Trajectory Determination for Autonomous Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional self-driving vehicles lack defined logic for decision-making in complex driving environments, leading to safety concerns due to uncertainty in behavior and inability to handle changes such as incoming pedestrians or moving vehicles, which affects their ability to plan safe trajectories.
Innovation Solution
A trajectory determination method that integrates a processing module with a trajectory generation module, obstacle detection module, lane space detection module, and road information module to assess time to collision, time gap, intersection, and safety flags, determining a target trajectory based on speed limits, block, intersection, lane-change, and safety flags to enhance safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional self-driving vehicles make decisions based only on current environment without defined logic, then the system complexity is reduced, but the safety and reliability of trajectory planning deteriorates
Solution Approach 1:
The decision-making process is segmented into multiple evaluation dimensions including time to collision (TTC), time gap, prediction time to collision (PTTC), block flags, intersection flags, lane-change flags, and safety flags. Each dimension independently evaluates a specific aspect of trajectory safety, and their combined results determine the final trajectory decision. This segmentation allows comprehensive safety assessment without requiring a single complex decision-making model.
2Adaptability or versatility
If conventional self-driving vehicles use fixed routes in enclosed areas, then the system reliability is improved, but the adaptability to general road environment deteriorates
Solution Approach 1:
The system dynamically adjusts trajectory planning based on real-time environmental conditions by continuously evaluating multiple flags and parameters. Unlike fixed routes, the system adapts its decision-making to current situations such as pedestrian presence, vehicle movements, and road conditions. The dynamic evaluation of TTC, time gap, PTTC, and various flags enables the system to maintain behavioral consistency through defined logic while adapting to diverse general road environments.
3Reliability
If conventional self-driving vehicles do not consider intended trajectory of obstacles, then the computational load is reduced, but the safety of planned trajectory deteriorates
Solution Approach 1:
The system performs preliminary evaluation of obstacle trajectories by calculating prediction time to collision (PTTC) and intended trajectory intersections before finalizing the vehicle's path. By assessing whether obstacle paths intersect with candidate trajectories in advance and computing potential collision times, the system identifies unsafe trajectories early in the evaluation process, eliminating the need for extensive computational analysis of obviously dangerous paths and focusing resources on viable options.
Data Source
AI summary
A trajectory determination method for a vehicle is provided. A target vehicle trajectory is determined from among multiple candidate vehicle trajectories by considering, for each of the candidate vehicle trajectories, presence or absence of a front obstacle, presence or absence of a potentially-colliding obstacle, and a condition related to lane change, so as to enhance driving safety of the vehicle.