Vehicle Trajectory Generation via Segmented Graph Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional algorithms for route selection in vehicle navigation are impractical due to the complexity of large road networks and are further compromised by factors like traffic, construction, and pedestrians, leading to inefficient routes and potential overload.
Innovation Solution
The generation of optimal vehicle trajectories is optimized by using processors to select travel segments based on operational metrics such as safety, reducing collisions, and minimizing changes in steering angle, which are ranked higher than other metrics like travel time, to ensure efficient and safe navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional algorithms are used for route selection in large road networks, then complete route coverage is achieved, but computational complexity becomes impractical
Solution Approach 1:
The patent segments the road network into a directed graph where intersections are nodes and road segments are edges. This segmentation transforms the complex problem of navigating entire road networks into manageable discrete units that can be processed efficiently by algorithms like A* or Dijkstra's, resolving the contradiction between complete route coverage and computational complexity.
Solution Approach 2:
The patent introduces a multi-dimensional cost function that evaluates routes based on multiple criteria simultaneously (travel time, distance, safety metrics, traffic conditions). By adding these additional evaluation dimensions, the system achieves comprehensive route assessment without requiring exhaustive exploration of all possible paths, thus managing computational complexity while maintaining reliability.
2Productivity
If greedy algorithms are used for route selection, then computational speed is improved, but route overload and traffic congestion increase
Solution Approach 1:
The patent incorporates real-time feedback mechanisms where operational metrics from multiple vehicles are continuously collected and used to update route cost functions. This feedback loop allows the system to dynamically adjust route recommendations based on current traffic conditions, preventing route overload while maintaining fast computation through efficient metric aggregation and propagation.
Solution Approach 2:
The patent dynamically changes the weight parameters in the cost function based on real-time operational metrics such as traffic density, vehicle speed, and safety conditions. By adjusting these parameters adaptively rather than using fixed greedy algorithms, the system achieves both computational efficiency and balanced route distribution across the network.
3Reliability
If multiple operational metrics are optimized equally, then comprehensive navigation quality is improved, but computational resources are excessively consumed
Solution Approach 1:
The patent changes the parameters of the cost function dynamically based on the current navigation context. For example, in safety-critical situations, safety metric weights are increased, while in normal conditions, travel time or distance weights dominate. This adaptive parameter adjustment maintains comprehensive navigation quality by considering all metrics while consuming fewer computational resources by focusing on the most relevant metrics for each situation.
Solution Approach 2:
The patent implements partial optimization by selectively evaluating and optimizing only the most critical operational metrics for each specific navigation scenario rather than equally optimizing all metrics simultaneously. This approach maintains adequate navigation quality across all dimensions while significantly reducing computational energy consumption by avoiding unnecessary calculations for less critical metrics.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Techniques for generation of an optimal trajectory for a vehicle include receiving, using one or more processors of the vehicle, an instruction for the vehicle to travel from an initial spatiotemporal location to a destination spatiotemporal location. One or more processors are used to generate a trajectory including travel segments. The trajectory begins at the initial spatiotemporal location and terminates at the destination spatiotemporal location. Each travel segment begins at a first spatiotemporal location and terminates at a second spatiotemporal location. Each travel segment is associated with operational metrics. The operational metrics are associated with navigating the vehicle from the first spatiotemporal location to the second spatiotemporal location. Each operational metric is optimized across the travel segments to generate the trajectory. Using a control module of the vehicle, the vehicle is navigated from the initial spatiotemporal location to the destination spatiotemporal location along the trajectory.