Ship Direction Estimation by Normal-Line Clustering at Berths
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for autonomous ship berthing require a measurement device like lidar to be mounted on the ship, which is not always feasible. To address this, a system with a measurement device at the berthing place is needed to accurately estimate the ship's direction based on measurement data.
Innovation Solution
An information processing device that acquires measurement data from a measurement device at the berthing place, generates clusters of measured points based on normal lines of the ship's surface, and estimates the ship's direction from the cluster with the largest number of points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a measurement device is installed at the berthing place instead of on the ship, then the system complexity on the ship is reduced and autonomous berthing can be achieved, but it becomes difficult to accurately identify the ship's direction from the measurement data
Solution Approach 1:
The patent segments the ship's measurement data into multiple clusters based on normal line directions. By dividing the point cloud data into distinct clusters corresponding to different ship surfaces, the system can identify the ship's direction through cluster analysis rather than attempting to directly measure the entire ship at once. This segmentation approach enables accurate direction identification while using a fixed measurement device at the berthing place.
Solution Approach 2:
The patent introduces a new dimension for analysis by calculating normal lines at each measured point and clustering based on normal line directions. Instead of analyzing only the spatial coordinates (x, y, z), the system adds the normal line direction as an additional dimensional feature, enabling more accurate ship direction identification from the measurement data obtained by the fixed device.
2Shape
If the ship contains many non-uniform shapes, then the ship can have complex structures, but it becomes more difficult to accurately estimate the ship's direction based on measurement data
Solution Approach 1:
The patent addresses complex ship shapes by segmenting the measurement data into multiple clusters, where each cluster represents a distinct surface or region of the ship. By analyzing the normal lines of points within each cluster separately and then synthesizing the results, the system can accurately determine the ship's overall direction despite the presence of non-uniform shapes and complex structures.
Solution Approach 2:
The patent applies local quality analysis by examining the normal line directions at local points on the ship's surface and clustering them based on their orientations. This local analysis approach allows the system to handle complex ship geometries by processing different regions with their unique characteristics separately, then combining the local results to determine the global ship direction.
Data Source
AI summary
The controller 13 of the information processing device 1A functions as the acquisition unit, the cluster generation unit, and the ship direction estimation unit. The acquisition unit is configured to acquire measurement data which is a set of data representing a plurality of measured points measured by a measurement device. The cluster generation unit is configured to generate, based on normal lines of the measured points of a ship, one or more clusters of the measured points of the ship. The ship direction estimation unit is configured to estimate a direction of the ship based on a first cluster which is a cluster having a largest number of the measured points of the ship.