Vehicle Trajectory Prediction for Collision-Risk Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current planning systems in autonomous and semi-autonomous vehicles face challenges in reliably determining actions due to the complexity of their systems, especially in intricate scenarios, which can lead to inefficiencies in processing and safety concerns.
Innovation Solution
Implementing a model on a vehicle computing device to predict trajectories of objects and vehicles, determining potential intersections and actions based on sensor data, map data, and control policies, utilizing parallel processing to enhance accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the planning system processes all detected objects and their trajectories to determine vehicle actions, then the reliability of collision avoidance improves, but the computational complexity and processing time increase significantly
Solution Approach 1:
The patent extracts and focuses computational resources on only those objects that pose a potential collision risk, rather than processing all detected objects. The system identifies objects with trajectories that may intersect with the vehicle's path and directs detailed analysis only to these relevant objects, eliminating unnecessary computational overhead from objects that cannot possibly collide.
Solution Approach 2:
The system performs preliminary filtering of object trajectories to identify potential collision risks before conducting detailed analysis. By pre-processing sensor data to flag only those objects whose trajectories intersect with the vehicle's intended path, the system prepares a reduced set of candidates for further evaluation, significantly reducing the computational burden on the planning system.
2Measurement precision
If the system analyzes all object trajectories in detail to ensure safety, then the measurement precision of collision risk assessment improves, but the processing time increases
Solution Approach 1:
The patent applies partial analysis by performing detailed trajectory intersection calculations only for objects that pass an initial screening threshold. Rather than analyzing all objects with equal depth, the system performs a first pass to identify potential risks, then applies more computationally intensive precision analysis only to those specific cases where collision cannot be ruled out by simpler criteria.
3Adaptability or versatility
If the planning system considers multiple potential actions for the vehicle, then the adaptability to complex scenarios improves, but the computational resources required increase
Solution Approach 1:
The system dynamically adjusts the level of analysis and number of candidate actions generated based on the complexity and risk level of the current scenario. In low-risk situations, the system uses simplified planning with fewer candidate actions to conserve computational energy. In high-risk or complex scenarios involving multiple objects with intersecting trajectories, the system automatically increases the depth of analysis and considers more potential actions, allocating computational resources adaptively to match the actual demand.
Data Source
AI summary
Techniques for accurately predicting and avoiding collisions with objects detected in an environment of a vehicle are discussed herein. A vehicle computing device can implement a model to output data indicating costs for potential intersection points between the object and the vehicle in the future. The model may employ a control policy and a time-step integrator to determine whether an object may intersect with the vehicle, in which case the techniques may include predicting vehicle actions by the vehicle computing device to control the vehicle.