Research On Monocular Visual-inertial SLAM And Path Planning Algorithm For Mobile Robots

With the rapid development of disciplines such as control theory,artificial intelligence,electronic information,and computer technology,mobile robots have been used unprecedentedly in various industries.For a mobile robot,the SLAM(Simultaneous Localization and Mapping)algorithm is used to determine the position of the robot in the environment and map the environment around the robot,and the path planning algorithm can ensure that the robot reaches the target position safely from the starting point without collision.The visual SLAM of the mobile robot and the research of path planning algorithm is of great significance.For the path planning algorithm part,this paper proposes a static path planning algorithm based on heuristic methods,a static path planning algorithm based on artificial potential field and A-star fusion,and a dynamic path planning algorithm based on an improved DWA algorithm;for the SLAM algorithm part,a monocular visual-inertial SLAM algorithm based on weighted median pre-integration is proposed in this article.Research on the technology of global path planning algorithm based on the heuristic method.In some cases,considering that the traditional A-star algorithm's own node search strategy has many shortcomings of path inflection points and large turning angles,and the feasible path generated by it is not optimal in theory,so this paper proposes a global path planning algorithm based on the heuristic method,we evaluate the performance of the path planning algorithm based on heuristic methods by designing different map environments,and define the evaluation indicators,and finally compare with the traditional A-star algorithm and the other two related algorithms to evaluate the effectiveness and feasibility of path planning algorithms based on heuristic methods.Research on the static path planning algorithm based on the fusion of artificial potential field and A-star algorithm.To solve the problem that the A-star algorithm cannot handle dynamic obstacles and the path generated by the artificial potential field method is not optimal,this paper proposes a fusion path planning algorithm based on the artificial potential field method and the A-star algorithm.Then,we designed different map layout characteristics to evaluate the performance of the static path planning algorithm based on the fusion of artificial potential field and A-star algorithm and compare the artificial potential field and A-star fusion static path planning algorithm proposed in this paper with other related methods.We conducted a comparative evaluation.The experimental results show that the proposed static path planning algorithm combining artificial potential field and A-star is effective and feasible.Research on dynamic path planning algorithm based on improved DWA(Dynamic Window Approach,referred to as DWA algorithm)algorithm.Considering that in dense obstacle areas,the dynamic window method has the behavior of bypassing the dense obstacle area when crossing dense obstacles,which will increase the total distance;at the same time,when encountering "C"-shaped obstacles,The target cost function of the standard DWA algorithm will be invalid,resulting in no path.Aiming at this problem,this paper introduces a constraint condition of the distance scoring between the current point and the target point based on the basic DWA algorithm.Then we designed the characteristics of the map layout in different situations.In the process of experimental evaluation,we used the standard DWA algorithm as a reference method to evaluate the performance of the dynamic path planning algorithm based on the improved DWA algorithm proposed in this article.The experimental results show that the dynamic path planning algorithm based on improved DWA is effective and feasible.Research on SLAM algorithm based on monocular vision inertial sensor.Considering that the existing pre-integration methods usually use Euler integration or median integration numerical integration methods,this integration method is simple to calculate,but the accuracy is lost in the integration process,which has an impact on the accuracy of the final positioning result.Therefore,this paper proposes a monocular visual-inertial odometer method based on weighted median pre-integration.We first introduced the experimental environment and experimental conditions,mainly including the Eu Ro C data set used for the performance evaluation of the monocular visual-inertial SLAM system and the hardware and software conditions required for the experiment.We use the systems VI?ORB?SLAM2,ROVIO,VINS-Mono as reference methods,Using the absolute position root mean square error RMSE as the evaluation index,the performance of the monocular visual-inertial SLAM system based on the weighted median pre-integration proposed in this paper is evaluated.The experimental results show that the monocular visual-inertial SLAM with weighted median pre-integration proposed in this paper is effective and feasible.