Occlusion Prediction Using Shared Temporal Maps for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating areas where objects are outside their sensor's observable or perceivable range, leading to reduced navigation confidence and impaired ability to predict behaviors of occluded objects.
Innovation Solution
Autonomous vehicles share and incorporate rich object information from other vehicles within a fleet through vehicle-to-vehicle communication and a remote joint map management system, using encoders to compress and timestamp data for efficient transfer and integration into a temporal map, enhancing perception, understanding, and prediction capabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If autonomous vehicles rely solely on their own sensors to detect objects, then the vehicle's navigation system remains simple, but the vehicle cannot perceive objects outside its sensor's observable range, reducing navigation confidence
Solution Approach 1:
The patent combines sensor data from multiple autonomous vehicles to create a composite perception of the environment. Each vehicle's sensor observations are merged into a shared understanding that allows any individual vehicle to infer the presence and state of objects outside its direct sensor range, thereby improving navigation confidence without requiring each vehicle to have omnidirectional sensors.
Solution Approach 2:
The patent introduces a communication system as an intermediary that transfers object information between vehicles. This mediator enables vehicles to share detections of occluded objects, allowing a vehicle to know about objects beyond its sensor range by receiving information from other vehicles that have direct line-of-sight to those objects.
2Reliability
If autonomous vehicles share detailed object information with other vehicles, then prediction capability for occluded objects improves, but data transmission time and communication bandwidth increase
Solution Approach 1:
The patent extracts only the essential object information needed for prediction from the full sensor data set. Instead of transmitting complete point clouds or high-resolution images, the system extracts key attributes such as object position, velocity, acceleration, and trajectory predictions, which are then shared with other vehicles. This extraction maintains prediction capability while dramatically reducing communication overhead.
Solution Approach 2:
The patent transforms raw sensor data into a different parameter representation that is more suitable for communication and prediction. By converting sensor observations into predicted trajectory parameters and occlusion state indicators, the system enables efficient data exchange that preserves the information needed for accurate prediction while minimizing transmission time.
3Loss of information
If autonomous vehicles incorporate object information from multiple sources, then understanding of occluded objects improves, but the complexity of integrating and managing data from multiple vehicles increases
Solution Approach 1:
The patent implements a universal data format and communication protocol that can accommodate information from multiple vehicle sources. This multi-functional framework allows the system to integrate data from any number of vehicles using a standardized interface, reducing the complexity of data management by providing a unified approach to handling heterogeneous sensor inputs from multiple sources.
Data Source
AI summary
Autonomous vehicles (AVs) utilize perception and understanding of objects on the road to predict behaviors of the objects, and to plan a trajectory for the vehicle. In some scenarios, an AV may benefit from having information about an object that is not within a perceivable area of the sensors of the AV. Another AV in the area that has the object within a perceivable area of the sensors of the other AV may share information with the AV. Rich information about the object determined by the other AV can be compressed using an encoder and transferred efficiently to the AV for inclusion in a temporal map that combines locally determined object information and transferred object information determined by other AV(s). The temporal map may be used in one or more parts of the software stack of the AV.