Vehicle Drivable Path Using Obstacle Confidence Scores
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle assistance systems face challenges in accurately determining drivable paths due to false positive detections from sensors, such as cracks and bumps in the road, leading to potential collisions and inefficient route adjustments, especially when driving the same route repeatedly.
Innovation Solution
A computer-implemented method that uses sensor data and a confidence score system to differentiate between overdrivable and non-overdrivable obstacles by storing and updating confidence scores based on previous encounters, allowing the vehicle to learn and adjust its path planning accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor data is used to determine drivable path, then obstacle detection capability is improved, but false positive detections increase
Solution Approach 1:
The system performs preliminary actions by storing sensor data and confidence scores from previous encounters with obstacles at specific geographical locations before making drivable path determinations. When the vehicle approaches a known obstacle location, the system retrieves historical data and uses it to inform the current drivable path calculation, preventing false positive reactions to previously encountered obstacles like road cracks or bumps.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring sensor data, determining confidence scores for detected obstacles, and storing this information for future reference. The confidence score feedback loop allows the system to learn from previous detections and improve its classification accuracy over time, adjusting drivable path determinations based on accumulated experience with specific obstacle types and locations.
2Measurement precision
If sensor data and machine learning models are used to detect objects, then detection accuracy is improved, but processing power requirements increase
Solution Approach 1:
The system applies local quality by focusing computational resources on specific geographical locations where obstacles have been previously detected. Instead of continuously applying heavy machine learning models to all sensor data, the system retrieves and processes only relevant historical data for current locations, reducing overall processing power requirements while maintaining detection accuracy for critical areas.
Solution Approach 2:
The system performs preliminary processing by pre-storing sensor data and confidence scores from previous encounters during less critical times or when computational resources are more available. This preliminary action allows the system to avoid real-time heavy processing during critical drivable path determination, reducing instantaneous processing power requirements while maintaining accuracy.
3Reliability
If the vehicle adjusts drivable path based on perceived obstacles, then collision avoidance is improved, but route efficiency decreases
Solution Approach 1:
The system performs preliminary analysis by retrieving historical obstacle data and confidence scores before making drivable path adjustments. By having this information pre-available from previous encounters, the system can quickly determine whether an obstacle requires path adjustment or can be safely ignored, reducing unnecessary route deviations and improving overall route efficiency while maintaining collision avoidance.
Solution Approach 2:
The system uses feedback from historical confidence scores to inform current drivable path decisions. When confidence scores indicate low risk for previously encountered obstacles, the system can maintain more efficient routes with fewer adjustments. This feedback mechanism allows the system to balance collision avoidance with route efficiency by learning from past experiences with specific obstacle types and locations.
Data Source
Figure 1a
Figure 1b
Figure 2
AI summary
The present disclosure relates to a computer-implemented method for determining a drivable path for a vehicle during a ride, the method comprising the steps of: obtaining sensor data indicating a perceived obstacle associated with a geographical location; retrieving, from a data storage, data associated with the perceived obstacle, the data comprising a confidence score; and if the confidence score is above a threshold, determining the perceived obstacle as overdrivable for the vehicle. In addition, the present disclosure relates to a corresponding apparatus, vehicle and computer program as well as to a system comprising one or more vehicles and a remote storage.