Autonomous Parking Obstacle Avoidance via Reverse Planning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for autonomous driving systems fail to handle complex obstacle situations with changing boundary conditions without interruption or manual intervention, particularly when encountering stationary or moving obstacles.

Innovation Solution

A method for fully or partially autonomous driving that involves detecting obstacles using sensors, reversing to a previously traveled waypoint, and then navigating past the obstacle on an avoidance trajectory, which reflects the dimensions and direction of the obstacle in both forward and reverse movements, allowing for flexible and safe path adjustments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a trained reference trajectory is used for autonomous parking, then the driving process can be executed quickly and simply, but the system fails when unexpected obstacles require deviations from the trained driving style

Engineering Contradiction:
Improvespeed of autonomous parkingVSAvoidability to handle unexpected obstacles
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The system performs preliminary actions by reversing the vehicle to a previously traveled waypoint before executing the avoidance maneuver. This preliminary repositioning allows the vehicle to approach the obstacle from a different spatial and temporal perspective, enabling the trained avoidance trajectory to be successfully executed even when obstacles appear unexpectedly during autonomous parking

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system inverts the normal forward-only trajectory execution by allowing reverse movement to a previous waypoint. This inversion of the temporal sequence (going backward in time/space along the trajectory) transforms a static trained path into a dynamic adaptive path that can handle obstacles by revisiting and re-evaluating previous positions

Inventive Principle:
Principle #13The other way round (Inversion)

2Reliability

If the vehicle reverses to a previous waypoint to avoid an obstacle, then collision-free paths can be achieved, but the driving time increases due to the additional reverse and forward movement

Engineering Contradiction:
Improvecollision-free pathVSAvoidadditional reversal time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system applies partial action by reversing only to a previous waypoint rather than completely aborting the parking maneuver or reversing to the very start. This partial reversal minimizes the additional time required while still achieving sufficient spatial separation from the obstacle to execute a safe avoidance trajectory

Inventive Principle:
Principle #16Partial or excessive action

3Measurement precision

If the starting point of the avoidance trajectory is selected opposite to the previous direction of movement, then the obstacle dimensions and direction can be better reflected, but the path planning complexity increases

Engineering Contradiction:
Improveobstacle detection accuracyVSAvoidpath planning complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system uses feedback by selecting the starting point of the avoidance trajectory based on the detected obstacle's dimensions and direction of movement. The trajectory planning incorporates real-time sensor data about the obstacle, creating a feedback loop where obstacle characteristics directly influence the choice of avoidance path parameters, improving measurement precision while managing complexity through structured decision rules

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3584141B1Object avoidance during trained parking by incorporating reverse planning
Publication Date: 2022.06.15 VALEO SCHALTER & SENSOREN GMBH
  • EP3584141B1 patent drawingFigure 1
  • EP3584141B1 patent drawingFigure 2
  • EP3584141B1 patent drawingFigure 3

AI summary

The present invention relates to a method for fully or partially autonomously driving a reference trajectory (6) with a vehicle (1, 2, 3), wherein, upon detection of an obstacle (5) in the area of ​​the reference trajectory (6) by means of at least one sensor, an alternative trajectory (7, 8) is driven, wherein as the starting point of the alternative trajectory a waypoint already driven back in the opposite direction of movement of the vehicle (2) is selected and the vehicle (2) is driven back to the starting point of the alternative trajectory (7) in the opposite direction of movement and then on the alternative trajectory (7) past the obstacle (5) in the original direction of movement onto the reference trajectory (6).