Mobile Object Trajectory Control Using Low-Risk Observation Points
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated driving systems face excessive processing loads, making it difficult to generate target trajectories in time for mobile objects due to exhaustive searches for low-risk points.
Innovation Solution
A mobile object control device and method that sets multiple observation points in the traveling direction, calculates risk grades based on weighted differences between these points, and corrects the target trajectory to minimize risk, allowing for rapid generation of low-risk paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an exhaustive search method is used to find low-risk points for target trajectory generation, then the accuracy of trajectory selection is improved, but the processing load increases excessively
Solution Approach 1:
The patent divides the road region into multiple divided regions and further segments the search process by setting observation points at specific intervals (e.g., every 5m in traveling direction, every 2m in lateral direction). This segmentation reduces the exhaustive search space while maintaining trajectory selection accuracy, directly resolving the contradiction between search accuracy and processing load.
2Measurement precision
If risk calculation is performed for each control cycle at multiple observation points, then the target trajectory can be generated accurately, but the control cycle cannot be speeded up sufficiently
Solution Approach 1:
The patent performs preliminary risk calculations only at specifically set observation points (not all possible points) before control cycle execution. By pre-defining observation points at strategic intervals and calculating risks only at these locations, the system maintains accurate risk evaluation while significantly reducing computation time per control cycle.
Solution Approach 2:
Instead of calculating risk at all possible points on the road, the patent applies partial action by selecting and calculating risk only at a subset of strategically positioned observation points. This partial calculation approach provides sufficient accuracy for trajectory generation while achieving the speedup needed for real-time control cycles.
3Ease of operation
If a sine wave perturbation method is used to shift observation points, then the search for low-risk points is simplified, but the processing load remains excessive for real-time control
Solution Approach 1:
The patent replaces the continuous sine wave perturbation approach with discrete segmentation of observation points at fixed intervals. This segmentation creates a manageable set of candidate points that are simple to process while maintaining the ability to find low-risk trajectories, resolving the contradiction between search simplicity and processing speed.
Data Source
AI summary
Provided is a mobile object control device including: a storage medium having computer-readable instructions stored therein; and a processor connected to the storage medium, wherein the processor executes the computer-readable instructions to recognize an object which is located near a mobile object, set a risk which is an index value indicating a degree to which the mobile object should avoid entry on the basis of a position of the object, and generate a target trajectory for the mobile object to travel along so as to pass through a point at which the risk is low, and generating the target trajectory includes setting a plurality of first observation points at intervals in a traveling direction of the mobile object, setting one or more second observation points in each of a left direction and a right direction as seen from the mobile object for each of the plurality of first observation points, and searching for a point at which the risk is low on the basis of the risk at an observation point group including a first observation point and a second observation point corresponding to each other.