Vehicle Lane Estimation Using Image and Radar Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object position detection systems, such as those using millimeter wave radar, struggle to accurately recognize vehicle lanes and object positions when the vehicle changes lanes or travels on curved roads, leading to potential erroneous warnings due to reflections from guardrails and incomplete lane recognition.
Innovation Solution
An object recognition apparatus that uses an image sensing device to detect vehicle lanes and stationary objects, combined with a travel history calculation to estimate the position of the vehicle lane and relative object position, allowing for accurate recognition of objects behind the vehicle, including those on oblique rear sides.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If millimeter wave radar is used to detect rear object position, then detection range is improved, but measurement precision deteriorates due to signal reflection from guardrails
Solution Approach 1:
The patent introduces an image sensing device as an intermediary to detect vehicle lanes and stationary objects. This mediator provides visual information that complements the radar data, allowing the system to distinguish between actual objects and false reflections from guardrails, thereby improving measurement precision while maintaining the wide detection range of the radar.
Solution Approach 2:
The patent merges the detection results from multiple sources: millimeter wave radar for wide-area detection, image sensing device for visual confirmation of lanes and objects, and travel history calculation for position estimation. By combining these different detection methods, the system achieves both wide detection range and high measurement precision simultaneously.
2Device complexity
If vehicle lane estimation is based on past travel position, then device complexity is reduced, but measurement precision deteriorates when vehicle changes lanes or travels on curved roads
Solution Approach 1:
The patent uses the image sensing device as an intermediary to directly detect vehicle lanes and stationary objects in the environment. This provides accurate lane position information without relying solely on travel history calculation, enabling precise lane recognition even when the vehicle changes lanes or travels on curved roads, while the system structure remains relatively simple.
3Device complexity
If detection relies only on radar signals, then device complexity is minimized, but reliability deteriorates due to false detections from signal reflection
Solution Approach 1:
The patent merges radar-based detection with image sensing device detection. The radar provides wide-area coverage and the image sensing device provides visual verification. By combining these two detection methods, the system achieves high reliability in object detection while avoiding false detections from signal reflections, without significantly increasing device 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
This solution enables accurate recognition of vehicle lanes and object positions, reducing erroneous warnings and improving safety by accounting for changes in vehicle lane position and curved road conditions.
Implementation Method 1
an image captured by the image sensing device capturing an image of an environment in front of the own vehicle
Implementation Method 2
a signal output from a radio wave radar such as a millimeter wave radar is reflected by the guardrail
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention provides an object recognition apparatus which, in a vehicle that detects an object present behind (including obliquely behind) the vehicle using a radio wave radar, is able to precisely recognize the position of the object present behind (including obliquely behind) the vehicle during traveling on a curve and a lane change, and a vehicle travel controller using the same. An object recognition apparatus is provided with: an image capturing unit (8) which captures an image of an environment in front of a vehicle; a lane detection unit (9) which detects a lane in front of the vehicle on the basis of the image captured by the image capturing unit (8); a lane position estimation unit (4) which estimates the position of a lane behind the vehicle on the basis of the lane detected by the lane detection unit (9) and the travel history of the vehicle; a rear object detection unit (7) which detects an object present behind the vehicle; and a relative position calculation unit (5) which calculates the relative position of the object detected by the rear object detection unit (7) with respect to the position of the lane estimated by the lane position estimation unit (4).