Obstacle Detection Model Using Point Cloud Shape for Automatic Doors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional object detection systems for automatic doors face challenges in accurately detecting obstacles on various road surfaces, leading to errors in determining the presence of obstacles and potentially causing the door to stop opening prematurely or collide with obstacles.
Innovation Solution
An object detection device that uses a sensor to acquire reflected waves from obstacles, generates a detection point cloud, and employs machine learning to create an object detection model. This model determines whether an object is an obstacle by analyzing the distribution shape of the detection point cloud, allowing for accurate obstacle detection regardless of the road surface.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional object detection sensors are used to detect obstacles for automatic door operation, then the door can perform opening operations, but detection errors occur on certain road surfaces (such as asphalt) causing false obstacle detection or premature door stopping
Solution Approach 1:
The patent transforms the raw detection point cloud data into a standardized three-dimensional coordinate system, converting detection results from various road surface conditions into a unified parameter space. This parameter transformation enables the machine learning model to learn invariant features that generalize across different road surfaces, thereby resolving the contradiction between detection reliability and measurement precision.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between the sensor data and the obstacle detection decision. This intermediary processes the raw detection point cloud through learned patterns, filtering out false detections caused by road surface characteristics while preserving true obstacle signals, thus improving both reliability and precision simultaneously.
2Reliability
If detection point cloud with noise is used for obstacle detection, then the door opening operation may be stopped prematurely or not performed at all, but this reduces the productivity and usability of the automatic door system
Solution Approach 1:
The patent performs preliminary processing of the detection point cloud by transforming it into a three-dimensional coordinate system and extracting relevant features before making the obstacle detection decision. This preliminary action filters out noise from road surface reflections while preserving true obstacle signals, enabling the system to maintain high safety standards without unnecessary door operation interruptions, thus improving productivity.
3Measurement precision
If machine learning model is used to detect obstacles by analyzing detection point cloud distribution, then high accuracy obstacle detection is achieved regardless of surrounding environments, but the device complexity increases
Solution Approach 1:
The patent replaces complex hardware modifications with a software-based machine learning model that processes the existing sensor data. By substituting mechanical or hardware complexity with computational algorithms, the system achieves high measurement precision while keeping the physical device structure relatively simple, thus resolving the contradiction between precision and complexity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system achieves high accuracy in detecting obstacles around vehicles, preventing collisions and ensuring smooth operation of automatic doors, even on diverse road surfaces.
Implementation Method 1
a sensor that transmits a probing wave and detects a reflected wave
Implementation Method 2
a reflected wave generated when a probing wave transmitted from a sensor installed in a door of a vehicle is reflected by an object around the vehicle
Data Source
AI summary
An object detection device according to an embodiment generates, in a learning phase, an object detection model by performing machine learning of a relationship between a feature vector indicating a distribution shape of a detection point cloud based on a reflected wave of a probing wave transmitted from a sensor and reflected by the object, and information indicating whether the object is an obstacle. In an estimation phase, the object detection device then calculates a detection point cloud as a position of the object on the basis of the plurality of acquired reception results, calculates a feature vector indicating a distribution shape of the detection point group on the basis of the calculated detection point cloud, determines whether the object is an obstacle on the basis of the calculated feature vector and the object detection model, and outputs a determination result.