Vehicle Position Estimation Using Landmark Orientation Differences
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In areas with fewer landmarks, such as suburban roads, the number of detectable landmarks is small, leading to challenges in vehicle position estimation, especially when using low-cost external sensors which reduce measurable distance and detection angle.
Innovation Solution
A position estimation device that acquires a predicted position and orientation of a moving body, calculates orientations based on measurement and map information, and corrects these values using a difference between measured and mapped orientations, employing a Kalman filter to accurately estimate position and orientation even with limited detectable landmarks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a low-cost external sensor is mounted, then device cost is reduced, but measurable distance and detection angle become limited
Solution Approach 1:
The patent changes the parameters used for position estimation from relying solely on distance measurements to incorporating orientation angle measurements. By using the difference between measured and map-based orientations of landmarks, the system achieves accurate position estimation even with limited detection range, effectively compensating for the reduced measurable distance of low-cost sensors.
2Ease of manufacture
If a low-cost external sensor is mounted, then device cost is reduced, but detection angle becomes narrow
Solution Approach 1:
The patent supplements the limited detection angle information with orientation data from map information. By calculating the difference between the measured orientation of a landmark and its orientation in the map, the system obtains additional angular information that expands the effective detection coverage, allowing accurate position estimation despite the narrow physical detection angle of the sensor.
3Reliability
If the number of detectable landmarks is small, then position estimation becomes difficult, but this is caused by limited sensor performance in low-landmark areas
Solution Approach 1:
The patent introduces map information as an intermediary that provides expected orientation data for landmarks. By comparing the actual measured orientation with the expected orientation from the map, the system creates a correction term that improves position estimation accuracy. This intermediary approach allows reliable position estimation even when few landmarks are detectable, as the map information provides additional constraints to resolve the ambiguity.
4Ease of operation
If only one landmark is detected, then position estimation can still be performed, but estimation accuracy deteriorates
Solution Approach 1:
The patent implements a feedback mechanism where the measured orientation of the detected landmark is compared with its expected orientation from map information. The difference between these orientations provides a correction that feeds back into the position estimation, improving accuracy even when only one landmark is detected. This feedback loop allows the system to maintain high estimation accuracy with minimal landmark detections.
Data Source
AI summary
The own vehicle position estimator 17 of the in-vehicle device 1 acquires the predicted own vehicle position X− (t) that is a predicted value of the present position and the orientation of the moving body. Then, the own vehicle position estimator 17 acquires: the relative angle Lψ (t) that is first orientation information indicative of the orientation of the landmark with respect to the vehicle specified based on the measurement data of the landmark by the lidar 2; and the relative angle L−ψ (t) that is second orientation information indicative of the orientation of the landmark with respect to the vehicle into which the orientation of the landmark indicated by the landmark information is converted. The own vehicle position estimator 17 corrects the predicted own vehicle position X−(t) based on the difference between the relative angle Lψ(t) and the relative angle L−ψ(t).