Virtual Obstacle Trajectory Planning for Sensor Occlusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle automation systems face challenges in determining safe trajectories when sensor limitations prevent complete detection of the surrounding environment, leading to potential safety risks during automated driving.
Innovation Solution
A method that detects observable obstacles, determines unobservable areas, adds virtual obstacles with associated hypothetical events and probability assignments, and calculates safe vehicle trajectories to account for both real and virtual obstacles, ensuring precautionary measures to prevent dangerous situations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the automation system uses sensors to monitor the surrounding environment, then the vehicle can detect observable obstacles and plan trajectories, but the sensors have inherent limitations and cannot detect everything in the environment
Solution Approach 1:
The system performs preliminary action by proactively placing virtual candidate objects in blocked areas before any collision can occur. When the sensor detects an occluded area, the system immediately generates virtual obstacle candidates in those regions, allowing the trajectory planning to account for potential hazards that are not yet visible to the sensor.
Solution Approach 2:
The system introduces virtual candidate objects as intermediaries between the sensor's limited detection capability and the trajectory planning system. These virtual objects serve as mediators that represent potential hazards in blocked areas, allowing the planner to make informed decisions without directly observing the actual obstacles.
2Reliability
If the system places virtual candidate objects in all blocked areas to account for potential obstacles, then collision avoidance is improved, but the complexity of trajectory planning increases
Solution Approach 1:
The system applies dynamics by making the set of candidate objects dynamic rather than static. The virtual candidate objects are continuously updated based on sensor feedback - when the sensor observes a blocked area, candidate objects are added; when the area becomes observable and no obstacle is detected, the candidates are removed. This dynamic adaptation reduces unnecessary complexity while maintaining safety.
Solution Approach 2:
The system changes parameters by adjusting the set of candidate objects based on sensor observations. The presence or absence of candidate objects in blocked areas is modified as a parameter, allowing the trajectory planning to adapt to current environmental conditions without requiring a completely complex replanning process.
3Reliability
If the system assumes obstacles may be present in blocked areas and places virtual candidates, then safety margins are improved, but the vehicle's speed and productivity are reduced
Solution Approach 1:
The system applies partial action by placing virtual candidate objects only in blocked areas where obstacles might exist, rather than assuming obstacles everywhere. This selective approach provides sufficient safety margins in critical areas while allowing the vehicle to maintain higher speeds in areas where the sensor can directly observe the environment and no candidates are present.
Solution Approach 2:
The system uses the sensor's own observations to determine where safety margins are needed. When the sensor can directly observe an area, no virtual candidates are placed and the vehicle can proceed at normal speed. When the sensor cannot observe an area (blocked area), the system automatically generates candidates to provide safety margins, making the safety mechanism self-regulating based on actual sensor capabilities.
Data Source
Figure 1a~1b
Figure 2a
Figure 2b~2c
AI summary
The present disclosure relates to a method and arrangement for determining safe vehicle trajectories for a vehicle (1) equipped with sensors (2) for monitoring the surrounding environment, taking into account sensing limitations, as well as a vehicle (1) comprising such an arrangement. The method comprises the steps of: detecting observable obstacles (3); detecting unobservable areas (4); adding virtual obstacles (5) in unobservable areas (4); associating each observable obstacle (3) and each virtual obstacle (5) with one or more hypothetical events and assigning an occurrence probability to each combination of obstacle (3, 5) and events; and determining safe vehicle (1) trajectories based on both observable obstacles (3) and virtual obstacles (5) and the occurrence probability of each combination of obstacle (3, 5) and events.