| As the population continues to grow and urbanization advances rapidly,the number of large buildings in cities has increased significantly,and their interiors are becoming increasingly complex.In the face of this intricate indoor environment,wayfinding,as a basic human spatial activity,is becoming more and more difficult,and indoor navigation with wayfinding as its purpose has a broad audience and potential."People-oriented" is the core of the scientific concept of development.Starting from human needs,paying attention to people's perception of navigation environment in complex indoor environment,improving people's acceptance and satisfaction of indoor navigation services,and establishing human-centered pedestrian navigation are becoming increasingly important research topics.Indoor navigation includes three closely related aspects: indoor model construction,indoor precise positioning,and indoor path planning.Supported by accurate indoor positioning technology,indoor model construction and path planning algorithm will be the key to indoor navigation.However,the current research on indoor navigation still has some problems in these two aspects that need to be solved:(1)Most of the existing studies are based on the 2D representation of the indoor environment or the spatial division of buildings without indoor spatial elements in a 3D framework.Few studies have discussed fine-grained and context-aware spatial division in 3D environments.Meanwhile,existing approaches ignore the influence of human behavior on the change of navigable space,while human interactions with indoor spatial elements occur all the time in daily life and constantly change the navigability of indoor space.(2)Most existing indoor spatial frameworks and navigation methods ignore the dimensional properties of the navigating user and implicitly treat the user as a point on the indoor plane.However,due to the complex indoor environment,differences in different dimensions among individuals can lead to different path choices even in the same navigation scenario.Therefore,ignoring the user's dimension may result in planned paths that lack sufficient and comfortable space for the desired movement due to indoor obstacles.(3)The existing navigation road network lacks a visual representation with scale characteristics,ignoring the cognitive and memory characteristics of navigation users for navigation road networks and navigation paths.In fact,the existing visual indicators also rely solely on the unique features of the roads themselves to attract attention,while lacking attention to the visual competition relationships that are crucial in the visual attention process.As a result,existing evaluation metrics can usually be considered as self-significant metrics that work at the local scale,while ignoring the cognitive hierarchy at the global scale,which leads to the poor ability of these metrics to identify grid-like path networks and cannot be applied to large grid-like indoor road networks.Path planning algorithms also mostly use a single shortest path algorithm,which is not conducive to users' cognition and memory of navigation paths.In this paper,corresponding solutions are proposed for the above problems,and relevant validations are conducted by experimental studies.(1)To address the shortcomings of the dynamic expression of the existing indoor space division framework,a dynamic space division framework is proposed.The framework can flexibly define different navigable spaces by considering human behavior rules and human perception of the environment.The details include: studying the reasons for the dynamic changes of indoor navigation environment,defining the affiliation and usage attributes of indoor space elements from fine-grained and coarse-grained perspectives,respectively,to determine their usability and navigability of interaction spaces;designing rules for overlapping subspace reclassification,reconstructing non-overlapping subspaces with multiple semantics and determining their navigability;judging the influence of navigation users on navigable spaces;developing a method for generating navigable spaces from coarse-grained to fine-grained by different granularity of navigable spaces.(2)To address the shortcomings of existing indoor navigation in considering the user dimension,a customized comfortable navigable space generation method is proposed.The method can generate customized comfortable navigable spaces for users by defining and considering the physical characteristics of indoor users in 3D space.The details include: analyzing and defining the object space of the navigation user;determining the interaction space of indoor spatial elements based on their function,location,size and the space occupied by the user;studying a method for generating a narrow gap between obstacles(which the user cannot pass comfortably)by considering the non-navigable interaction spaces and the object spaces of all indoor spatial elements as obstacles,and generating customized comfortable navigable spaces accordingly.(3)To address the shortcomings of the existing visual representation of the navigation path network for indoor navigation,a visual hierarchical approach to indoor navigation path networks and a corresponding path planning algorithm are proposed.These methods utilize the human visual attention mechanism to enhance the visual features of the grid-like path network and provide help for users to construct mental maps and memorize paths.The details include: studying the degree of attraction of geometric and psychological competing factors to human visual attention at local and global scales by simulating the visual attention mechanism;constructing a mutual saliency metric framework for path networks to improve the visual recognition of grid-like navigation path networks;developing a hierarchical approach and path planning algorithm based on the visual saliency of path networks.(4)This paper verifies through simulation experiments that the dynamic space division framework can provide suitable navigation space for specific users in dynamic indoor environment;the proposed navigable space generation method can provide personalized adaptive navigation space and friendly paths in comfortable navigation environment for users with different dimensions;the proposed visual hierarchical method can simulate human eyes' perception of grid-like path networks,and the generated navigation paths are in line with users' cognitive habits and can help users remember the path network structure and paths. |
PDF Full Text Request