Intersection Occlusion Handling with Phantom Obstacles for AVs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles (AVs) face challenges in navigating intersections due to occlusions, which can limit their ability to detect all objects approaching the intersection, leading to potential safety issues.
Innovation Solution
The system determines whether occluded objects are approaching an intersection by analyzing sensor data to identify occlusions. If an occlusion is detected, the system generates a phantom obstacle behind the occlusion and predicts its speed to determine if it will reach the conflict region before the vehicle. Based on this analysis, the system decides whether to proceed through the intersection or take a safety-related action.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the AV stops or requires human takeover when approaching an intersection with occlusions, then safety is improved, but productivity and operational efficiency deteriorate
Solution Approach 1:
The system performs preliminary actions by generating phantom obstacles before the AV actually encounters potential hazards. The visibility grid analysis and phantom obstacle generation occur in advance, allowing the system to prepare safety measures without requiring immediate stops or human intervention, thus maintaining both safety and operational efficiency
Solution Approach 2:
The phantom obstacle acts as an intermediary between the occluded real obstacle and the AV's decision-making system. By creating a virtual representation of potential hazards, the system can evaluate risks and make informed decisions without needing direct sensor contact with actual obstacles, enabling continuous operation while maintaining safety
2Measurement precision
If the AV uses multiple sensors to detect objects, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The visibility grid serves as an intermediary data structure that consolidates information from multiple sensors. Instead of processing raw data from numerous sensors separately, the system integrates sensor inputs into a unified visibility grid representation, simplifying the processing architecture while maintaining comprehensive object detection capabilities
Solution Approach 2:
The system segments the environment into discrete visibility grid cells, allowing independent analysis of each region. This segmentation approach enables the system to handle complex multi-sensor data by breaking it down into manageable units, improving measurement precision without proportionally increasing system complexity
3Reliability
If the AV generates phantom obstacles for all occluded regions, then reliability is improved, but device complexity and computational load increase
Solution Approach 1:
The system applies different processing qualities to different regions of the visibility grid. Phantom obstacles are generated selectively based on local conditions - only in occluded regions where they are actually needed for safety, rather than uniformly across all areas. This localized approach maintains reliability while reducing unnecessary computational complexity
Solution Approach 2:
The system changes parameters dynamically based on the visibility grid analysis. Phantom obstacles are generated with specific parameters (position, speed, trajectory) that match the local occlusion characteristics. This parameter adaptation allows the system to maintain high reliability with computationally efficient, context-appropriate phantom obstacle generation
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An autonomous vehicle navigates an intersection in which occlusions block the vehicle's ability to detect moving objects. The vehicle handles this by generating a phantom obstacle behind the occlusion. The vehicle will predict the speed of the phantom obstacle and use the predicted speed to assess whether the phantom obstacle may collide with the vehicle. If a collision is a risk, the vehicle will slow or stop until it confirms that either (a) the phantom obstacle is not a real obstacle or (b) the vehicle can proceed at a speed that avoids the collision. To determine which occlusions shield real objects, the system may use a rasterized visibility grid of the area to identify occlusions that may accommodate the object.