Locked Pedestrian Heading Detection for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving vehicles face challenges in accurately detecting and predicting pedestrian movements, particularly when pedestrians turn or swing their arms, leading to erroneous perceptions of changing heading directions and potential collisions.
Innovation Solution
A system that uses image capturing devices to determine a pedestrian's facing direction and projects them onto the nearest lane, locking their heading direction based on a predetermined condition to minimize false predictions and ensure safe navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the system tracks pedestrian heading direction based on real-time sensor data, then it can detect actual pedestrian movements, but it erroneously detects changing heading when pedestrians turn or swing arms, leading to false collision predictions
Solution Approach 1:
The system performs preliminary classification of pedestrian movements by analyzing whether detected heading changes correspond to actual direction changes or natural body movements. It pre-establishes criteria to distinguish between meaningful navigation changes and erroneous detections from arm swinging or body turning, applying these rules before generating collision predictions to filter out false positives
Solution Approach 2:
The system introduces an intermediary analysis layer between raw sensor data and collision prediction. This intermediary module processes heading detection data through additional validation steps, cross-referencing multiple sensor inputs and temporal patterns to verify whether detected heading changes represent true directional changes before passing information to the prediction system
2Productivity
If the system responds to every detected heading change, then it can react to actual pedestrian direction changes, but it produces erratic planning trajectories due to false detections from natural pedestrian movements
Solution Approach 1:
The system performs preliminary filtering of heading change detections before generating trajectory adjustments. It pre-establishes validation criteria that assess whether detected changes warrant trajectory modifications, filtering out false positives from natural movements while maintaining rapid response to genuine directional changes through pre-computed response protocols
Solution Approach 2:
The system dynamically adjusts its response threshold based on contextual factors such as pedestrian location, speed, and environmental conditions. Rather than applying fixed response rules, it modulates sensitivity levels to match situational requirements, maintaining trajectory stability during normal pedestrian behavior while enabling rapid adaptation when genuine directional changes occur
Data Source
AI summary
Embodiments is disclosed to detect a locked heading direction of a pedestrian and to predict a path for the pedestrian using the locked heading direction. According to one embodiment, a system perceives an environment of an autonomous driving vehicle (ADV) using one or more image capturing devices. The system detects a pedestrian in the perceived environment. The system determines a facing direction of the pedestrian relative to the ADV as one of left/right side, front, or back. If the facing direction of the pedestrian is determined to be front or back facing, the system determines a lane nearest to the pedestrian. The system projects the pedestrian onto the nearest lane to determine a lane direction at the projection. The system determines a heading direction for the pedestrian locking to the lane direction of the nearest lane based on a predetermined condition.