Automated Vehicle Path Planning With Modified Environmental Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Partially automated vehicles require significant resources, including bandwidth and human intervention, to navigate uninterpretable situations, as they struggle to reliably interpret environmental obstacles, often necessitating teleoperation.
Innovation Solution
The vehicle generates a modified environmental model by adjusting object entries to propose paths that initially appear impossible or unsafe, allowing a remote computing unit to select a feasible path, reducing the need for full teleoperation and conserving resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the vehicle computing system generates paths based on the original environmental model including all detected objects, then the path planning is conservative and safe, but the vehicle cannot propose potentially feasible paths that appear impossible due to misinterpretation or uncertainty
Solution Approach 1:
The system changes parameters of the environmental model by generating modified versions with different object interpretations. The vehicle computing system creates multiple environmental models where objects may be interpreted differently (e.g., different object types, positions, or states), allowing path planning to explore a broader range of possible paths while maintaining safety through validation.
2Reliability
If the vehicle requests teleoperation for all uninterpretable situations, then control safety is maintained, but significant bandwidth and human resources are consumed
Solution Approach 1:
Instead of requesting teleoperation for all uncertain situations, the system applies partial action by only seeking human intervention when necessary. The vehicle computing system autonomously handles path planning using modified environmental models, and only requests teleoperation when no feasible path can be determined, thus reducing bandwidth and human resource consumption while maintaining safety.
Solution Approach 2:
The vehicle computing system serves itself by autonomously generating and evaluating multiple path options based on modified environmental models. It can independently resolve many uninterpretable situations without human intervention, only escalating to teleoperation when truly necessary, thereby reducing dependency on human operators and network bandwidth.
3Stability of the object's composition
If the vehicle strictly follows the original environmental model, then interpretation consistency is maintained, but potentially safe paths are rejected as impossible
Solution Approach 1:
The system performs preliminary actions by generating modified environmental models before final path selection. It creates multiple interpretations of the environment in advance, allowing path planning to explore various possibilities, and only commits to a path after validation, thus improving path finding efficiency without compromising model consistency.
Data Source
Figure 1
Figure 2
Figure 2
AI summary
According to a method for controlling an at least partly automated vehicle (1), a vehicle computing system (2) of the vehicle (1) is used to generate an environmental model containing an entry for an object (6) in an environment of the vehicle (1). The vehicle computing system (2) is used to modify the entry for the object (6) to generate a modified environmental model and to generate at least one proposed path (8, 9, 10) for the vehicle (1) based on the modified environmental model and transmit the at least one proposed path (8, 9, 10) to a remote computing unit (11). The vehicle computing system (2) is used to receive information concerning a selected path (8) from the remote computing unit (11) and to control the vehicle (1) to drive along the selected path (8).