Autonomous Vehicle Trajectory Regions for Admissible Driving Decisions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional autonomous vehicle control systems face computational challenges in determining safe trajectories that achieve driving objectives while avoiding obstacles, as they require extensive online trajectory generation and collision checking, leading to high computational expense.
Innovation Solution
The system employs trajectory-generating functions parameterized by a low-dimensional parameter vector to generate trajectories that achieve driving decisions while avoiding collisions, using achieving and colliding regions to ensure safety and adherence to traffic rules, without separate collision and goal achievability checks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional control systems generate candidate trajectories online and perform collision checking at multiple points along trajectories, then the vehicle can achieve driving objectives while avoiding obstacles, but the computational expense becomes excessively high
Solution Approach 1:
The system pre-computes achieving regions and colliding regions offline before actual driving decisions are needed. These regions are stored and reused during real-time operation, eliminating the need for online trajectory generation and collision checking. This preliminary preparation resolves the contradiction by shifting computational burden from runtime to offline setup.
Solution Approach 2:
Instead of generating actual trajectories and performing detailed collision checks, the system uses simplified region representations (achieving regions and colliding regions) that capture the essential safety and goal-achievability information. These regional copies allow rapid decision-making without the computational cost of full trajectory analysis.
2Reliability
If the system determines all trajectories a vehicle can follow to result in a certain decision while checking for collisions, then complete safety verification is achieved, but the computational complexity becomes unmanageable
Solution Approach 1:
The system transforms the problem from trajectory space to state space by defining achieving regions and colliding regions in the vehicle's state space rather than analyzing individual trajectories. This dimensional transformation reduces computational complexity while maintaining completeness of safety verification through set-based reasoning.
Solution Approach 2:
The achieving regions and colliding regions are pre-computed offline for all possible driving decisions. During real-time operation, the system only needs to check which pre-computed regions contain the current state, eliminating the need for complex online trajectory analysis while ensuring complete safety verification.
Data Source
AI summary
A control system for controlling a motion of an ego-vehicle traveling to a target destination is provided. The control system includes a memory and a processor to execute instruction stored by the memory. The memory stores multiple trajectory-generating functions corresponding to a maneuver defined by a parameter vector associated with a driving decision. The parameter vector is defined by one or multiple parameters. Each of the multiple trajectory-generating functions is configured to generate an achieving sequence of regions of states and values of the parameter vector reaching an input target region within a prediction horizon. The stored instructions cause the control system to test control admissibility of at least some of the driving decisions consistent with the target destination of the ego-vehicle at a current state. The stored instruction also caused the control system to control the ego-vehicle according to one of the admissible driving decisions.