Vehicle Steering Trajectory Smoothing for Tight Obstacle Maneuvers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for steering vehicles autonomously in tight environments fail to account for time and dynamically adjust to obstacles, leading to potential collisions and inefficiencies, as they rely on geometrical path planning without considering the vehicle's physical capabilities and comfort constraints.
Innovation Solution
A method that computes joint lateral and longitudinal guidance trajectories, incorporating time-stamped segments and dynamic state variables like velocity and acceleration, to ensure collision-free and comfortable vehicle movement, using a structure-exploiting nonlinear programming solver for real-time implementation, and includes distance computation, vehicle modeling, path planning, and path improvement to account for static and moving obstacles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If geometrical path planning is used without considering vehicle physical capabilities, then path computation is simpler, but trajectory feasibility and safety are compromised
Solution Approach 1:
The patent transforms static geometrical path planning into dynamic trajectory optimization by incorporating time-stamped segments and vehicle dynamic states (velocity, acceleration, jerk). The trajectory is computed as a sequence of time-parameterized states that evolve according to vehicle dynamics models, allowing the system to adapt to changing conditions while ensuring physical feasibility at each time step.
Solution Approach 2:
The patent changes the parameterization of the trajectory from pure spatial coordinates to time-stamped states including position, velocity, acceleration, and jerk. This parameter transformation allows the optimization to directly enforce vehicle physical constraints (maximum acceleration, jerk limits) while maintaining computational tractability through structured parameterization.
2Loss of time
If traditional path planning methods are used, then computation time is reduced, but obstacle avoidance and collision prevention are insufficient
Solution Approach 1:
The patent performs preliminary computation of a collision-free reference trajectory using optimized path planning algorithms before executing the maneuver. This reference trajectory serves as a constrained optimization problem that guarantees obstacle avoidance, and subsequent real-time control only needs to track this pre-computed safe path with minor adjustments, reducing real-time computation burden while maintaining safety.
Solution Approach 2:
The patent extends the planning problem from 2D spatial path planning to 4D trajectory optimization by adding time and vertical dynamics dimensions. This dimensional expansion allows the system to compute comprehensive collision-free trajectories that account for temporal evolution and vehicle dynamics, while the structured optimization approach maintains computational efficiency through exploitation of problem structure.
3Reliability
If joint lateral and longitudinal trajectory optimization is performed, then vehicle comfort and safety are improved, but computational complexity increases
Solution Approach 1:
The patent segments the trajectory optimization into discrete time-stamped segments where each segment connects successive states. This segmentation transforms the continuous optimization problem into a sequence of constrained sub-problems that can be solved efficiently using dynamic programming or sequential quadratic programming, reducing overall computational complexity while maintaining joint optimization of lateral and longitudinal dynamics.
Solution Approach 2:
The patent creates a universal trajectory optimization framework that simultaneously handles multiple objectives (safety, comfort, efficiency) and multiple vehicle dynamics constraints (lateral and longitudinal) through a unified cost function and constraint formulation. This multi-functional approach avoids separate optimization passes and achieves comprehensive control in a single computational framework.
4Device complexity
If path planning does not consider time factors, then planning is simpler, but maneuvering efficiency and responsiveness are reduced
Solution Approach 1:
The patent implements periodic recalculation and optimization of the trajectory at discrete time intervals, allowing the system to adapt to changing environmental conditions and vehicle states. This periodic update mechanism maintains planning simplicity at each time step while achieving overall maneuvering efficiency through cumulative optimization over the maneuver duration.
Data Source
AI summary
A method for steering a vehicle along a path in a driveway and around obstacles between a starting position into a target position, comprises the steps of determining the vehicle dimensions, steering and driving capabilities, carrying out a path optimization step to evaluate, based on a predetermined cost function, the least costly path between the starting position and the target position avoiding any collisions with obstacles. The method further comprises the further step of applying a path improver step, smoothening the trajectory obtained by the path optimization method by means of numerical optimization while fulfilling dynamical constraints on acceleration and steering rate of the vehicle through planning lateral and longitudinal movement of the vehicle in a joint optimization problem or by means of separate optimization problems.