Parameterized Engine for Vehicle Distance Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automotive systems for estimating the distance between a motor vehicle and external objects using camera images lack sufficient accuracy, which can be critical for collision avoidance, especially in autonomous driving scenarios, and often incur high computational costs.
Innovation Solution
A method employing a parameterized calculation engine trained by machine learning, utilizing both digital image characteristics and reference points from sensors like LiDAR and radar, to provide corrected-position information for external objects, thereby improving distance estimation accuracy while reducing computational expenses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If processing of digital images is applied to detect external objects and estimate their characteristics, then object detection capability is provided, but positioning reliability is insufficient
Solution Approach 1:
The patent introduces a parameterized calculation engine as an intermediary component that receives image processing characteristics and sensor reference points, then computes corrected position information. This intermediary bridges the gap between unreliable image-based positioning and accurate sensor-based positioning, synthesizing both data sources to achieve both detection capability and positioning reliability.
Solution Approach 2:
The patent replaces pure image processing mechanisms with a hybrid system that incorporates sensor reference points (from radar, LiDAR, or ultrasonic sensors) to correct image-based position estimates. This substitution of mechanical/optical processing with sensor-fusion-based computation significantly improves positioning accuracy while maintaining object detection capabilities.
2Measurement precision
If multiple sensors and processing algorithms are used to improve positioning accuracy, then distance estimation accuracy is improved, but computational costs increase
Solution Approach 1:
The patent applies local quality by processing sensor data and image characteristics differently based on their specific properties and roles. Image processing characteristics provide object detection and preliminary positioning, while sensor reference points provide correction data. The parameterized calculation engine applies appropriate weighting and correction factors locally to each data source, optimizing accuracy while managing computational resources efficiently.
Solution Approach 2:
The patent utilizes parameter changes through the parameterized calculation engine that adjusts correction factors and positioning weights based on input data characteristics. By dynamically changing computational parameters based on data quality and availability, the system achieves high accuracy when possible while reducing computational costs when data is insufficient or unreliable.
3Difficulty of detecting and measuring
If image processing is used to estimate object characteristics, then object detection is achieved, but position information accuracy is insufficient for collision avoidance
Solution Approach 1:
The patent merges image processing characteristics (providing object detection and preliminary position) with sensor reference points (providing accurate position correction) through the parameterized calculation engine. This combination allows the system to maintain object detection capability while significantly improving position information accuracy to levels suitable for collision avoidance.
Solution Approach 2:
The patent applies preliminary action by using image processing to first detect objects and obtain preliminary position information, then using sensor reference points to correct these preliminary estimates. This two-stage approach ensures object detection is performed efficiently first, followed by accurate position correction, separating detection and precision functions.
Data Source
AI summary
A method and apparatus for estimating a distance between a motor vehicle and an external object, the vehicle being equipped with at least one camera for capturing digital images in a capture field, the vehicle having an onboard calculation device adapted to calculate estimated characteristics of at least one external object located in the capture field, including, for each external object, first estimated-position information about the object. A parameterized calculation engine takes as an input the estimated characteristics and provides second corrected-position information about the external object, used to estimate the distance between the motor vehicle and the external object. The values of the parameters of the calculation engine are obtained in a prior learning phase, taking into account, for each external object, the characteristics estimated on the basis of the digital images and reference points obtained by at least one remote detection sensor.