Search

This paper introduces a general state estimation framework fusing multiple sensor information for hybrid wheeled-legged robots performing mobile manipulation tasks. At the core of the state estimator is a novel unified odometry for hybrid locomotion which can seamlessly maintain tracking and has no need to switch between stepping and rolling modes. To the best of our knowledge, the proposed odometry is the first work in this area. It is calculated based on the robot kinematics and instantaneous contact points of wheels with sensor inputs from IMU, joint encoders, joint torque sensors estimating wheel contact status, as well as RGB-D camera detecting geometric features of the terrain (e.g. elevation and surface normal vector). Subsequently, the odometry output is utilized as the motion model of a 3D Lidar map-based Monte Carlo Localization module for drift-free state estimation. As part of the framework, visual localization is integrated to provide high precision guidance for the robot movement relative to an object of interest. The proposed approach was verified thoroughly by two experiments conducted on the Pholus robot with OptiTrack measurements as ground truth.

In order to cope with the operational instability and scheduling complexity brought by wind power access to the grid, this paper aims at multi-domain optimal dispatching of wind power systems based on voltage stability constraints. The wind power system is divided into different grid operation domains, and the system is divided into normal domains, abnormal domains, and emergency domains according to the regulation capacity of conventional units and their changing trends. The optimal scheduling model considering voltage stability constraints in different operation regions is established. Through the simulation calculation of a typical example of IEEE-30 node, it is verified that the optimized scheduling method can improve the economy, security and wind power capacity of the system operation.

For the actual shopping scenes in supermarkets, there will be more complex background environment, more diverse detection objects, and more changeable actual scenes, so the traditional image recognition network cannot meet the existing needs. In this paper, Convolutional Neural Network technology and YOLOv3 target detection algorithm are adopted to realize the recognition and detection of target objects in images, automated analysis and processing of these product image data with computer vision algorithms. The convolution neural network and COCO data set are used to train the neural network. Finally, the well-trained network is tested. Tested, the effect is good.

With the increasing global popularity of self-driving cars, there is an immediate need for challenging real-world datasets for benchmarking and training various computer vision tasks such as 3D object detection. Existing datasets either represent simple scenarios or provide only day-time data. In this paper, we introduce a new challenging A*3D dataset which consists of RGB images and LiDAR data with significant diversity of scene, time, and weather. The dataset consists of high-density images ($\approx~10$ times more than the pioneering KITTI dataset), heavy occlusions, a large number of night-time frames ($\approx~3$ times the nuScenes dataset), addressing the gaps in the existing datasets to push the boundaries of tasks in autonomous driving research to more challenging highly diverse environments. The dataset contains $39\text{K}$ frames, $7$ classes, and $230\text{K}$ 3D object annotations. An extensive 3D object detection benchmark evaluation on the A*3D dataset for various attributes such as high density, day-time/night-time, gives interesting insights into the advantages and limitations of training and testing 3D object detection in real-world setting., Comment: A new 3D dataset by I2R, A*STAR for autonomous driving

Seismic communication might promise to revolutionize the theory of seismic waves. However, one of the greatest challenges to its widespread adoption is the difficulty of signal extraction because the seismic waves in the vibration environments, such as seas, streets, city centers and subways, are very complex. Here, we employ segmented correlation technology with Morse code (SCTMC), which extracts the target signal by cutting the collected data into a series of segments and makes these segments cross-correlate with the decoded signal to process the collected data. To test the effectiveness of the technology, a seismic communication system composed of vibroseis sources and geophones was built in an environment full of other vibration signals. Most notably, it improves the signal-to-noise ratio (SNR), extending the relay distance and suppressing other vibration signals by using technology to deal with seismic data generated by the system.

Large numbers of seismic channels and high-density energy-efficient acquisition systems are the development trend of seismic instruments and have attracted high R&amp, D interests in recent years. The combination of remote sensing and wireless sensor network technology provides superior observation capabilities for high-density seismic exploration. However, large-scale and multi-node acquisition methods place higher requirements on time synchronization performance. Seismic data with poor time synchronization will cause considerable errors in the interpretation of seismic data and even have no practical significance. Thus, the strict time synchronization performance is the prerequisite and basis for the application of cable-less storage seismograph in high-density seismic array applications. The existing time synchronization methods have high power consumption and poor time synchronization accuracy, which is not suitable for the long-time task. In addition, these methods are affected by the number of nodes and the distance. This paper presents an energy-efficient time-sharing indexed interpolation intercept method for the seismic data synchronization. The time synchronization method uses the high-precision TCXO as the main clock and records GPS time in the SD card at intervals to achieve the high-precision timestamp for the seismic data. Then the seismic data is intercepted intermittently based on precise time stamps, which achieves the strict seismic data synchronization. Performance analysis shows that the time synchronization accuracy of the proposed method is 0.6 &mu, s and saves 73% energy of the time-sync periods compared to the common GPS timing method. The field measurement results indicate that the time synchronization accuracy is not associated with the working time and the distance between nodes so that the proposed synchronization method is suitable for the high-density seismic survey.

The operational instability and scheduling complexity is brought by the high proportion of renewable energy access to the grid. In order to cope with this problem, this paper proposes a method based on the mean voltage stability L-index analysis of stochastic voltage stability and the grid operating domain. This method can quickly find the weakest node in the whole network and help the dispatchers monitor and control the weak nodes. Based on the observation of the voltage stability of the whole network, the grid operation domain division index is proposed, and the grid is divided into different operation domains based on the voltage stability constraint. The validity and correctness of the proposed scheme and indicators are verified by the typical example of IEEE-30 node.

New development of classification system and safety-protection strategy of geologic survey data is summarized for searching optimized solutions of disaster tolerance technology for mass geologic survey data acquired in recent years by Chinese geological organizations. The new disaster tolerance technology solution which is building on the basis of extensive research and review of empirical data from national and international best practices and comparing the pros and cons of data backup and disaster tolerance technologies can provide some basic network supports for social services of geological data.

In this paper, taking the slice method as slope stability analysis model, a new method was presented to evaluate the slope stability combining with the self-adaptive genetic algorithm (SGA) simulating biological evolution. This method optimized the slope stability analysis model to overcome the local convergence, avoid precocity phenomenon and calculate the minimum safety factor by searching out the most dangerous sliding surface with the search variables of slip circle center and radius. Finally, it is pointed that the method is reasonable and applicable in project application.