Occlusion Probability Mapping for Autonomous Vehicle Point Clouds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous ground vehicles face challenges in navigating through environments with occlusions in sensor data, such as dust, precipitation, or terrain, which can obscure obstacles and impassable terrain, leading to gaps in the sensor field of view and lack of explicit information about occluded areas.
Innovation Solution
A method for updating an occlusion probability map using probabilistic models based on sensor data from systems like LiDAR, radar, or cameras, determining cell observation status, and applying an occlusion probability update function to infer occlusions and guide vehicle navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensors are used to detect obstacles and terrain, then obstacle detection capability is improved, but sensor data is obscured by dust, precipitation, or terrain creating occlusions that reduce detection reliability
Solution Approach 1:
The system performs preliminary actions by maintaining and updating an occlusion probability map before navigation decisions are made. The map is continuously updated with occlusion probabilities for different areas, allowing the vehicle to anticipate potential occlusions and plan routes that avoid high-probability occlusion zones, rather than reacting to occlusions after they are encountered
Solution Approach 2:
The occlusion probability map serves as an intermediary data structure that mediates between raw sensor data and navigation decisions. Instead of directly using sensor data to make navigation decisions, the system first processes sensor data to update the occlusion probability map, which then guides path planning and obstacle avoidance behaviors
2Area of stationary object
If sensor field of view is increased to reduce occlusions, then coverage area is improved, but device complexity and cost increase
Solution Approach 1:
The system transitions from a two-dimensional sensor field of view to a three-dimensional occlusion probability map that includes probability values across the coverage area. This dimensional transformation allows the system to represent not just what is visible, but also the likelihood of occlusions in each region, enabling more informed navigation decisions without requiring additional sensors
3Loss of information
If probabilistic modeling is applied to infer occlusions, then information about occluded areas is improved, but computational complexity increases
Solution Approach 1:
The system changes parameters by representing occlusion information as probability values in a structured map format. Each cell in the occlusion probability map contains a probability value indicating the likelihood of occlusion, transforming qualitative occlusion concepts into quantifiable parameters that can be efficiently processed and used for navigation decisions
Data Source
AI summary
Some embodiments of the invention include a method for updating an occlusion probability map. An occlusion probability map represents the probability that a given portion of the sensor field is occluded from one or more sensors. In some embodiments, a method may include receiving field of view data from a sensor system; producing a probabilistic model of the sensor field of view; and updating an occlusion probability map using the probabilistic model and field of view data.