Implicit Occupancy Motion Planning for Real-Time Collision Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous systems face challenges in accurately and timely predicting future object occupancy in dynamic environments, leading to potential collisions, especially when traditional object-based and grid-based methods are resource-intensive and inflexible.
Innovation Solution
Implementing an implicit occupancy approach using an implicit decoder model to predict geographic location occupancy without building a full occupancy grid, by discretizing trajectories into query points and generating point attributes for safe trajectory selection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an object-based system is used to detect and predict object positions, then prediction accuracy is improved, but computational resources and processing time are increased
Solution Approach 1:
The patent extracts only the essential occupancy information needed for trajectory planning rather than processing complete object models and trajectories. By querying only points along the planned trajectory and retrieving minimal occupancy attributes, the system achieves accurate collision prediction while dramatically reducing computational overhead compared to full object-based prediction systems.
Solution Approach 2:
The system performs partial occupancy prediction only at query points along the trajectory rather than predicting occupancy for the entire geographic region. This selective approach provides sufficient information for safe navigation while avoiding the excessive computational resources required for complete region-wide prediction.
2Reliability
If a grid-based approach is used to predict occupancy for each cell, then collision avoidance is improved, but device complexity and resource consumption are increased
Solution Approach 1:
The patent extracts occupancy information only at specific query points along the trajectory rather than maintaining and processing a complete occupancy grid for the entire region. This eliminates the complexity of grid management while retaining the essential collision avoidance functionality by querying occupancy only where needed for safety decisions.
Solution Approach 2:
The system performs occupancy prediction partially by querying only the necessary points along the trajectory rather than computing occupancy for all grid cells. This provides sufficient reliability for collision avoidance while dramatically reducing the device complexity associated with maintaining and processing complete occupancy grids.
3Measurement precision
If a complete occupancy grid is constructed for prediction, then measurement precision is improved, but use of energy and computational resources are increased
Solution Approach 1:
The patent extracts only the occupancy attributes needed for trajectory safety at query points along the path rather than constructing and processing a complete occupancy grid. This extraction approach maintains prediction accuracy for collision avoidance while minimizing computational energy consumption by avoiding unnecessary processing of unvisited regions.
Solution Approach 2:
The system performs occupancy prediction partially by querying only the essential points along the trajectory rather than computing occupancy for the entire grid. This provides sufficient measurement precision for safe navigation while dramatically reducing the computational energy required compared to complete grid construction and processing.
4Device complexity
If fixed resolution grid cells are used for prediction, then device complexity is reduced, but adaptability to different scenarios is worsened
Solution Approach 1:
The patent implements dynamic query point generation that adapts to different trajectories and scenarios. Rather than using a fixed resolution grid, the system dynamically determines query points based on the specific trajectory being evaluated, allowing the prediction system to adapt to various scenarios (different speeds, directions, environments) while maintaining simple query-based architecture that avoids grid complexity.
Data Source
AI summary
Motion planning with implicit occupancy for autonomous systems includes obtaining a set of trajectories through a geographic region for an autonomous system, and generating, for each trajectory in the set of trajectories, a set of points of interest in the geographic region to obtains sets of points of interest. Motion planning further includes quantizing the sets of points of interest to obtain a set of query points in the geographic region and querying the implicit decoder model with the set of query points to obtain point attributes for the set of query points. Motion planning further includes processing, for each trajectory of a least a subset of trajectories, the point attributes corresponding to the set of points of interest to obtain a trajectory cost for the trajectory. From the set of trajectories, a selected trajectory is selected according to trajectory cost.