Lane Assignment Using Object Trajectories and Radar
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional lane assignment methods in Advanced Driver Assistance Systems (ADAS) are costly, error-prone, and limited by weather conditions and sensor range, often requiring multiple sensors like cameras and radars, which are not reliable in poor visibility and have limited applicability for various ADAS functions.
Innovation Solution
A method that uses trajectories of moving objects' positions over time to assign lanes without the need for separate lane detection, relying on ranging data from systems like radar or lidar, allowing lane assignments for multiple objects and lanes relative to the ego-vehicle, even in adverse weather conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If camera-based lane detection and radar-based object detection are combined for lane assignment, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent combines camera-based lane marking detection with radar-based object detection into a unified lane assignment system. The camera detects lane markings and determines road course, while radar detects object positions. These data streams are merged through coordinate transformation and mapping to achieve accurate lane assignment without requiring separate independent systems.
Solution Approach 2:
The system uses a multi-functional approach where the camera serves both lane detection and road course estimation, while radar provides object detection and positioning. The central processing unit performs multiple functions including coordinate transformation, lane marking extraction, and lane assignment determination, reducing the need for dedicated specialized components.
2Adaptability or versatility
If camera-based road course estimation is used for lane assignment, then adaptability to road conditions is improved, but reliability deteriorates under poor visibility conditions
Solution Approach 1:
The patent introduces an intermediary processing approach where camera-detected lane markings serve as intermediate references for determining road course. This intermediate road course information is then used as a mediator to transform radar object positions into lane assignments, allowing the system to maintain reliability under poor visibility by relying on the intermediate road structure information when available.
Solution Approach 2:
The system prepares for poor visibility conditions by having multiple detection pathways ready. When camera-based lane marking detection is successful, it provides a cushion of reliability. When visibility deteriorates and lane markings become undetectable, the system can switch to alternative methods using pre-established road course models and object trajectory analysis, cushioning against the reliability loss.
3Measurement precision
If separate detection systems for lane position and object position are used, then measurement precision is improved, but loss of time increases due to multiple processing steps
Solution Approach 1:
The system performs preliminary actions by continuously maintaining updated road course models and lane marking detections in advance. The camera continuously tracks lane markings and updates the road course representation, so when objects are detected by radar, the lane assignment can be quickly determined by mapping object positions to the pre-computed lane structures, reducing real-time processing time.
Solution Approach 2:
The patent transforms the problem from separate 2D detections (lane plane and object plane) into a unified 3D spatial mapping problem. By introducing the depth dimension and using coordinate transformations, the system maps radar object positions directly into the camera's lane marking coordinate system, enabling simultaneous determination of both lane position and object position in a unified spatial framework rather than sequential separate detections.
4Measurement precision
If camera and radar systems are deployed for lane assignment, then measurement precision is improved, but cost increases
Solution Approach 1:
The patent merges the functionality of camera and radar systems into a coordinated lane assignment architecture. Rather than having redundant separate systems, the camera and radar are integrated where the camera provides lane structure information and the radar provides object positioning, with a shared processing unit that performs coordinate transformations and unified lane assignment, reducing overall system cost through functional integration.
Data Source
AI summary
A technique for assigning lanes on a road to objects moving in a vicinity of a vehicle on the road is proposed. A method embodiment of the invention comprises the steps of providing trajectories, wherein the or each trajectory represents a time sequence of positions of a moving object; selecting first and second objects and determining a distance between a current position of the first object and the trajectory of the second object; comparing the distance with a predefined threshold; and providing, based on a result of the comparison, a lane assignment indicating a lane to which the second object is assigned.