Obstacle Avoidance Trajectory Using Sinusoidal Curvature Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving systems struggle to generate collision-free trajectories in emergency maneuvers, failing to utilize vehicle grip limits, calculate quickly, and ensure kinematic and dynamic feasibility, leading to unsafe and unstable maneuvers.
Innovation Solution
A method and system for generating obstacle avoidance trajectories that consider vehicle grip limits, calculate trajectories rapidly, and ensure dynamic feasibility by using sinusoidal and constant curvature patterns within the vehicle's dynamic response limits, allowing for safe and efficient lane changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional trajectory generation algorithms are used, then collision-free paths can be generated, but the calculation time is too long for emergency maneuvers
Solution Approach 1:
The trajectory generation is segmented into two distinct phases: a fast phase that generates an initial collision-free trajectory using simplified assumptions, and a refinement phase that optimizes the trajectory considering vehicle dynamics. This segmentation allows the system to first ensure safety by quickly avoiding collisions, then improve maneuver quality without exceeding computational time limits for emergency situations.
Solution Approach 2:
The method performs preliminary action by first generating a trajectory that satisfies basic collision-free requirements before refining it. The initial trajectory is calculated using simplified models that can be computed rapidly, establishing a safe baseline path that is then optimized in subsequent steps considering more complex vehicle dynamics and grip limits.
2Reliability
If trajectory optimization considers vehicle dynamics, then maneuver quality improves, but computational complexity increases
Solution Approach 1:
The system applies dynamics by progressively incorporating vehicle dynamic constraints into the trajectory optimization. After generating an initial trajectory, the method refines it by considering vehicle grip limits, lateral acceleration constraints, and steering angle limitations. This dynamic approach ensures the final trajectory is not only collision-free but also dynamically feasible for the specific vehicle.
Solution Approach 2:
The optimization process changes parameters iteratively, starting with basic geometric constraints and progressively incorporating dynamic parameters such as lateral acceleration limits, steering rate constraints, and friction circle limitations. This parameter changes approach allows the system to balance computational complexity with dynamic feasibility by adding complexity only as needed.
3Reliability
If emergency avoidance maneuvers are executed, then collision risk is reduced, but vehicle stability may be compromised
Solution Approach 1:
The method applies beforehand cushioning by incorporating vehicle stability constraints and grip limits into the trajectory optimization before executing the maneuver. By pre-defining acceptable lateral acceleration limits and steering angle constraints based on vehicle dynamics, the system ensures that emergency avoidance maneuvers remain within stable operating boundaries, cushioning against potential instability.
Solution Approach 2:
The system uses feedback by continuously monitoring vehicle state and adjusting the trajectory to maintain stability. The optimization process incorporates feedback from vehicle dynamics models to ensure that commanded maneuvers do not exceed stable operating limits, allowing collision avoidance while maintaining vehicle control and stability throughout the maneuver.
Data Source
Figure 1~3
Figure 2
Figure 4(a)~5
AI summary
A method is described for the generation of a trajectory for the avoidance of a detected obstacle for a land vehicle driving along a planned path, particularly in a scenario of assisted or autonomous driving, wherein the avoidance trajectory includes at least one curvature with respect to a current reference direction of the vehicle, characterized in that the curvature has an amplitude that evolves over time with a sinusoidal pattern or with a composite pattern including sinusoidal and constant value sections, the sinusoidal pattern having a frequency lower than a predetermined maximum admissible frequency, lower than the maximum bandwidth of a control system of the dynamics of the vehicle arranged for the tracking of the avoidance trajectory by means of driving actuator means of the vehicle comprising at least one steering system.