Autonomous Vehicle Speed Planning for Spatial-Temporal Derate Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion planning systems for autonomous vehicles face inaccuracies in transforming spatial derate intervals to the temporal domain, leading to premature or delayed slowing down due to approximate speed profile inaccuracies, affecting ride comfort and collision avoidance.
Innovation Solution
The method involves generating a spatial domain speed profile by combining derate intervals with the spatial domain speed profile, determining minimum speeds or weighted averages along the path, and transforming this derate profile accurately into the temporal domain to ensure precise speed control, using a processor-based system that considers occlusions, map, and perception information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If an approximate speed profile is used to transform spatial derate intervals to the temporal domain, then the transformation process is simplified and can be executed efficiently, but the transformation accuracy deteriorates leading to premature or delayed speed adjustments
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing derate profiles in the spatial domain before motion planning execution. These pre-computed profiles contain advance information about required speed reductions at specific spatial locations, which are then transformed to the temporal domain using the approximate speed profile. This approach allows the system to prepare transformation data in advance, improving both efficiency and accuracy during real-time execution.
2Reliability
If accurate speed profile transformation is achieved, then collision avoidance and ride comfort are improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent reduces computational complexity during real-time execution by pre-computing derate profiles in the spatial domain and storing them for later use. During motion planning, the system only needs to transform these pre-prepared profiles to the temporal domain using a simple approximate speed profile, rather than performing complex iterative calculations. This preliminary preparation shifts computational burden to an offline or pre-processing stage.
Solution Approach 2:
The patent introduces derate profiles as an intermediary data structure that bridges spatial path information and temporal speed control. These profiles serve as a mediator that contains pre-processed derating information, allowing the system to avoid direct complex transformations between spatial and temporal domains during real-time operation. The intermediary profile simplifies the coupling between path geometry and speed planning.
3Measurement precision
If derate intervals are determined based on comprehensive information including occlusions, map data, and perception information, then the accuracy of speed planning is improved, but the information processing time and system complexity increase
Solution Approach 1:
The patent applies preliminary action by determining derate intervals and creating derate profiles in advance using comprehensive information from occlusions, map data, and perception sources. This pre-processing step consolidates information from multiple sources before motion planning execution, so that during real-time operation the system only needs to transform the already-prepared derate profiles to the temporal domain, reducing processing time during critical execution phases.
Data Source
Figure 1A~1B
Figure 1C
Figure 2
AI summary
Systems and methods for generating a trajectory of a vehicle are disclosed. The methods include generating a spatial domain speed profile for a path represented as a sequence of poses of a vehicle between a first point and a second point. Generation of the spatial domain speed profile uses a longitudinal problem while excluding a derate interval on the path when speed of the vehicle cannot exceed a threshold. The methods further include generating a derate profile using the spatial domain speed profile and the derate interval, and transforming the derate interval to a temporal derate interval. The temporal derate interval may include time steps at which the speed of the vehicle cannot exceed the threshold while traversing the path, and may be used as an input to the longitudinal problem for generating a trajectory of the vehicle for navigating between the first point and the second point.