Font Size: a A A

Research On Function Allocation Of Robotic Tractor Teleoperation System

Posted on:2020-06-11Degree:MasterType:Thesis
Country:ChinaCandidate:S X DongFull Text:PDF
GTID:2492306314484674Subject:Vehicle Engineering
Abstract/Summary:PDF Full Text Request
The combination of teleoperation technology and robotic tractors means the robot’s intelligence is integrated with human intelligence through the way of"human-in-the-loop".And robot tractors will have a new development trend-based on human-machine collaboration,which means operators need to interact with the automation system and cooperate with the automation system to complete tasks.This will not only can fully play the role of human beings,but also improve the efficiency of the task and the intelligent level of the automation system greatly.Therefore,the rational allocation of functions between the operator and the automation system is significant to the robotic tractor teleoperation system premised on safe and efficient operation.By reading literatures,we can see that there are two approaches used to allocate function between man and machine in the fields of industrial automation,aerospace and aircraft:one is static function allocation and the other one is dynamic function allocation.Since the static function allocation may generate some problems,such as the phenomenon of "human out of the loop",the operator lack of situational awareness or alertness due to excessive dependence on automation system.Therefore,it is necessary to complete the dynamic allocation functions,that is,to reallocate functions according to the changes of real-time environment conditions.In this paper,the concept of man-machine cooperation is integrated into the intelligent agricultural robot vehicle system,and the function allocation in the man-machine cooperation system of tractor robot teleoperation is explored.The main contents of this article are shown below:1.Design of human-machine collaborative model.The teleoperation platform built in the early stage of the laboratory is introduced.According to the characteristics of the teleoperation system,a cooperative operation mode was proposed,which integrates the principle of situational awareness and the principle of human-machine function allocation.The human-machine operation system was designed,including user interface design,data processing design and communication design.2.The analyses and decomposition of function.The teleoperation platform built in the early stage of the laboratory is introduced.According to the characteristics of the teleoperation system,three remote operation modes was proposed,which integrates the principle of situational awareness and the principle of human-machine function allocation.The remote operation system was designed,including user interface design,data processing design and communication design.3.Detection and location of obstacles.By analyzing the development of object detection technology,the method based on information fusion is selected to complete obstacle detection and location.The structure of Faster R-CNN network and data processing process of lidar are described in detail.A object detection model of Faster R-CNN based on VGG-16 was built.The sample training set and test times were established.The detection accuracy of pedestrians and tractors is 90.2%and 86.3%respectively.The detection results of the two methods were fused by ICP algorithm to complete the detection and location of obstacles.Therefore,the robotic tractor can know the task requirements.4.Research on Function Allocation.According to the comparative analysis of man-machine capabilities,static function allocation is completed.Then,according to the real-time environmental factors and actual requirements of agricultural operation,the principles of when to change the automation level and how to change the automation level are summarized.The flow of BP neural network,genetic algorithm and adaptive algorithm were expounded,and the automation level prediction model of adaptive genetic BP neural network was established.The simulation results show that the prediction effect of the adaptive genetic BP neural network is the best.
Keywords/Search Tags:teleoperation, agricultural robot, function allocation, adaptive genetic algorithm, BP neural network
PDF Full Text Request
Related items