Research On Adaptive Enhanced Display Of Human-Machine Interface In Wide-Field Supervisory Tasks

With the rapid development of computer and information technology,intelligent operation control system presents the trend of digitalization of information.Wide field of vision and multi-screen digital human-machine integrated supervisory system has been widely used in rail transit,military command,traffic management,aerospace,schools and hospitals and other security-critical areas.However,due to the rapid increase of information,it becomes extremely difficult for users to effectively supervise various complex system information in real time.The missing of critical system information will bring serious potential risks to the system operation.In view of the above contradictions in human-machine interaction,this thesis optimized the interface presentation and human-machine interaction process based on adaptive interface enhancement technology according to the attention behavior characteristics of supervisors and the real-time characteristics of task process,in order to make up for the bottleneck of information transmission between users and wide-field screen under supervisory tasks.The main work of the thesis includes the following sections:1.Research on gaze tracking algorithm in wide field of viewIn order to track users' attention behavior in real time in wide-field supervisory scenarios,it is necessary to construct a gaze tracking algorithm for wide-field supervisory tasks.Through data conversion based on human spatial coordinate expression,the three-dimensional coordinate system of human vision was obtained,and then the visual gaze points of the supervisors on the screen were solved.The error sources of the algorithm were discussed in depth,and the calculation process of the algorithm was simplified.The algorithm of spatial depth distance was compared with other existing schemes.Finally,the feasibility of the algorithm was proved by simulation experiments.2.Application of gaze tracking algorithmThis thesis studied how to apply the above gaze tracking algorithm to the actual production environment.The characteristics of input data of wide-field gaze tracking algorithm were analyzed,and the limitations and requirements of input equipment and data were put forward.The weighted average filter algorithm was used to restrain the random errors introduced by the input devices.An application example of system hardware and software suitable for real environment was constructed.The error test was carried out in the actual scene to verify the feasibility and stability of the system.Finally,the real-time performance of the system and the error sources of the system were analyzed.This provided a basic data solution for the input layer of the adaptive interface.3.Attention allocation model for supervisorsUsing the importance of all the information in supervisory tasks,the information entropy function could be quantified as the priority of information.Based on the psychological entropy of supervisors,a prediction model of attention allocation for supervisory tasks was constructed.The mechanism of value priority optimization in the model was discussed in depth.Finally,the ergonomic experiment proved the validity of the model,and compared it with the existing model.This provided a theoretical support for the modelling in the decision layer of the adaptive interface.4.Research on adaptive enhancement display interfaceThe construction process of adaptive enhanced display interface in wide field of view was studied.Firstly,based on the entropy method,the information priority in system tasks and the eye movement and gaze behavior of supervisors were weighted,and the comprehensive influencing factors of information value were obtained.Through these factors,the information of the interface was tracked in real time.The saliency of the interface,i.e.the output layer of the adaptive interface,was adaptively changed based on the attention allocation prediction model in order to guide the user's attention behavior.Finally,experiments showed that this adaptive interface model would effectively enhance the task performance of supervisors.