Vehicle Collision Prediction Using Combined Position Probabilities
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current advanced driver assistance systems (ADAS) primarily focus on forward-facing traffic participants and use simplistic constant velocity predictions for Time to Collision (TTC) calculations, failing to account for lateral or rear collisions and lacking in accuracy.
Innovation Solution
The system employs a processing circuitry with a neural network to predict the future positions of both the vehicle and external traffic participants, incorporating motion state information, map data, and interaction models to calculate combined probability measures, enabling more accurate and realistic predictions of potential collisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a front looking camera is used for Forward Collision Warning, then the system can identify traffic participants in the front direction, but the system fails to consider traffic participants in other directions (lateral, rear)
Solution Approach 1:
The patent applies universality by using a single front-looking camera to perform multiple functions: detecting traffic participants in front, lateral, and rear directions through image processing and coordinate transformation, eliminating the need for separate cameras for each direction
Solution Approach 2:
The patent transforms the detection problem from a 3D spatial coverage issue to a 2D image processing issue by using coordinate transformation to map detected objects from front camera views to their actual positions in the environment, enabling multi-directional detection without additional cameras
2Measurement precision
If constant velocity prediction is used for TTC calculation, then the calculation is simple, but the prediction accuracy is low
Solution Approach 1:
The system performs preliminary detection and tracking of traffic participants before collision risk assessment, maintaining continuous motion state information that feeds into the prediction model, enabling more accurate predictions without complex real-time calculations
Solution Approach 2:
The patent changes the prediction approach from simple constant velocity to a more sophisticated model that incorporates acceleration and interaction forces, improving prediction accuracy while managing complexity through efficient calculation methods
3Reliability
If only forward-facing traffic participants are considered, then the system remains simple, but collision risk assessment is incomplete
Solution Approach 1:
The patent segments the detection task by using a single camera to capture different fields of view at different times, processing images sequentially to build a comprehensive picture of the environment, dividing the complex multi-directional detection into manageable temporal segments
Solution Approach 2:
The system uses coordinate transformation and image processing algorithms as intermediaries to translate front camera data into comprehensive environmental awareness, acting as a mediator between the limited camera input and the comprehensive collision risk assessment requirement
Data Source
AI summary
An apparatus for assisting a control system and/or a driver of a vehicle is provided. The apparatus comprises a processing circuitry configured to: determine for a plurality of positions within a vicinity of the vehicle, based on motion state information of the vehicle, a respective first probability measure that the vehicle will be at the respective position at one or more future time instants; determine for the plurality of positions, based on motion state information of at least one traffic object, a respective second probability measure that the at least one traffic object will be at the respective position at the one or more future time instants; and determine for the plurality of positions, a respective combined probability measure that the vehicle and the at least one traffic object will be at the respective position at the one or more future time instants.