| With the continuous promotion of agricultural facilities and modernization,greenhouse robots with automatic navigation technology have gradually become one of the research hotspots for intelligent agricultural equipment.Due to the problems of narrow operating paths and GPS signal loss in the facility environment,this paper uses the multi-sensor fusion SLAM(Simultaneous Localization and Mapping)as the localization system and a four-wheel independent steering and four-wheel independent drive(4WIS-4WID)structural experimental prototype is used to carry out the navigation study of greenhouse robot and make improvements.The specific works are as follows.(1)First,the experimental prototype and control system were built and improved to meet the requirements of the subject.To address the problem of low accuracy of traditional image feature matching in visual navigation,an image feature matching algorithm based on the combination of improved uniform ORB and LK optical flow is proposed.Based on this feature matching algorithm,a lightweight visual odometry is proposed in combination with a depth camera,and the visual odometry is fused with laser SLAM to build a multi-sensor fusion SLAM system.(2)Second,in order to reduce the number of inflection points of the planned path and improve the quality of the tracking path,an improved path planning algorithm is proposed based on the A* algorithm.The evaluation function of the A* algorithm is improved to optimize the path planning results by designing the penalty relationship function of the inflection points on the planned paths.The effectiveness of the algorithm is verified by MATLAB simulation tests.(3)Again,a path tracking algorithm based on 4WIS-4WID structure and model predictive control is proposed to improve the accuracy and efficiency of path tracking for greenhouse robots with 4WIS-4WID structure.Based on the analysis of the relationship and defects between the traditional pure tracking algorithm and the vehicle structure,a model predictive controller with the deviation degree of the vehicle as the state quantity and the steering center position as the control quantity is designed.The algorithm is tested by MATLAB software simulation to verify its feasibility.(4)Finally,the greenhouse robot navigation experiments were conducted.The feature extraction and matching experiments show that this algorithm can improve the feature matching accuracy by 19.73% on average and shorten the total matching time by 56.34% on average compared with the traditional method;the positioning accuracy experiment shows that the average positioning error of this positioning system and the ultra-wideband system is 4.31 cm,and the positioning accuracy is between 5.69 cm and 14.31 cm;the path tracking experiments show that the average deviation of the path tracking based on this algorithm is between 98 and 132 mm.The average deviation of the path tracking based on this algorithm is between98 and 132 mm,and the stabilization time after line change is between 2.3 and 4.2 s,which is much better than the traditional pure tracking algorithm.The greenhouse robot based on multi-sensor fusion SLAM proposed in this paper has good steering flexibility and can perform autonomous localization and map building.All experimental results verify the effectiveness and advantages of this research method,and it can meet the demand of automatic navigation operation in greenhouse,and provide technical reference for automatic navigation technology of facility mobile robot. |
PDF Full Text Request