Autonomous Vehicle Trajectory Planning in Spatio-Temporal Convex Corridors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion planning systems for autonomous vehicles face challenges in optimizing the spatio-temporal search space due to time-dependent constraints from dynamic objects, leading to suboptimal trajectory solutions and increased computation costs when using rectangular or trapezoidal corridors.
Innovation Solution
The use of spatio-temporal convex corridors with complex shapes allows for increased search space coverage and optimized trajectory planning by ensuring that Bezier curve segments remain within the convex hull, reducing computation costs and improving vehicle operation efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional motion planning algorithms are used, then motion plans can be generated for robotic systems, but the computational complexity increases exponentially with the number of degrees of freedom
Solution Approach 1:
The patent segments the high-dimensional motion planning problem into multiple lower-dimensional subproblems by decomposing the configuration space into manageable components. This allows the system to handle complex robotic systems with many degrees of freedom by solving simpler subproblems independently and combining their solutions, thereby reducing overall computational complexity while maintaining motion planning capability.
Solution Approach 2:
The patent transforms the complex high-dimensional motion planning problem into a lower-dimensional problem by projecting the configuration space onto a reduced-dimensional manifold or using dimensionality reduction techniques. This enables efficient computation by solving the planning problem in fewer dimensions while still capturing the essential motion characteristics of the robotic system.
2Reliability
If motion plans are generated for robotic systems with many degrees of freedom, then comprehensive motion coverage is achieved, but the planning time increases significantly
Solution Approach 1:
The patent performs preliminary computations to pre-process the configuration space, pre-compute valid motion sequences, or create lookup tables of feasible motions before actual planning is needed. This preliminary action reduces the computational burden during real-time planning, enabling the system to generate complete motion plans faster by leveraging pre-computed information.
Solution Approach 2:
The patent changes key parameters of the motion planning algorithm, such as adjusting the resolution of the configuration space grid, modifying search depth limits, or changing cost function weights, to optimize the balance between planning completeness and planning time. These parameter adjustments allow the system to achieve sufficient motion coverage within acceptable timeframes.
3Manufacturing precision
If exact motion planning algorithms are used, then optimal motion plans are guaranteed, but the computational resources required become prohibitive
Solution Approach 1:
The patent applies partial action by using approximation algorithms or heuristic search methods that do not guarantee complete optimality but provide sufficiently good motion plans with much lower computational resource consumption. This approach accepts some loss in optimality in exchange for dramatically reduced energy and computational requirements, making the system practical for real-world deployment.
Data Source
Figure 1
Figure 2A~2B
Figure 3~4
AI summary
Systems, methods and computer-readable media for path planning for an autonomous vehicle, comprising receiving data defining an upper bound of a spatio-temporal convex corridor and lower bound of the spatio-temporal convex corridor; computing a plurality of control point times within a corridor start time and a corridor end time, the control point times including at least a start control point time, an end control point time and an intermediate control point time between the start control point time and the end control point time; determining, for each of the control point times, respective control point maximum values and control point minimum values; computing, based on the respective control point maximum values and the respective control point minimum values, respective control point values for each of the control point times, the control point values defining a curve segment that is within the upper bound and the lower bound of the spatio-temporal convex corridor.