Autonomous Vehicle Trajectory Ranking for Safer Path Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating safely and accurately on roadways due to the need to process various types of data, including visual information, radar, and GPS data, while identifying location, avoiding obstacles, and responding to traffic signals and signs, without effective integration of these data sources.
Innovation Solution
A system comprising cameras, processing units, and sensors that analyze visual information, GPS data, and sensor data to provide navigational responses, using image processing and machine learning algorithms to detect lane markings, vehicles, pedestrians, and traffic signals, and make decisions on steering, braking, and acceleration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple data sources (visual information, radar, GPS) are processed simultaneously to improve navigation accuracy, then the reliability of autonomous vehicle navigation is improved, but the device complexity increases
Solution Approach 1:
The system segments navigation data processing into multiple specialized modules: visual information processing module, radar data processing module, GPS data processing module, and trajectory generation module. Each module handles specific data types independently, improving reliability through specialized processing while managing complexity through modular architecture.
Solution Approach 2:
The system merges multiple data sources (cameras, radar, GPS) into a unified trajectory generation system that processes all inputs simultaneously. This integration allows the system to cross-validate information from different sensors, improving navigation safety and reliability while making decisions based on comprehensive environmental understanding.
2Measurement precision
If multiple sensors and data sources are integrated to improve obstacle detection accuracy, then the measurement precision is improved, but the device complexity increases
Solution Approach 1:
The system segments obstacle detection into specialized processing paths: visual obstacle detection using camera data, radar-based obstacle detection, and GPS-based location verification. Each sensor type processes specific aspects of obstacle detection, improving overall accuracy through multi-modal verification while managing complexity through dedicated processing modules.
Solution Approach 2:
The system implements feedback loops where detected obstacles from one sensor type validate and refine detections from other sensors. For example, radar-detected obstacles are cross-checked with visual data from cameras, and GPS location data provides contextual verification. This feedback mechanism improves measurement precision by eliminating false positives and confirming true obstacles through multiple independent verification channels.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
A navigation system for a host vehicle may include at least one processor programmed to receive, from a camera, a plurality of images representative of an environment of the host vehicle. The processor may also be programmed to analyze at least one of the plurality of images to identify navigational state information associated with the host vehicle; determine a plurality of potential trajectories for the host vehicle based on the navigational state information; perform a preliminary analysis relative to each of the plurality of potential trajectories and assign to each of the plurality of potential trajectories, based on the preliminary analysis, at least one indicator of relative ranking; select, based on the at least one indicator of relative ranking assigned to each of the plurality of potential trajectories, a subset of the plurality of potential trajectories, wherein the subset of the plurality of potential trajectories includes fewer potential trajectories than the plurality of potential trajectories; perform a secondary analysis relative to the subset of the plurality of potential trajectories, and based on the secondary analysis, select one of the subset of the plurality of potential trajectories as a planned trajectory for the host vehicle; determine one or more navigational actions for the host vehicle based on the planned trajectory selected from among the subset of the plurality of potential trajectories; and cause at least one adjustment of a navigational actuator of the host vehicle to implement the one or more navigational actions for the host vehicle.