Design And Mechanical Characteristics Analysis Of Underwater Two-Arm Robot

As mobile robotics continues to iterate,SLAM technology,one of its core technologies,has also made great strides.The pluralism required and the complexity of the application scene makes it hardship for a singleton sensor to guarantee high precision,real-time and robust environment perception for mobile robots in a variety of environments.Visual SLAM is a mature technology for mobile robot perception,but it is heavily influenced by illumination and performs poorly in weakly textured environments and scenarios with missing feature points.Inertial measurement unit(IMU)can provide scale constraints for the vision sensor,and it has good complementarity with it.To address this issue,a Grand Slam system based on eye vision and IMU is proposed to improve the accuracy and robustness of the localization of mobile robots.The main work on the content of this paper is as follows.1)The mathematical model has been studied to constitute the principle and error source to solve the problem of position and status estimation of the persistence of visual inertia in motion.Further,the binocular camera model and its distortion processing are investigated,and calibration experiments for the double-eye camera.The experimental results show that the reprojection error of the left-eye and right-eye camera is distributed within the interval of plus or minus 0.5 pixels,and the calibration results meet the requirements of the SLAM system.2)The SLAM algorithm based on point-line integration features is proposed to solve the problem of insufficient function and the problem points of SLAM loss based on a single function,combined with the sensing characteristics of the front-end mileage table.First extract and match this feature point,and then use the improved LSD line function to extract the image and estimate pose of the camera.Experimental results show that compared with the original LSD algorithm,the improved LSD algorithm will reduce detection time by 51%,the number of extracted line segments shrinks by 62% and the point-line feature fusion SLAM system based on the improved LSD algorithm achieves higher localization accuracy.3)To address the problem of binocular vision and IMU information processing,a tightly coupled approach is used to fuse the data from both by a nonlinear optimization method.The IMU measurement model and error sources are investigated,and calibration experiments are conducted on the IMU.Experimental results show that the Gaussian white noise of the accelerometer and gyroscope is very close to the true value.Then execute the IMU pre-fusion process to ensure that the IMU is aligned with the image data,and initialize the two eye cameras and the IMU for the system.In back-end optimization,nonlinear optimization arithmetic based on sliding windows is used to solve the optimal bit pose and the cumulative error is eliminated based on the loop probe module based on the word bag model.4)The SLAM system validation problem for the fusion of binocular camera and IMU is tested in the Eu Ro C public dataset and in realistic corridor scenarios.The quality of the extraction line segment has been significantly improved.In the data set comparing the mainstream robust schemes and algorithms in this article,the physical verification platform of the SLAM system is built through the selection of mobile platforms and sensors,etc.,the joint multi-sensor calibration experiments,and finally the SLAM system positioning experiments and related performance verification experiments are designed and completed.The experimental results show that the line features are added to the visual inertia Grand Slam system,its positioning precision can be improved by up to 7.6cm,and there is no feature loss,and the SLAM can be basically completed in the real corridor environment,which verifies the effectiveness of the system.