Autonomous Vehicle Path Planning Using Posterior Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous driving systems face challenges in accurately predicting the future movement of moving objects due to insufficient consideration of interactions between the autonomous vehicle and other objects, leading to inadequate motion planning and potential collisions.
Innovation Solution
The system employs a prediction module that performs posterior analysis based on the current state of the autonomous vehicle, including relative location, speed, and heading direction, to adjust object path predictions using machine-learning models and cost functions, calculating centripetal acceleration and collision costs to determine the most likely trajectory of moving objects and plan an optimal vehicle path to avoid collisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional prediction methods are used to predict moving object trajectories, then the prediction process is simple and fast, but the prediction accuracy is insufficient and collision avoidance is inadequate
Solution Approach 1:
The prediction system is segmented into multiple independent modules: a prediction module that generates initial trajectories, a cost calculation module that computes collision and centripetal acceleration costs, and a planning module that selects optimal paths. This segmentation allows each module to specialize in specific tasks, improving overall prediction accuracy while maintaining computational efficiency through modular architecture.
Solution Approach 2:
The system performs preliminary cost calculations for multiple candidate trajectories before final path selection. By pre-computing collision costs and centripetal acceleration costs for various possible paths, the system prepares prediction data in advance, enabling faster and more accurate real-time decision-making without increasing operational complexity.
2Reliability
If interaction between autonomous vehicle and moving objects is not considered, then the prediction process is simpler, but the motion planning is inadequate and collisions may occur
Solution Approach 1:
The system incorporates feedback mechanisms by calculating collision costs based on the autonomous vehicle's current state and predicted moving object trajectories. The cost function continuously evaluates the interaction between the vehicle and moving objects, adjusting path predictions in real-time to avoid collisions while accounting for dynamic environmental factors.
Solution Approach 2:
The prediction model dynamically adjusts parameters such as centripetal acceleration costs and collision probabilities based on the autonomous vehicle's state (position, speed, heading) and the detected moving objects' characteristics. This parameter adaptation allows the system to reliably predict interactions without requiring overly complex fixed models.
Data Source
AI summary
In response to perceiving a moving object, one or more possible object paths of the moving object are determined based on the prior movement predictions of the moving object, for example, using a machine-learning model, which may be created based on a large amount of driving statistics of different vehicles. For each of the possible object paths, a set of trajectory candidates is generated based on a set of predetermined accelerations. Each of the trajectory candidates corresponds to one of the predetermined accelerations. A trajectory cost is calculated for each of the trajectory candidates using a predetermined cost function. One of the trajectory candidates having the lowest trajectory cost amongst the trajectory candidates is selected. An ADV path is planned to navigate the ADV to avoid collision with the moving object based on the lowest costs of the possible object paths of the moving object.