Vehicle Spatial Occupancy Prediction for Complex Maneuvers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing prediction techniques for autonomous vehicles are inadequate for detecting and safely maneuvering around vehicles performing complex, longer-term maneuvers, as they rely on timing-based predictions that are insufficient for accurate behavior prediction.
Innovation Solution
Training machine-learning models to predict regions of space likely to be occupied by vehicles while they perform complex maneuvers, allowing autonomous vehicles to safely navigate and avoid these regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If timing-based prediction techniques are used for autonomous vehicles, then the system complexity is reduced, but the prediction accuracy deteriorates for complex maneuvers
Solution Approach 1:
The patent changes the prediction parameter from time-based to spatial-based. Instead of predicting when a vehicle will arrive at a location, the system predicts the spatial region (occupancy grid) that a vehicle is likely to occupy during complex maneuvers. This parameter transformation enables accurate prediction of vehicle positions throughout the maneuver duration, resolving the accuracy issue while maintaining reasonable system complexity through grid-based representation.
2Measurement precision
If spatial prediction models are trained to predict occupied regions, then the prediction accuracy improves, but the device complexity increases
Solution Approach 1:
The patent segments the spatial environment into discrete grid cells forming an occupancy grid. This segmentation transforms the continuous spatial prediction problem into a discrete classification problem where each grid cell is independently predicted. The segmentation simplifies the model training by breaking down complex spatial relationships into manageable discrete units, reducing the overall device complexity while maintaining high prediction accuracy.
Solution Approach 2:
The patent introduces a spatial dimension to the prediction problem by predicting occupancy across a 2D grid space rather than single-point temporal predictions. This dimensional transformation allows the model to capture the full spatial extent of vehicle maneuvers, improving accuracy while the grid structure provides a regularized framework that manages computational complexity.
3Speed
If traditional time-based prediction is used, then the processing speed is maintained, but the ability to detect complex maneuvers deteriorates
Solution Approach 1:
The patent performs preliminary spatial mapping by creating an occupancy grid that predicts all possible vehicle positions throughout the maneuver duration in advance. This preliminary spatial action allows the system to proactively identify regions that may be occupied during complex maneuvers, enabling faster real-time decision-making by eliminating the need for repeated temporal predictions and improving reliability of complex maneuver detection.
Data Source
AI summary
Techniques relating to determining regions based on intents of objects are described. In an example, a computing device onboard a first vehicle can receive sensor data associated with an environment of the first vehicle. The computing device can determine, based on the sensor data, a region associated with a second vehicle proximate the first vehicle that is to be occupied by the second vehicle while the vehicle performs a maneuver. Further, the computing device can determine an instruction for controlling the first vehicle based at least in part on the region.