Jaywalker Trajectory Prediction Using Entry-Point Likelihoods
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in predicting the trajectories of uncertain road users in urban environments due to the complexity of understanding their behavior, which is influenced by demographics, traffic dynamics, and environmental conditions, leading to difficulties in navigating and avoiding potential collisions.
Innovation Solution
The system generates trajectories for uncertain road users by detecting them within a drivable area and using perception information to identify potential entry points on the drivable area boundary, analyzing features such as distance, time, and directional alignment to determine the likelihood of entry, and controlling navigation to avoid collisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the autonomous vehicle uses detailed map information and contextual information to predict URU trajectories, then the prediction accuracy improves, but the computational complexity and processing time increase
Solution Approach 1:
The system segments the drivable area boundary into multiple discrete entry points, allowing the prediction algorithm to focus on specific locations rather than continuously analyzing the entire boundary. This segmentation enables more manageable computation while maintaining prediction accuracy by evaluating features at each discrete entry point independently
Solution Approach 2:
The system performs preliminary identification of potential entry points and calculates relevant features (distance, time, directional alignment) before generating trajectory predictions. This preliminary processing organizes data in advance, reducing the computational burden during real-time prediction and improving processing efficiency
2Reliability
If the system analyzes multiple features (distance, time, directional alignment) for each entry point, then the prediction reliability improves, but the processing time increases
Solution Approach 1:
The system applies different feature analysis depths to different entry points based on their relevance and risk level. Entry points closer to the vehicle's path or with higher predicted likelihood receive more detailed feature analysis, while less critical entry points use simplified evaluation, optimizing the balance between reliability and processing time
Solution Approach 2:
The system dynamically adjusts the threshold for feature analysis based on current driving conditions, vehicle speed, and environmental factors. When the vehicle is moving slowly or in high-risk zones, more comprehensive feature analysis is performed; when moving quickly in low-risk zones, analysis is streamlined, reducing processing time while maintaining adequate reliability
3Adaptability or versatility
If the autonomous vehicle identifies multiple potential entry points on the drivable area boundary, then the coverage of prediction improves, but the device complexity increases
Solution Approach 1:
The drivable area boundary is divided into discrete entry point segments, transforming a continuous complex boundary analysis problem into manageable discrete point evaluations. This segmentation enables comprehensive coverage of all potential entry locations while simplifying the algorithm by treating each point independently with standardized feature calculation procedures
Solution Approach 2:
The system identifies and analyzes a finite set of potential entry points that covers the necessary prediction scope, rather than attempting to analyze every possible location on the boundary. This partial action approach provides sufficient prediction coverage for safety while avoiding the excessive complexity of exhaustive analysis
Data Source
AI summary
Methods and systems for controlling navigation of a vehicle are disclosed. The system will first detect a URU within a threshold distance of a drivable area that a vehicle is traversing or will traverse. The system will then receive perception information relating to the URU, and use a plurality of features associated with each of a plurality of entry points on a drivable area boundary that the URU can use to enter the drivable area to determine a likelihood that the URU will enter the drivable area from that entry point. The system will then generate a trajectory of the URU using the plurality of entry points and the corresponding likelihoods, and control navigation of the vehicle while traversing the drivable area to avoid collision with the URU.