Mobile Body Path Control Using Split Costmaps for Stable Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Path generation based on costmaps that consider object movement can compromise the safety and comfort of mobile bodies, such as vehicles, due to unstable movement predictions.
Innovation Solution
A control system that generates a first costmap based on detected object positions and a second costmap based on predicted object movements, using these to stabilize path generation and incorporate stopping or deceleration when necessary, while also considering object attributes and lane predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a path is generated based on a costmap that considers predicted movement of objects, then the path may avoid future collision risks, but the path becomes unstable and compromises safety and comfort due to prediction uncertainties
Solution Approach 1:
The costmap is segmented into multiple layers: a first costmap layer representing current object positions and a second costmap layer representing predicted future positions. This segmentation allows the path planning system to separately handle static and dynamic cost components, reducing the impact of prediction uncertainties on overall path stability while maintaining collision avoidance capabilities.
Solution Approach 2:
The system dynamically adjusts the weight or influence of the second costmap layer based on prediction confidence levels. When movement predictions are highly uncertain, the system reduces their impact on path generation, thereby maintaining path stability. When predictions are confident, the system leverages them for better collision avoidance, achieving adaptive balance between stability and safety.
2Reliability
If movement predictions are reflected onto the costmap to avoid future collisions, then collision risk decreases, but path comfort deteriorates due to frequent path adjustments
Solution Approach 1:
The system performs preliminary path planning using the first costmap layer (current positions) to establish a stable base path. The second costmap layer (predicted positions) is then used to make incremental adjustments only where necessary, rather than completely re-planning the path. This preliminary action approach maintains comfort by minimizing frequent full path recalculations while still achieving collision avoidance.
Solution Approach 2:
The system applies movement prediction adjustments locally only in regions where predicted object movements create potential collision risks, rather than globally adjusting the entire path. This localized adjustment maintains path comfort in stable regions while ensuring safety in dynamic risk zones, reducing unnecessary path fluctuations that would compromise comfort.
Data Source
Figure 1A
Figure 1B
Figure 2
AI summary
A control system (10) includes an obtainer (11) that obtains sensing data output from a sensor that performs sensing of an outside of a mobile body, a detector (12) that detects a position of an object outside the mobile body based on the sensing data, a movement predictor (13) that predicts a movement of the object based on the sensing data, a costmap generator (14) that generates a first costmap based on the position detected of the object and a second costmap based on the movement predicted of the object, a path generator (15) that generates a path for the mobile body based on the first costmap, a determination generator (16) that generates a movement determination of the mobile body based on the second costmap and the path generated, and a controller (17) that controls the movement of the mobile body in accordance with the path generated and the movement determination.