With the development of technology, unmanned aerial vehicles (UAVs) are playing an increasingly important role in the inspection of high-voltage power transmission line. The traditional inspection method relies on the operator to manually control the drone for inspection. Although many companies are using real-time dynamic carrier phase differencing technology to achieve high-precision positioning of UAVs, when UAVs fly autonomously at high altitudes to photograph specific objects, the objects tend to deviate from the center of the picture. To address this error, in this paper, an autonomous UAV inspection system based on object detection is designed. (1) To detect inspection objects, the corresponding dataset is established on the basis of the UAV autonomous inspection task. (2) To obtain the position information of the target object, a lightweight object detector based on the YOLOX network model is designed. First, the backbone is replaced with MobileNetv3. Next, in the neck structure, the Ghost module is introduced and depthwise convolution is applied instead of normal convolution. Then, to embed the location information into the channel attention, coordinate attention is introduced after the output feature layer of the backbone, enabling the lightweight network to operate on a larger area of focus. Finally, to improve the accuracy of the bounding box regression, the α-DIOU loss function is introduced. (3) To obtain the best image acquisition position, the results of object detection combined with the real-time status of the UAV are used. (4) To enable the UAV to complete the final corrections, position control and altitude control are used. Compared to the original YOLOX tiny, the new model improves the mAP_0.5:0.95 metric by about 2 percentage points, with a significant reduction in the number of parameters and computation, while running at 56 FPS on Nvidia NX. This system can effectively solve the problem of the target deviating from the center of the picture when the UAV takes pictures during a high-altitude autonomous inspection, verified by many actual flight experiments.

In recent years, multi-rotor unmanned aerial vehicles (UAV) have been widely applied for various applications; however, they are yet to be as commonly utilized in certain industrial transportation applications. Thus, this work designed and implemented a reference model-based integral sliding mode control (SMC) method applied to the velocity controller of a multi-rotor UAV. The designed controller was compared with an integral SMC scheme, and then the controller and modeling robustness were verified. Finally, the proposed method was applied to an industrial six-rotor UAV. Three experiments involving target-tracking, fixed-point hovering, and robustness verification were executed. A load of approximately 81.5% of the UAV’s self-weight was used to verify the robustness of the proposed scheme against parameter uncertainty. This work will serve as a meaningful reference for the application of the SMC in practical industrial applications.

Sliding mode control, an algorithm known for its stability and robustness, has been widely used in designing robot controllers. Such controllers inevitably exhibit chattering; numerous methods have been proposed to deal with this problem in the past decade. However, in most scenarios, ensuring that the specified form and the parameters selected are optimal for the system is challenging. In this work, the reinforcement-learning method is adopted to explore the optimal nonlinear function to reduce chattering. Based on a conventional reference model for sliding mode control, the network output directly participates in the controller calculation without any restrictions. Additionally, a two-step verification method is proposed, including simulation under input delay and external disturbance and actual experiments using a quadrotor. Two types of classic chattering reduction methods are implemented on the same basic controller for comparison. The experiment results indicate that the proposed method could effectively reduce chattering and exhibit better tracking performance.

With the development of science and technology, the traditional industrial structures are constantly being upgraded. As far as drones are concerned, an increasing number of researchers are using reinforcement learning or deep learning to make drones more intelligent. At present, there are many algorithms for object detection. Although many models have a high accuracy of detection, these models have many parameters and high complexity, making them unable to perform real-time detection. Therefore, it is particularly important to design a lightweight object detection algorithm that is able to meet the needs of real-time detection using UAVs. In response to the above problems, this paper establishes a dataset of six animals in grassland from different angles and during different time periods on the basis of the remote sensing images of drones. In addition, on the basis of the Yolov5s network model, a lightweight object detector is designed. First, Squeeze-and-Excitation Networks are introduced to improve the expressiveness of the network model. Secondly, the convolutional layer of branch 2 in the BottleNeckCSP structure is deleted, and 3/4 of its input channels are directly merged with the results of branch 1 processing, which reduces the number of model parameters. Next, in the SPP module of the network model, a 3 × 3 maximum pooling layer is added to improve the receptive field of the model. Finally, the trained model is applied to NVIDIA-TX2 processor for real-time object detection. After testing, the optimized YOLOv5 grassland animal detection model was able to effectively identify six different forms of grassland animal. Compared with the YOLOv3, EfficientDet-D0, YOLOv4 and YOLOv5s network models, the mAP_0.5 value was improved by 0.186, 0.03, 0.007 and 0.011, respectively, and the mAP_0.5:0.95 value was improved by 0.216, 0.066, 0.034 and 0.051, respectively, with an average detection speed of 26 fps. The experimental results show that the grassland animal detection model based on the YOLOv5 network has high detection accuracy, good robustness, and faster calculation speed in different time periods and at different viewing angles.

As the scale of the power grid continues to expand, drone inspection operations are becoming increasingly popular. However, most of the existing inspection drones are for transmission line inspection in the open field environment, with the characteristics of large size and high quality, which is difficult to be directly applied to transmission line inspection around the city area. To address the above issues, in this paper, a small unmanned aerial vehicle (UAV) inspection system is designed with the aim of achieving autonomous inspection of overhead ground wires in urban peripheral areas, combined with image processing technology, with a total weight of less than 400g. Specifically, during the inspection of small UAV, Raspberry Pi uses traditional image processing methods such as Hough transform to obtain the position information and pixel error of ground wire from real time video stream; the flight control system uses the results of image processing combined with data from millimetre wave radar to achieve conversion from pixel error to actual distance error; finally, the ground wire is made to be in the center of the video as much as possible through the correction strategy, thus realizing the autonomous inspection task of the small UAV along the line. The experimental results show that the small UAV can stably identify the target transmission lines and achieve autonomous flight along the lines with horizontal deviation within plus or minus 0.3 m and height deviation within plus or minus 0.1 m, which is of great reference value for the application of small UAV in urban transmission line inspection.