robust feature representation of the teacher network. At the same time, the residual attention mechanism enhances the student network's ability to characterize the output of normal data features, amplifying the feature differences between the student and teacher networks for abnormal data. In summary, this paper addresses the problem of difficult detection of small targets in anomaly detection. Experimental validation has been performed on the industrial anomaly detection dataset MVTec-AD, which ultimately achieved 96.70% and 97.94% results for detecting and localizing, respectively. Compared with the other anomaly detection methods, the method proposed in this paper achieves a significant performance improvement.
In this paper, a novel adaptive terminal sliding-mode control (ATSMC) method is developed for road vehicles with uncertain dynamics. It is shown that the designed adaptive laws can recursively update the controller parameters based on the estimation values of the unknown parameter bound in the Lyapunov sense, and then the lateral position error can be driven to zero in a finite time. In comparison to the conventional sliding-mode-based vehicle lateral control method, the proposed ATSMC not only ensures fast error convergence, but also eliminates the requirement for known bounded information in controller design. Simulation results are presented to demonstrate the good performance of the proposed ATSMC scheme.
With the development of deep learning, object detection technology has been widely used in various areas. However, the large size of the model and the high computation cost limit the deployment on the devices of the robot. To address these problems, this study proposes a slight cross-stage object detection model YOLO-OR for the dish recycling robot. In detail, the cross-stage octave convolution combines with the residual structure to build a novel convolution block ORBlock, improving feature extraction capability through cross-stage convolution operations. Moreover, ORBlock significantly reduces the parameters via splitting the traditional convolution into two-dimension. In particular, the four outputs of the backbone’s last two ORBlocks are merged and transmitted to the YOLO neck for feature fusion. Based on the Dish-20 dataset, YOLO-OR achieves excellent performance with 99.56% mAP, 189.42 FPS, 4.64 G FLOPs and 3.29 MB parameters. Meanwhile, the performance of YOLO-OR outperforms other state-of-the-art YOLO models. These demonstrate the effectiveness of the proposal.
In recent years, there has been an increase in the number of patients with motor dysfunction due to neurological disorders such as stroke and spinal cord injury. Existing rehabilitation medical resources are unable to provide one-on-one rehabilitation treatment for patients. Lower limb rehabilitation robots (LLRRs) are used as assistants to rehabilitation physicians to help patients with rehabilitation training by simulating the walking gait of a normal person. Each patient’s individual degree of impairment varies from person to person, and the parameters of the lower extremity differ, which leads to different rehabilitation training programs used by the patient. Therefore, how to tailor the reference trajectory for patients according to their physical and gait parameters becomes the focus of research at this stage. In this paper, a gait generation model based on GLS network is designed to provide continuous personalized gait trajectories for patients. The GLS network is composed of an LSTM (Long Short-Term Memory) network based on the gap loss function and a sparrow search algorithm (SSA).The gap loss is presented based on the difference between the initial angle and the end joint angle of each cycle. The experimental results are analyzed by mean absolute deviation (MAD). The experimental results show that the gait trajectory generated by the GLS network reduces the experimental error by 16.9 % compared to the trajectory generated by the conventional algorithm (PSO). Therefore, our proposed GLS network-based generation of personalized continuous gait trajectories generates gait trajectories that are more similar to the actual gait trajectories.
