Autonomous Vehicle Trajectory Planning With Spatial-Temporal Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing trajectory planning methods for autonomous vehicles often result in jerky paths due to non-continuous trajectories, are time-consuming when managing obstacles in real-time, and require predefined global waypoints, while graph-based approaches can be computationally intensive and lack effective incorporation of road and kinematic constraints.
Innovation Solution
A novel spatial-temporal graph neural network architecture that processes graphs with virtual nodes positioned along the road and kinematic constraints, incorporating personalized potential fields for obstacle avoidance and strategic decision-making, using a neural network to generate smooth and safe trajectories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If graph-based approaches are used for trajectory planning, then the ability to handle real-time obstacles is improved, but computational complexity increases making the system time-consuming
Solution Approach 1:
The graph is segmented into spatial components (nodes representing positions) and temporal components (edges representing transitions over time steps). This segmentation allows the system to process trajectory planning in discrete, manageable units that can be evaluated efficiently for real-time obstacle handling while reducing overall computational burden.
Solution Approach 2:
The graph structure is made dynamic by allowing nodes and edges to be added, removed, or modified based on real-time obstacle detection and changing environmental conditions. This dynamic adaptation enables the system to respond to real-time obstacles without requiring complete recomputation of the entire trajectory plan.
2Productivity
If traditional trajectory planning methods are used, then computational efficiency is improved, but the resulting paths are jerky and not continuous
Solution Approach 1:
The spatial-temporal graph ensures continuity by requiring that trajectory transitions follow connected edges between nodes, representing continuous spatial-temporal paths. This structure guarantees smooth, continuous trajectories while maintaining computational efficiency through graph-based search algorithms.
3Stability of the object's composition
If interpolating curve planners are used to generate smooth trajectories, then trajectory continuity is improved, but the requirement for global waypoints makes the system time-consuming for real-time obstacle management
Solution Approach 1:
The system transitions from traditional 2D spatial trajectory planning to 3D spatial-temporal planning by adding the time dimension as the third axis. This allows the graph to represent not only spatial paths but also temporal evolution of the trajectory, enabling smooth continuous paths to be generated and updated in real-time based on obstacle conditions without requiring global waypoint redefinition.
Data Source
AI summary
An autonomous vehicle has several modules responsible for one or more of the aforementioned items, including a trajectory planner which plays a pivotal role in the safety of the vehicle, the comfort of its passengers, for respecting kinematic constraints and any applicable road constraints. A spatial-temporal graph trajectory planner generates safe and comfortable trajectories. A spatial-temporal graph uses the autonomous vehicle, its surrounding vehicles, and virtual nodes along the road. The graph is then forwarded into a sequential network to obtain the desired states. A simple behavioural layer is also presented that determines kinematic constraints for the planner and a novel potential function trains the network. The proposed planner is tested on three different complex driving tasks and the performance is compared with two frequently used methods. The planner generates safe and feasible trajectories, while achieving similar or longer distance in the forward direction and comparable comfort ride.