Continuing Lane Driving Prediction Using Roadgraph and ML Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in predicting and responding to atypical behaviors of road users, such as jaywalking pedestrians or vehicles running red lights, particularly when determining the likelihood and geometry of continuing lane driving, which can lead to rapid acceleration or braking maneuvers.
Innovation Solution
The use of machine learning models to identify and predict continuing lane driving behaviors by analyzing sensor data from various sources, including map information and agent features, to generate accurate behavior predictions and control vehicle operations accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional detection and prediction methods are used for road users, then the system is simpler to implement, but the precision and recall of behavior predictions deteriorate, leading to inability to detect atypical behaviors like continuing lane driving
Solution Approach 1:
The prediction system is segmented into multiple specialized machine learning models: a first trained model identifies candidate continuing lanes, while a second trained model generates behavior predictions. This segmentation allows each model to focus on specific aspects of atypical behavior detection, improving overall precision without requiring a single overly complex system
Solution Approach 2:
The system performs preliminary identification of candidate continuing lanes using map information and lane geometry analysis before generating behavior predictions. This preliminary action filters the search space and prepares structured data that enhances the accuracy of subsequent behavior prediction models
2Reliability
If the system predicts atypical behaviors like continuing lane driving, then safety improves by avoiding sudden maneuvers, but the difficulty of detecting and measuring such behaviors increases
Solution Approach 1:
The system detects continuing lane driving by analyzing lane geometry and spatial relationships between source lanes and candidate continuing lanes. This geometric dimension provides an additional detection avenue beyond traditional behavior pattern recognition, making it easier to identify atypical behaviors that violate expected driving patterns
Solution Approach 2:
Map information and lane geometry data serve as intermediaries that bridge the gap between observable vehicle positions and unobservable driver intentions. These intermediaries provide structured contextual information that helps the prediction models infer atypical behaviors more easily
3Measurement precision
If the system uses multiple trained machine learning models for prediction, then the accuracy of behavior predictions improves, but the computing resources and processing time increase
Solution Approach 1:
The computing workload is segmented across two specialized models rather than one monolithic model. The first model handles candidate lane identification using map data, while the second model focuses on behavior prediction using sensor data. This segmentation allows for more efficient resource utilization compared to a single comprehensive model
Solution Approach 2:
The system performs preliminary filtering and candidate selection using the first model before engaging the more computationally intensive second model for behavior prediction. This preliminary action reduces the input complexity for the second model, thereby reducing its computational energy requirements
Data Source
AI summary
The technology relates to controlling a vehicle in an autonomous driving mode in accordance with behavior predictions for other road users in the vehicle's vicinity. In particular, the vehicle's onboard computing system may predict whether another road user will perform a “continuing” lane driving operation, such as going straight in a turn-only lane. Sensor data from detected/observed objects in the vehicle's nearby environment may be evaluated in view of one or more possible behaviors for different types of objects. In addition, roadway features, in particular whether lane segments are connected in a roadgraph, are also evaluated to determine probabilities of whether other road users may make an improper continuing lane driving operation. This is used to generate more accurate behavior predictions, which the vehicle can use to take alternative (e.g., corrective) driving actions.