Neural Vision Collision Avoidance for Personal Mobility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Personal mobility devices such as electric kickboards and bicycles face a high risk of collision accidents due to their operation on roads and sidewalks, necessitating effective collision avoidance methods to ensure user safety.
Innovation Solution
A collision avoidance system that utilizes a neural network model to recognize objects in a driving image, calculate relative distances, and determine required collision times, providing collision warnings or controlling the device to brake when necessary, incorporating features like object tracking and Kalman filter prediction for accurate distance calculation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If personal mobility devices operate on roads and sidewalks, then transportation convenience is improved, but collision accident risk increases
Solution Approach 1:
The system performs preliminary object recognition and collision risk assessment before actual collision occurs. The neural network model continuously monitors the environment and predicts potential collision scenarios, enabling preventive actions to be taken in advance, such as warning signals or automatic braking interventions.
Solution Approach 2:
The system implements continuous feedback loops through real-time image acquisition, object recognition, and collision risk calculation. The calculated collision risk information feeds back to control the moving body's operation, creating a closed-loop safety system that continuously adjusts based on environmental conditions.
2Reliability
If object recognition and collision calculation are performed in real-time, then collision avoidance capability is improved, but system complexity increases
Solution Approach 1:
The system replaces complex mechanical sensing and measurement systems with neural network-based image recognition. Instead of using multiple physical sensors to detect objects and calculate distances, the system uses visual information processing through neural networks to achieve object recognition and collision risk assessment.
Solution Approach 2:
The system creates a virtual representation of the physical environment through image capture and neural network processing. The bounding boxes and recognized object data form a digital copy of the scene that can be analyzed for collision risks without requiring direct physical measurement of all objects.
3Measurement precision
If neural network model is used for object recognition, then recognition accuracy is improved, but processing time increases
Solution Approach 1:
The system uses bounding boxes to partially represent objects rather than full detailed recognition. The neural network focuses on identifying key features and spatial relationships necessary for collision assessment, performing sufficient recognition to determine collision risk without exhaustive object analysis.
Data Source
AI summary
There is provided a collision avoidance method of a moving body collision avoidance device including acquiring a driving image of the moving body, recognizing an object in the acquired driving image using a neural network model, calculating a relative distance between the moving body and the object based on the recognized object, calculating a required collision time between the object and the moving body based on the calculated relative distance, and controlling an operation of the moving body based on the calculated required collision time.