Vehicle Recognition Control for Adaptive Pedestrian Path Blocking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in determining whether pedestrians will cross a crosswalk, leading to overly cautious path estimation that can block their own path and result in unnecessary route recalculations and avoidance maneuvers.
Innovation Solution
A recognition device and method that includes a recognition unit for identifying objects in the vehicle's path, a determination unit to assess path blockage, an acquisition unit to gather information on object movement from an operator, and a change unit to adjust the estimated movement region based on this information, allowing for more precise path management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the autonomous vehicle estimates a sufficiently large region for pedestrian movement to prevent accidents, then safety is improved, but the vehicle's own path may be blocked and movement is interfered with
Solution Approach 1:
The system dynamically adjusts the estimated movement region based on the detected state of the object. When a pedestrian is detected as stationary (not moving), the system reduces the estimated movement region compared to when the pedestrian is detected as moving. This dynamic adjustment resolves the contradiction by making the safety buffer adaptive rather than static, allowing the vehicle to maintain safety while minimizing path blockage.
2Reliability
If the autonomous vehicle uses a large estimated movement region to ensure safety, then accident prevention is improved, but unnecessary avoidance travel and route recalculations occur
Solution Approach 1:
The system dynamically adjusts the estimated movement region based on the detected state of the object. When a pedestrian is detected as stationary (not moving), the system reduces the estimated movement region compared to when the pedestrian is detected as moving. This dynamic adjustment resolves the contradiction by making the safety buffer adaptive rather than static, allowing the vehicle to maintain safety while minimizing path blockage.
Solution Approach 2:
The system changes the parameter of the estimated movement region size based on the detected state of the pedestrian. By switching between different size parameters (large when moving, small when stationary), the system optimizes both safety and efficiency, preventing unnecessary avoidance maneuvers while maintaining adequate safety margins when needed.
3Reliability
If the autonomous vehicle frequently recalculates travel path in response to detected objects, then safety is improved, but system efficiency deteriorates
Solution Approach 1:
The system applies different levels of path recalculation based on the local situation. When a stationary pedestrian is detected, the system makes minimal adjustments to the travel path rather than performing full recalculation. This localized response maintains safety while preserving system efficiency by avoiding unnecessary computational overhead.
Solution Approach 2:
The system changes the parameter of path recalculation frequency and depth based on the detected object state. For stationary objects, it uses minimal recalculation parameters, while for moving objects it employs more comprehensive recalculation. This adaptive parameter adjustment resolves the contradiction between safety and efficiency.
Data Source
AI summary
In a recognition device, a recognition unit recognizes an object present in a direction of travel of an own vehicle and an estimated movement region of the object, a determination unit determines whether a path of the own vehicle is blocked by the object by determining whether the estimated movement region of the object overlaps with the path of the vehicle, based on a recognition result obtained by the recognition unit, in response to the determination unit determining that the path is blocked, an acquisition unit acquires information indicating whether the object is a moving object or a stationary object from an operator, and a change unit changes the estimated movement region recognized by the recognition unit, based on the information acquired by the acquisition unit.