Vehicle Localization Using Camera, Map, and Velocity Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle detection systems require sophisticated sensors that are not readily available, limiting their applicability to high-end products and are not suitable for broad usage, especially for autonomous or partially autonomous mobile robots operating in pedestrian environments.
Innovation Solution
A method that combines image data from cameras with additional data such as map data and velocity sensor data, like Doppler radar, to determine the location of vehicles efficiently and accurately, using intersections between roads and image-derived directions, and velocity hypotheses to resolve ambiguities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sophisticated sensors (e.g., LIDAR, multiple distance measurement devices) are used for vehicle detection and localization, then measurement precision and reliability are improved, but device complexity and cost increase significantly
Solution Approach 1:
The patent combines image data from a camera with additional data (map data, velocity sensor data) to determine vehicle location. This merging of multiple data sources compensates for the lack of sophisticated distance measurement sensors, achieving accurate localization without requiring complex sensor systems like LIDAR or multiple distance measurement devices.
Solution Approach 2:
The patent introduces map data as an intermediary element to resolve the location determination problem. By comparing image-derived directions with pre-stored map data, the system can determine vehicle location without direct distance measurements, thus avoiding the need for complex sensing hardware.
2Reliability
If sophisticated sensors and complex processing algorithms are deployed, then vehicle detection reliability is improved, but ease of manufacture and deployment deteriorates
Solution Approach 1:
The patent makes the vehicle detection system universal by using a camera (a common, inexpensive component) combined with map data and velocity sensor data. This multi-functional approach allows the same system to be deployed in various mobile robots with different capabilities, enhancing ease of manufacture and deployment while maintaining reliability through data fusion.
Solution Approach 2:
The patent replaces expensive, complex sensors with cheaper alternatives: a standard camera instead of LIDAR, and uses readily available map data and velocity sensor data. This substitution of inexpensive components achieves reliable vehicle detection without the high cost and complexity of sophisticated sensing systems.
3Measurement precision
If distance measurements are obtained using sophisticated sensors, then location determination accuracy is improved, but loss of time in data acquisition increases
Solution Approach 1:
The patent performs preliminary actions by pre-storing map data and using velocity sensor data to predict vehicle position. This allows the system to quickly determine location by comparing current image data with pre-prepared information, reducing the time needed for data acquisition while maintaining accuracy without relying on slow distance measurement sensors.
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
Enables fast, efficient, and accurate vehicle localization for mobile robots, even without distance measurements, making it suitable for low-cost, low-speed robots with enhanced safety and fail-safety.
Implementation Method 1
processing image data to determine a direction between a camera capturing an image and the object
Implementation Method 2
velocity sensor data, like Doppler radar
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates to a method for determining a location of an object (300), the method comprising processing image data to determine a direction between a camera capturing an image and the object (300); processing additional data comprising at least one of map data and velocity sensor data; and combining information based on the image data and the additional data to arrive at a location of the object (300). The present invention also relates to a corresponding robot configured to carry out such a method.