Depth Perceived Trajectory Mapping using Precedent Obstacle Avoidance (POA) for Optimized AGV Locomotion

Machine Vision builds up a significant role in the act of achieving an instantial near perfect collision-free locomotion solution in AGV (Automated Guided Vehicle). For conservation of memory structured from earlier explored non-optimum path by the concerned mobile agent, VI (Visual Inference) constructs a technical precision in decision making to choose optimized path. Representation of transformed Geometric configurations of on-route obstacles in a specific indoor environment into spatial coordinates is achieved by Visual Perception. This paper basically focuses on collection of 3D depth scanned data of concerned indoor surrounding using RGB-D (Red, Green, Blue Depth) sensor for proper detection and avoidance of obstacles. The identification of obstacles is categorized using specially proposed technique of Sector Divided Precedence Obstacle Avoidance (POA). The work extends its findings with the incorporation of the results obtained from implementing the proposed 3D Depth SLAM (Simultaneous Localization And Mapping) for trajectory mapping. The subsequent sections portray structural working architecture, and experimental analysis with a brief numerical description of the constructed required algorithms, which would serve as an important memory for accomplishing optimum PTP (Point-To-Point) navigation by AGV in a specific GPS (Global Positioning System)-denied indoor architecture as well as outdoor environment. Visual and tabular presentation of experimental output confirms the reliability of the proposed research procedure to be carried out further in the domain of MRN (Mobile Robot Navigation).