Local-Sensor Lane Change Control Without GPS or HD Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing assisted driving systems face challenges in performing lane changes without relying on costly global localization tools like GPS receivers and high-definition maps, which may not be available in all areas, leading to limitations in driving autonomy.
Innovation Solution
An assisted-driving system that initiates a lane change based on local information by defining a parametric representation of the target lane, receiving local information from perception components, and comparing it to the parametric representation to provide references for actuating the lane change, allowing the vehicle to follow a global path while relying solely on local sensor data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS receiver and high-definition map are used for vehicle localization, then positioning accuracy is improved, but system cost increases
Solution Approach 1:
The patent extracts the essential function of localization from expensive GPS and HD map systems, implementing lane change control using only local sensor data and parametric lane representations. This removes the dependency on costly global positioning infrastructure while maintaining the core functionality of accurate lane positioning and vehicle control.
Solution Approach 2:
The patent replaces expensive, permanent infrastructure (GPS receivers and high-definition maps) with inexpensive, transient local sensor measurements. By using temporary local information from onboard sensors combined with parametric lane models, the system achieves comparable positioning accuracy without requiring costly persistent global localization infrastructure.
2Extent of automation
If high-definition map is used for assisted driving, then driving autonomy is improved, but availability decreases due to limited coverage areas
Solution Approach 1:
The patent segments the global localization problem into local lane-level tasks. Instead of requiring comprehensive global map data and GPS positioning, the system divides the problem into localized lane detection and tracking using onboard sensors. This segmentation enables assisted driving functionality in areas without HD map coverage by focusing computational resources on immediate local environment understanding.
Solution Approach 2:
The patent inverts the traditional approach by not starting with global positioning and then determining local lane context. Instead, it begins with local sensor observations of lane markings and geometry, constructs parametric representations of the target lane, and uses these local models to guide lane changes. This inversion eliminates dependency on global infrastructure availability.
3Device complexity
If parametric representation with local information is used for lane change, then system cost is reduced, but path accuracy may worsen
Solution Approach 1:
The patent performs preliminary construction of parametric representations of the target lane using available local sensor data before executing the lane change maneuver. By pre-defining the geometric model of the target lane and comparing it with real-time sensor observations, the system proactively compensates for potential accuracy losses, ensuring the vehicle follows the intended path even without expensive global positioning infrastructure.
Solution Approach 2:
The patent implements continuous feedback by comparing real-time sensor observations of the target lane with the predefined parametric representation. This feedback loop allows the system to detect deviations from the planned path and make corrective adjustments, maintaining path accuracy despite using cheaper local information sources instead of expensive GPS and HD map systems.
Data Source
AI summary
A computer-implemented method comprises: initiating, by an assisted-driving system that is controlling motion of a vehicle, a lane change of the vehicle from an ego lane to a target lane, the lane change to be performed based only on local information of the vehicle; defining, by the assisted-driving system, a parametric representation of the target lane, the parametric representation based on first waypoints of the target lane; receiving, by the assisted-driving system and during the lane change, a first output of a perception component of the vehicle, the first output being the local information and reflecting second waypoints of the target lane; performing, by the assisted-driving system, a comparison based on the first output and the parametric representation; and providing, by the assisted-driving system, references of a first global path to controllers of the vehicle for actuating the lane change, the references selected based on the comparison.