Autonomous Vehicle Driving Environment Sub-Environment Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in emulating the human decision-making process for navigating through complex driving environments, as they struggle to consolidate perceptions of multiple objects and scenarios into coherent driving decisions.
Innovation Solution
The system constructs a driving environment divided into sub-environments, using sensor data to identify and handle specific scenarios, with dedicated driving scenario handlers determining driving conditions and consolidating decisions for safe and appropriate vehicle control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If autonomous vehicles use a comprehensive perception system to detect all objects in the driving environment, then the completeness of environmental understanding is improved, but the system complexity and computational burden increase significantly
Solution Approach 1:
The patent divides the complex driving environment into multiple sub-environments (e.g., road environment, pedestrian environment, vehicle environment). Each sub-environment is processed by dedicated scenario handlers that focus on specific aspects, reducing the computational burden on any single system component while maintaining comprehensive environmental understanding.
Solution Approach 2:
The patent introduces scenario handlers as intermediary components between the sensor data and the decision-making system. These handlers act as mediators that preprocess and organize information from multiple sensors, filtering and structuring data before it reaches the core decision-making algorithms, thereby reducing overall system complexity.
2Speed
If autonomous vehicles process all sensor data in real-time to make instant driving decisions, then the responsiveness of decision-making is improved, but the computational load and processing time increase
Solution Approach 1:
The patent segments the decision-making process into multiple parallel streams, with each scenario handler processing specific types of scenarios independently. This allows simultaneous processing of different environmental aspects without sequential bottlenecks, improving overall responsiveness while distributing computational load across multiple specialized modules.
Solution Approach 2:
The patent implements selective processing where scenario handlers only activate and process data relevant to current driving scenarios. Not all handlers are active simultaneously, and each handler processes only the subset of sensor data pertinent to its specific scenario type, reducing overall computational load while maintaining real-time responsiveness for critical decisions.
3Adaptability or versatility
If autonomous vehicles implement a human-like synthesized decision-making process that consolidates individual object decisions, then the naturalness of driving behavior is improved, but the complexity of the decision-making system increases
Solution Approach 1:
The patent breaks down the synthesized decision-making process into discrete scenario handlers, each responsible for specific driving scenarios. This segmentation allows the system to emulate human-like decision-making through modular components rather than a monolithic complex system, where each handler contributes to the overall natural driving behavior in its specific domain.
Solution Approach 2:
The patent designs scenario handlers with universal interfaces and standardized processing pipelines that can handle multiple types of scenarios through a common framework. This multi-functionality allows the same basic architecture to serve different driving scenarios, reducing overall system complexity while maintaining adaptable, human-like driving behavior across diverse situations.
Data Source
AI summary
In response to sensor data received from sensors mounted on an autonomous vehicle, a surrounding environment is perceived based on the sensor data. The surrounding environment includes multiple sub-environments. For each of the sub-environments, one of a plurality of driving scenario handlers associated with the sub-environment is identified, each driving scenario handler corresponding to one of a plurality of driving scenarios. The identified driving scenario handler is invoked to determine an individual driving condition within the corresponding sub-environment. An overall driving condition for the surrounding environment is determined based on the individual driving conditions provided by the identified driving scenario handlers. A route segment is planned based on the overall driving condition of the surrounding environment, the route segment being one of a plurality of route segments associated with a route. The autonomous vehicle is controlled and driven based on the planned route segment.